Podcasts about cloud act

Cloud computing solved for scale by centralizing everything into a handful of massive buildings. Edge computing is the argument that scale was never the only thing worth optimizing for.On this episode, Jeff MacMillan of Green Edge Computing Corp makes the case that sovereignty was never really about geography. It was about how many entities you'd have to strong-arm before someone gave in.The conversation moves from mining sites with unreliable wireless to a Cloud Act clause most people have never read. It lands on an uncomfortable question: when decentralizing compute means decentralizing oversight too, who ends up accountable for what happens in the box nobody's watching.The same hardware standard the U.S. military spent fifteen years refining now sits in clinics and 7-Elevens, and has implication and applications for bringing cutting edge software to the most remote areas. What does it mean when you can bring your own compute literally anywhere?Mentioned: Jevon's paradox

Vad betyder egentligen AI Act, Cloud Act och nya EU-regler för svenska kommuner och regioner? Julia Altenhofer, VP Public Affairs på Evroc, hjälper oss att navigera i den föränderliga världen av digital lagstiftning, molntjänster och AI – och vad utvecklingen innebär för framtidens välfärd. Inspelat på ett loft i Visby under Almedalsveckan.

Gerard and Laurent welcome Michel Boutouil, co-founder and CEO of Polarise, a leading European AI infrastructure provider and NVIDIA Cloud Partner based in Berlin. After discussing about what happens outside of a datacenter, it is time to dive inside one.  Polarise is one of the few genuinely European NeoCloud companies — essentially a European counterpart to CoreWeave — specializing in GPU infrastructure for AI inference. Through its partnership with NVIDIA, Polarise designs its datacenters around the GPU rack itself, using liquid cooling from the outset rather than starting with a traditional real estate-first approach. The company has already developed AI factories in Germany, Norway and the UK.  In the conversation, we explore the growing commoditization of large language models and why the real long-term value may lie in AI factories — facilities that are fundamentally different from conventional datacenters. Given Europe's notoriously long grid-connection timelines, Polarise focuses on refurbishing brownfield sites with under 50MW of grid access instead of pursuing massive gigawatt-scale campuses. It's a pragmatic “pod” strategy: adapt to the grid's constraints rather than try to reshape the entire energy system.  We also tackle the thorny issue of digital sovereignty. With the U.S. CLOUD Act allowing U.S. authorities access to data managed by American tech companies, it is fair to ask what hyperscalers are doing with European data — and whether Europe needs its own sovereign AI infrastructure. Polarise has secured €1 billion in backing from Swiss investor SWI Stoneweg Icona, but even that is modest compared with the hyperscalers' spending power. For comparison, SpaceX has reportedly invested around $40 billion in Colossus 1 and 2 alone.  So, what does the future of the European AI ecosystem look like? Michel's answer is clear: Europe should not try to outspend China or the United States head-on. Instead, it should play to its strengths — smart execution, agility, flexibility, and the ability to learn quickly from the mistakes being made elsewhere.    “Today's show is supported by the BMW Foundation Herbert Quandt. The BMW Foundation unites leaders across sectors to develop solutions that foster an innovative economy and a future-proof society. A key focus is "Energy Transition & Climate Change," where the Foundation drives "International collaboration to accelerate the energy transition." With rising energy demands from AI and data centres, new partnerships, effective collaboration, and the exchange of science-based solutions and strategies are essential.”  

Spezialgast David Rosenthal und Martin Steiger diskutieren Rechtsfragen rund um die Nutzung von KI in der Schweiz. Themen im zweiten Gesprächsteil sind unter anderem die Datensicherheit, der CLOUD Act, die Bedeutung von Bauchgefühl und Emotionen, und die Überforderung der Tech-Unternehmen.

L'Europa sta cercando di staccarsi dal cordone ombelicale tecnologico americano, ma il prezzo della sovranità è più alto di quanto sembri. In questo video di 12 minuti analizziamo il "Tech Sovereignty Package": dalle istituzioni che abbandonano Google e Microsoft per Qwant ed Euro-Office, fino al rischio legale del CLOUD Act sui tuoi dati personali. Ma c'è un paradosso che nessuno ti svela: mentre Bruxelles scrive leggi per l'indipendenza, i tuoi risparmi (tramite ETF come VWCE o MSCI World) stanno finanziando proprio quei giganti americani. La dipendenza digitale non è solo un tema politico, è l'architettura stessa dei tuoi investimenti. La disclosure dinamica aggiornata è disponibile qui: www.marcocasario.com/disclosure Vuoi lanciare il tuo negozio online?Vai su https://www.shopify.com/it e scopri come iniziare con Shopify in pochi minuti.

In dieser Folge dreht sich alles darum, wer was lesen kann – und wer nicht. Patrick bekommt eine verschlüsselte Mail von seinem Stromanbieter und kann sie nicht öffnen. Die Niederlande stellen fest, dass Microsoft Behördennamen ungeschwärzt an den US-Senat weitergegeben hat – dank CLOUD Act völlig legal, aber trotzdem eine schlechte Idee. Und die US-Cybersicherheitsbehörde CISA legt AWS-Keys, Kubernetes-Configs und Klartext-Passwörter in einem öffentlichen GitHub-Repository ab. Das Repository heißt "Private-CISA". Verschlüsselung hilft. Manchmal. Wenn man's richtig macht.

L'Europe relance la bataille pour sa souveraineté numérique face aux GAFAM • Le Parlement européen adopte Qwant comme moteur par défaut • L'IA affronte le droit d'auteur et Mistral monte au créneau • Anthropic relance le débat sur une pause mondiale de l'IA • SoftBank investit 75 milliards d'euros dans des data centers en France • Microsoft dévoile ses nouveaux modèles et ses agents autonomes ⭐️ Découvrez Frogans à Vivatech 2026

C'est un véritable coup de théâtre auquel vient d'assister en France le secteur de l'enseignement supérieur.Sous pression, la prestigieuse École polytechnique vient d'annuler sa migration vers Microsoft 365.Le cadre juridique se durcit sérieusementCe revirement spectaculaire est d'abord le signe que le cadre juridique se durcit sérieusement. En tentant d'imposer l'écosystème Microsoft, Polytechnique s'est heurtée de plein fouet au Code de l'éducation, qui exige de prioriser les logiciels libres.Mais surtout, en migrant vers ces outils, l'école exposait les données sensibles de la recherche française aux lois extraterritoriales américaines, notamment le fameux Cloud Act.Sous la menace d'une action en justice menée par le Conseil National du Logiciel Libre, et face à la fronde de ses propres chercheurs, la direction a dû capituler.Ce dossier dépasse largement les murs de l'écoleMais attention, ce dossier dépasse largement les murs de l'école.Car au même moment, l'autorité de protection des données autrichienne édicte que la version éducation de Microsoft 365 viole purement et simplement le RGPD, notamment en pistant les utilisateurs à des fins commerciales.Il est donc temps de sortir du déni. Les promesses d'hébergement localisé ne suffisent plus lorsque la société mère reste contrainte de livrer ses données sur simple injonction d'un juge outre-Atlantique.Quelle alternative s'offre aux décideurs ?Face à ce constat, quelle alternative s'offre aux décideurs ? Longtemps perçu comme complexe, le logiciel libre pourrait servir de levier stratégique.Le Conseil National du Logiciel Libre assure structurer une véritable filière d'accompagnement.L'objectif est d'aider les organisations à migrer vers des outils ouverts, souverains et interopérables.Hébergé par Ausha. Visitez ausha.co/politique-de-confidentialite pour plus d'informations.

Ronald, Marco en Jelle zijn terug met DigiD, device-code-phishing, residential proxies en de vraag of AI cyberaanvallers echt onhoudbaar maakt. Eerst kort: Marco repareert tijdens een nachtwacht Home Assistant-data met Claude, Jelle bouwt met AI een lesdashboard, en Ronald rijdt in Kaapstad een fox hunt met antennes op de auto. Daarna DigiD. Staatssecretaris Willemijn Aerdts blokkeert de Amerikaanse overname van Solvinity door Kyndryl. Ronald legt uit waarom dit via de Wet ongewenste zeggenschap telecommunicatie loopt, waarom dat juridisch anders is dan VIFO, en waarom Nederland hiermee feitelijk zegt: Amerikaanse jurisdictie en CLOUD Act-risico's zijn voor DigiD te groot. Marco bespreekt RSI, recursive self-improvement, als nieuwe AI-hypeterm. Het idee: AI die zijn eigen training verbetert. De nuchtere conclusie blijft: losse stappen automatiseren lukt steeds beter, maar richting houden, controleren of iets klopt en echt autonoom onderzoek doen blijft lastig. Jelle pakt Kali365: phishing via Microsoft 365 device-code-flows. Het slachtoffer logt in op de echte Microsoft-site, maar autoriseert het apparaat van de aanvaller. Domeinchecken is dus niet genoeg als de context rond de login vergiftigd is. Het eerste hoofdverhaal: ASocks en residential proxies. Politie en NCSC verstoren een botnet met minstens 17 miljoen besmette apparaten, aangestuurd via ongeveer 200 servers in Nederland. Marco vat het scherp samen: het botnet is de infrastructuur, de residential proxy is het product. Aanvallers kopen verkeer vanaf normale thuisverbindingen in plaats van herkenbare datacenters of Tor-exitnodes. Daardoor lijken phishing, credential stuffing, DDoS en brute-force-pogingen op gewoon verkeer van echte gebruikers. Open vraag: zijn de apparaten echt opgeschoond, of vooral de aansturing geraakt? Jelle sluit af met Lennart Maschmeyers paper Deception and Detection. Maschmeyer stelt dat AI aanval en verdediging helpt, maar verdedigers structureel meer kunnen winnen: verdediging draait veel om detectie en patroonherkenning, aanval verderop in de kill chain om misleiding, context en gecontroleerde effecten. De drie zijn kritisch op zijn dwell-time-argument, maar herkennen de kern: je wilt geen autonome agent die in een vijandelijk netwerk creatief gaat improviseren. Tegelijk maakt AI aanvallers wel sneller als copiloot, codegenerator, parser van scanoutput en phishinghulp. Vooral lagere en middelmatige actoren kunnen daarmee sneller opschalen. *Bronnen* DigiD / Solvinity - NOS: https://nos.nl/artikel/2615885-staatssecretaris-verbiedt-amerikaanse-overname-solvinity-bedrijf-achter-digid - Wet OZT: https://wetten.overheid.nl/BWBR0045423 - Wet VIFO: https://wetten.overheid.nl/BWBR0046686 RSI - TechCrunch: https://techcrunch.com/2026/05/28/rsi-is-the-new-agi-and-its-just-as-hard-to-pin-down/ Kali365 - FBI IC3: https://www.ic3.gov/PSA/2026/PSA260521 - BleepingComputer: https://www.bleepingcomputer.com/news/security/fbi-warns-of-kali365-phishing-service-targeting-microsoft-365-accounts/ ASocks / residential proxies - Politie: https://www.politie.nl/nieuws/2026/mei/28/06-politie-en-ncsc-halen-groot-botnetwerk-offline.html - NCSC expertblog: https://www.ncsc.nl/expertblogs/residential-proxies-en-hun-grote-impact-op-de-digitale-veiligheid-in-nederland - NCSC nieuws: https://www.ncsc.nl/nieuws/gezamenlijke-actie-politie-en-ncsc-legt-groot-botnetwerk-plat - Security.nl: https://www.security.nl/posting/938396/Proxy-botnet+van+17+miljoen+apparaten+na+actie+politie+en+NCSC+offline?channel=rss Maschmeyer / AI - CV Maschmeyer: https://www.lennartmaschmeyer.com/CV_Lennart_Maschmeyer.pdf - Paper: https://doi.org/10.1162/isec.a.398 - M-Trends 2025: https://cloud.google.com/security/resources/m-trends

De Europese Commissie wil de cloudmarkt aanpakken en de eigen techsector een handje helpen. Op dit moment is deze markt voor bijna tweederde in handen van bedrijven zoals Amazon, Google en Microsoft. Onder de Cloud and AI Development Act wil Eurocommissaris Henna Virkkunen, die technologie in haar portefeuille heeft, helpen door woensdag nieuwe maatregelen aan te kondigen. Dat staat in een conceptversie waar persbureau Reuters over schrijft. Rosanne Peters vertelt erover in deze Tech Update. De belangrijkste reden is Europa minder afhankelijk te maken van technologie uit de Verenigde Staten. Dat is niet alleen cruciaal voor ons concurrentievermogen, maar ook voor onze veiligheid. Er blijft een reële angst dat de VS onder hun eigen Cloud Act toegang kan krijgen tot opgeslagen data. Verder in deze Tech Update: Apple stelt release van meerdere producten uit door Siri die - opnieuw - tekort schiet See omnystudio.com/listener for privacy information.

Umiestnenie amerických serverov v Európe je len falošný pocit bezpečia. Neslávne známy zákon Cloud Act jasne hovorí: Ak ste americká firma, musíte vydať dáta vašej vláde, bez ohľadu na to, v ktorej časti sveta sa fyzicky nachádzajú. Európa si preto konečne začína uvedomovať obrovské riziko, ktoré prináša masívne spoliehanie sa na služby od gigantov ako Amazon, Google či Microsoft.Už to dávno nie sú len teoretické hrozby. Tlak na Dánsko ohľadom anexie Grónska či tvrdé digitálne odstrihnutie Medzinárodného trestného súdu OSN ukázali, ako rýchlo môže Európa prísť o svoju infraštruktúru, ak nevlastní jej zdrojový kód. Liekom na túto závislosť majú byť „suverénne cloudy“ a masívny prechod na open-source riešenia. O tom, prečo je aj na Slovensku potrebné budovať vlastnú digitálnu nezávislosť a ako sa reťazec Lidl stal poskytovateľom cloudu, sa v podcaste SHARE rozpráva Maroš Žofčin s riaditeľom spoločnosti dNation Martinom Pilkom.Podcast vznikol v spolupráci so spoločnosťou dNation.Pripravte sa na budúcnosť s knihou od redaktorov Živé.sk „Umelá inteligencia: Pripravte sa na budúcnosť“. Teraz ju máme aj v elektronickej verzii. Nájdete ju na obchod.aktuality.sk.TIP: https://zive.aktuality.sk/clanok/0RfdZVW/nahliadnite-do-buducnosti-vydavame-knihu-o-umelej-inteligencii/V podcaste sa dozviete aj o týchto témach:Prečo fyzické umiestnenie servera v EÚ nestačí a ako funguje sporný zákon Cloud Act.Ako v priebehu 24 hodín úplne zmizla z internetu celá digitálna identita Medzinárodného trestného súdu.Prečo je otvorený zdrojový kód kľúčom k tomu, aby ste neboli rukojemníkom poskytovateľa licencií.Ako nástroje ako Nextcloud dokážu plnohodnotne nahradiť obľúbený Office 365 či Google Disk.Ako si známy európsky supermarket vybudoval cloud pre e-shop a dnes ho predáva ako infraštruktúru.Prečo slovenskí výskumníci stavajú vlastný privátny cloud s mimoriadne citlivými dátami.Podcast SHARE pripravuje magazín Živé.sk.

Nigel Brown, CTO of Microserve Not every voice at Dell Technologies World last week belonged to a vendor. For a partner perspective on the week’s biggest themes, In The Channel sat down with Nigel Brown, CTO of Microserve – a Burnaby, BC-based solution provider, Dell Titanium partner, and Dell’s Client Solutions Partner of the Year in Canada in consecutive years. Brown walked away from DTW with deskside agentic AI as his headline takeaway, particularly after hands-on time in a Dell lab showcasing NemoClaw – NVIDIA‘s enterprise-governance take on the OpenClaw open-source agent framework. “They’ve set it up closed by default – it can’t leave the box,” Brown says. “That’s a safety net that really opens the conversation.” That said, he’s clear-eyed about where most of his public sector and enterprise clients actually are. “Broad scope, it’s ahead. The hardware is going to follow it.” The tokenomics reality landed hard too. Brown shared a personal story about spending a hundred dollars testing Claude on a single flight – a relatable example he’s started using to frame the real cost implications of unmanaged AI usage, well before any on-premises or local inference conversation begins. On cyber resilience, Brown says he’s had to evolve his approach: “I got to be more of a jerk. I was being too nice.” His firm’s managed backup practice has seen firsthand the damage when clients – and even other MSPs – treat backup as a checkbox. When you show up after a ransomware event to find the backup server was on the same domain and hit just as hard, the conversation changes. And on Canadian data sovereignty, Brown goes beyond the standard data-residency talking points. FISA Section 702 and the CLOUD Act, he argues, represent far more serious legal exposure than most clients realize – even those who believe a Canadian cloud region is sufficient protection. The conversation also covers the AI PC refresh cycle colliding with supply chain pressure, the end-user adoption gap that’s undermining Copilot investments, and what Dell’s revised partner incentive structure signals about where the growth opportunities are. Read Full Transcript Robert Dutt: Hello and welcome to In the Channel from ChannelBuzz.ca, bringing news and information to the Canadian IT channel community for the last 16 years. I’m Robert Dutt, editor of ChannelBuzz.ca and your host for the show. Last week, I was at Dell Technologies World in Las Vegas, Dell’s big annual customer and partner event. Over the course of the week, I had a number of conversations that I’ll be bringing here on In the Channel. Last week, we featured three Dell executives. This week, we’re bringing you some partners. Today, we start on that partner perspective, specifically from one of Canada’s top Dell partners. Nigel Brown is CTO of Microserve, a Burnaby, BC-based solution provider that has earned Titanium status with Dell and taken home Dell’s Client Solutions Partner of the Year in Canada in consecutive years. Microserve serves an enterprise and public sector-heavy client base, which means Nigel’s job is regularly about taking what gets announced on a stage in Las Vegas and translating it into something that makes sense for organizations that don’t necessarily move at conference speed. I caught up with Nigel on site at DTW last week. We covered a lot of ground – deskside agentic AI and what it’s actually going to take to make that real for customers, the very real cost of token economics, why he’s had to be, as he put it, more of a jerk about cyber resilience, and why the Canadian data sovereignty conversation is more urgent than most people realize. Let’s get right to it. My chat with Nigel Brown. Nigel, thanks for taking the time. Appreciate it. Nigel Brown: Happy to be here. Thanks for having me. Robert Dutt: So you guys are here, obviously, as a Titanium-level Dell partner, consecutive years as the Client Solutions Partner of the Year in Canada. What’s your overall read on this week? What made your ears stand up? What caught your attention? What are you taking back to both your team and to your customers when you go back to Burnaby? Nigel Brown: That’s a really good question. It’s also a big one. There’s been a lot of announcements, a lot of dialogue over the last couple of days. I’m trying to process that a little bit, assuming you were going to ask me that. I think the biggest takeaway I had – everybody’s heard of OpenClaw, everybody’s heard all the IT people are terrified of it, so it’s more, how do we get rid of it in our environments? Seeing this whole push around deskside agentic AI, especially given our market where we play a lot with clients – I actually had the opportunity, I did the lab today because I couldn’t resist seeing what it’s like. The governance and security wrapper on it totally makes sense and it’s opened my eyes. I think that’s probably the biggest. Beyond that, I would say the Dell hardware being able to run frontier models, seeing Gemini running local for sovereignty conversations – I think that’s a really good thing to see as well. Robert Dutt: Along those lines, obviously you touched on one of the big stories this week, which is deskside AI – the idea of physical infrastructure that’s at or near the customer’s desk, either in the data center or right there in the PC, that’s processing the models locally. It sounds like something that you’re interested in. I’m curious where it lands for your customers. Is it something that’s a conversation point, or is it ahead of where they are in the AI discussion at this point? Nigel Brown: I would say broad scope – I don’t want to lump all my customers into one bucket – but broad scope, it’s ahead. I don’t think you’re seeing a lot of organizations ready for it. We also deal heavily with public sector enterprise accounts, for example. We’re doing more and more in the commercial market where you’re going to see a little bit more playing and adoption within tech teams. But in ours, yeah, I’d say we’re definitely ahead right now. So it gives you a chance to get in there and pitch the idea as something new and plant those seeds. Once I get it past my IT and security folks, then that’s where it’s all going to start. If I can’t get it through mine in a good conversation, then I’m never going to be able to with our clients. Robert Dutt: But it sounds like there’s at least that – from your comments on OpenClaw, it sounds like there’s that door, that area of interest. Nigel Brown: Seeing it today under the NemoClaw and Viya umbrella – yeah, I think there’s definitely something there. They’ve set it up closed by default. It can’t leave the box. That’s what I saw in the lab today. So until you set up essentially like a firewall rule to allow it to do something, it’s a safety net that I think really opens the conversation and allows the idea of end users actually playing. Those are really early adopters anyway. And how could I integrate agentic AI into organizations? Robert Dutt: Man, how often does it come back down to governance with AI? Nigel Brown: Oh, absolutely. That’s pretty much the name of the game everywhere. And so we’re doing it well, and many are still scrambling. Robert Dutt: You touch on you guys having a lot of public sector, healthcare, education, all those kinds of verticals – not always the fastest to move on new tech. Along the lines of the previous questions, but sort of taken out a notch – how much of what the AI announcements we’ve heard this week translate directly to where your customers are at, versus how much needs to be, shall we say, adapted for the reality of your accounts? Nigel Brown: Well, you go to any of these events and it’s, “We’re behind if we’re not doing agentic AI everywhere.” Reality is, it’s just not true. I think it’s very forward-thinking – or very optimistic – to think we’re all moving that fast. It’s headed in that direction quicker and quicker. Executive tables are always the ones sitting there going, “We want it, we need it in our organizations, we’re going to get left behind.” So it’s very top of mind. But some organizations have very niche deployments – they’re figuring out the right solutions. Healthcare – I’ve seen it, they’ve done some phenomenal things in radiology and other areas. So it’s picking up. We’re dealing with one client right now that’s looking at online pharmacy and they’re looking at a huge Dell compute cluster to run AI on. So you see it, but it’s not commonplace. It’s not every organization. Certainly as you get into municipalities and things like that, it’s Copilot at best – that’s really where they’re trying to play – and their user base just isn’t adopting, not even close. Robert Dutt: So it sounds like there are at least a couple of steps that need to happen to get to the point of, A, using what’s already in place and, B, potentially looking at building out something internally – and the stuff that’s been talked about here a lot, the idea of running those AI workloads internally on the data center side. Nigel Brown: Yeah. I think it’s going to get there for sure. Right now the conversation has to be outcomes – not “I want AI.” And right now it’s so heavily, “Well, I know I need it, I don’t know what for yet.” I’ve seen it even in some peer groups – the dialogue is, “Well, we’re going to do AI, we’re going to build agents.” So, what for? And then there’s a long pause. Driving outcomes conversations is where it’s going to start, in my opinion. The hardware is going to follow it. And that really ties into, well, where are you going to run it? Do you understand token economics – or tokenomics, whatever the buzzword is right now – and that’s a really big deal. For me, getting that message out really loud and clear around the cost of tokens – I’ve done it, I’ve gotten burned. I spent a hundred bucks on a plane because I wanted to see Claude do something cool. And you’re going, wow, if I can do that in 10 minutes, think of what larger organizations will spend if they don’t find a smarter way to run it. Robert Dutt: That’s a good point – it’s not something you necessarily understand, but it’s something you can sure feel if you start to have adventures with the stuff. Nigel Brown: Well, exactly. And all it’s going to take – like I said, a lot of organizations started with Copilot under the Microsoft umbrella, because it was like an easy button. It was there for them, it was already set up. I am worried about some of those days changing, where that subscription turns into usage-based models. And we’ll see where that goes. You’re seeing it with Anthropic, you’re seeing it with Perplexity. I bounce off my limits all the time. Most of what I’m doing I can wait till tomorrow – but it’s easy to get out of control. Robert Dutt: And user computing is pretty core to what you guys do. There are a few things going on there – Windows 11 end-of-life support coming in October, the AI PC push coming from every direction at the same time. I’m curious if those two things are coming together in customer conversations as one refresh decision, or are they still separate tracks – the need to modernize for the Windows upgrade versus the need to modernize to get the most out of AI workloads? Nigel Brown: I think the end-of-support conversation and hardware refresh, honestly, is the biggest driver of the conversation that I’ve seen. And then that leads into, well, do I need an AI PC, and why, and what’s going to run on it? Everybody’s exploring and curious about it. There’s more skepticism about whether you need it now. Robert Dutt: How is that hitting along with the current fun situation with hardware constraints and prices spiking? And we’re hearing pretty directly from Jeff Clarke that, you know, telling customers, let us know what you want as early in the process as you can. I think the natural addendum to that is, make decisions knowing you might have this machine for a little bit longer than you previously expected. Nigel Brown: Totally right. So it’s very much my dialogue with our clients – it’s future-proofing. You better do it now. You don’t want to be stuck with a machine that can’t run an NPU for the next five years. So even if right now there’s skepticism about how much is going to run on it today, I think it is an important conversation to have and make sure that we’re ready for the moments where we’re really seeing workloads and inferencing running on device. You have to have that conversation now and pre-plan for it. But yeah, it’s been – especially in public sector – a hard conversation to have right now. Supply chain – we’re like a broken record. It still surprises me how many clients we talk to that haven’t seen this coming, that don’t know it’s real, or you get the ones going, “Well, I think it’s going to clear up in September, I’ll just wait till then.” Oh man. Brace for it. We’ve got to be ready. It just feels like a conversation on repeat these days – and it’s more than worth it, making sure we’re doing model selection with the future in mind. Robert Dutt: I find it’s a fun time to be a partner in that particular space. Nigel Brown: Well, you know, quote volume has quadrupled, because that same customer deal might take four different passes before they’ve made it through, especially in government. Pricing validity is a real challenge. It’s a moving target – no decision ever gets made fast. Robert Dutt: I want to talk a little about cyber resilience – another big topic here at the event. You guys run a managed backup practice, I understand, and you’re doing a lot of what vendors are asking MSPs to evolve towards. When you get into a customer environment today, what’s the most common gap between what they think their backup situation looks like and the reality of the situation? Nigel Brown: That’s an interesting question. It’s a real mixed bag. I always start with, “How confident are you in your ability to recover?” And most leaders – business leaders, outside of IT – there’s like a long pause. “Well, I don’t know.” Okay. Have you ever tested your recovery capability? No. Well, that’s where we’re going to start. And in other dialogues, they think they’ve got the backups running, but nobody’s been looking at them – they’re coming from doing it themselves, or maybe a mom-and-pop IT person taking care of it. They’re not watching, they’re not looking at tools, they’re not getting alert notifications on whether it’s keeping up and whether they’re protected. So that’s very foundational. Warning new clients – it’s just, let’s take them on that journey, do an assessment of the whole environment, make sure we’re protected. And a lot of conversations are, “Do you know that you’re not protected? Like, if you got ransomware tomorrow, there’s nothing I could do to help you, even though I’m your MSP.” That’s a scary reality. I’ve seen that have to go back to boards and make some tough decisions, find budget and solve it. They usually do – they react fast – but you’ve got to make the risk abundantly clear. Robert Dutt: That makes sense. In talking to Rob Emsley, who’s on the marketing team for the cyber resilience side at Dell, he was saying that 97% of cyber attacks now are specifically targeting backup infrastructure – because it turns out that’s where all the stuff is. Does that match what you’re seeing, and has that shift changed what you’re recommending to customers about what being protected really means for them? Nigel Brown: I wouldn’t say it’s really changed our messaging. I’d like to think we were maybe ahead of the curve in talking about storage and immutability – some of these key elements of, well, you just need it. That’s how we run our hosted service for clients that use it. And if we’re building out an architecture for another client, it’s just fundamental these days. You can’t even consider a solution that doesn’t include immutability protection, being able to spot bad things happening. But I’ve seen it – we’ve come into a disaster client where, “Hey, we got ransomware, can you help us recover?” And you go to the backup server to find out it was ransomwared too. “Do you have any tapes floating around?” It’s a tough chat to have. You see that less these days, but you definitely see the attempts – people trying to do it. And even other MSPs – I hate to say it – they’re not mature enough in how they’re protecting. They took the backup server, joined it to the domain – it’s just another device on the network. And sure enough, that’s exactly what gets hit because they didn’t plan it out. So it’s all planning and doing it right in the first place. Robert Dutt: It’s a checkbox as opposed to something that’s more firmly thought through. Given that, how do you approach it with customers? Do you come at it as, “This is something you should do, these are the reasons why, this is the potential downside” – or is it a thou-shalt kind of conversation? Nigel Brown: You know, a pile of years ago, after seeing an incident hit a new customer, I kind of resolved – I’ve got to be more of a jerk. I hate to say it. I got to be a lot tougher in my stance. I was being too nice. So yeah, in all things on this, my position is to generally take a pretty firm line. It’s all about risk, though. And to business leaders especially, that’s a term they understand. I’m not telling them, “Okay, you need this type of backup solution and it’s going to do these things.” It’s all about, how do we address the risk that you have right now? Leave it to us to figure out the details as we design the solution. Rarely do we get into the weeds of it unless it’s a larger client where we’re dealing with a large IT team that has opinions. But usually in those larger environments, there are groups that are already aligned – they know what they should be doing, maybe just haven’t done it themselves yet. The new architecture is absolutely going to include all those steps. So it’s an easier conversation to have. In some ways, it’s giving them permission if you’re coming in as a new supplier – it’s the stuff they’ve wanted to do, but haven’t really had the air cover to make the case. Robert Dutt: Yeah, you come in as that outside opinion to say, this is how it needs to be. Nigel Brown: And our job is often more of just a translator for those IT teams to their leadership – to help support the business case. Robert Dutt: I want to talk about the Modern Partner Platform and some of the partner program changes that have rolled out this week. One of the big things is obviously the revised incentive structure, with cyber resilience particularly called out as a premium rebate area. From your seat as a Titanium partner, what does the new structure tell you about where Dell sees the biggest growth opportunities for partners? Nigel Brown: Well, I think it does exactly that – it says where the growth opportunities are. And largely there was no surprise. In my opinion, when you look at it, it aligns to how we want to lead deals, it aligns with the conversations we’re already going to have. Now it’s just helping incentivize that dialogue. Nothing surprising there – I just see better alignment. Robert Dutt: Let’s play a little bit of “anything can happen here.” Vendors like Dell are starting to build agentic AI into their programs, their portals, their tools – all the stuff you guys work with every day. Where do you see the most genuine value for an organization like your own in vendors – agentifying, for want of a better word – their partner programs and tools? And the flip side: are there any potholes you’re watching out for as that rolls out? Nigel Brown: You know, the more the merrier – more tools you can bring in is great. We’re always excited to see what they come up with. But to me, the bottom line is back to outcomes. It’s about reducing friction in the sales process. What do we want our sellers to do? We want them out selling. Living in a partner portal trying to find what they need, deal registration, all of those things that can be painful – sometimes it’s just admin work taking you away from conversations with clients. Reduce friction – that’s the name of the game. Do I want to see more AI-generated marketing content? No. We can do that ourselves – one prompt, feed something in, done. To me, the more you can expose what matters to us and reduce friction, the better. It keeps us doing what we should be doing and not sitting there doing admin work. Robert Dutt: It sounds like based on that comment, what Dell and a lot of its peers are doing is already on track – because I’m sure they’re asking these exact same questions of partners around the world right now. Nigel Brown: Oh, they’ve got way smarter people than me working in these massive organizations. They know the outcomes we want to achieve. And I’m excited that we’re at a point in time where we can see some of this come to fruition. Ten years ago, this was never a reality. Robert Dutt: What’s the biggest misconception you think your customers have about what it means to be AI ready right now? Nigel Brown: I think it depends on who the conversation is centered around. If it’s C-suite leadership, it’s back to, “We want AI, I don’t know what for, I don’t know what it is, but I know I need it.” There are tough conversations to be had. AI readiness is really, is your data ready? We heard that on stage this morning. Most organizations we walk into – it turns out they’ve got no data governance. So, let’s define some of this, let’s build some process, look at the right tools. In the Microsoft lens, we do a lot around Microsoft 365 and modern workplace. Well, then it’s a Purview conversation. And they get confused – “Why are you talking about DLP and Purview? I thought we were talking about AI readiness.” That’s exactly what it’s all about. The other big one I think they’re not taking seriously enough is the end-user adoption side. I’ve seen organizations – you go into their portals and have a look with them – their adoption of Copilot, where they’ve spent a whole pile of money, is abysmal. So then the dialogue is, “What you actually need to do is get your users excited. Train them, show them the cool things.” I think we’ve been really successful doing that inside our own organization, and now that’s something we deliver to our clients as well – we need to get your teams ready and thinking differently. At a C-suite level, they’re usually surprised at the path it takes, or in some cases how long it might take to get there. “Your data is in such rough shape – you’re two years away. You need to build a foundation before you can really consume it.” Now, some of the announcements this morning – okay, that starts changing the equation. We could get there faster if we have the right infrastructure in place. Robert Dutt: For a variety of reasons, the Canadian data sovereignty question feels like it’s getting louder. And I have to imagine, especially in your public sector footprint, how are you helping customers think through AI infrastructure decisions when data residency and compliance are an increasing part of the equation? Nigel Brown: It’s a non-negotiable for most of our enterprise and public sector clients. It’s going to run on-prem. They cannot afford to run on cloud. Yes, they want the latest models, the frontier models, the cool bells and whistles as we all do. But really – I presented at a conference last year on exactly this topic, why it’s important to bring it back on-prem. Never mind the tokenomics conversation – now there’s just more ammunition. I chatted with one IT leader, a commercial client, not public sector, who was all proud of how he’d migrated everything to cloud. We were in a session where they talked through the tokenomics challenge and another reason why sovereignty matters. And you watch the look on his face go, “Wow, I’m going to have to start building a data center again. I thought I got out of that.” And he was sitting there with his CEO in the room for that conversation. Kind of a wake-up call. So my dialogue is, let’s talk through what does the Patriot Act mean? What does FISA Section 702 mean? It’s a little bit scary, and people are shocked – “I thought running in Google Cloud or AWS, running it in a Canadian location was good enough.” No. That provider has access to your data. Have you heard of the CLOUD Act? That’s nothing compared to FISA 702 – they don’t even need to ask. They can just go and get it. And that’s pretty scary. So yeah, a lot of our job now is just sharing and communicating the right things to our clients and making sure they’re aware. Robert Dutt: Aside from your efforts to bring that education – do you find that the level of general awareness is on the rise? Are we getting to more of a discussion about how to solve for this, rather than still defining the scope of the problem? Nigel Brown: I would love to say it’s more mature. The reality is no – it’s still early-stage conversations. You get anomalies. We were with some clients who are way ahead and have just deployed Azure Local on Dell infrastructure. They’re doing amazing things, moving fast. So now it’s more, “How can I partner with you to go share this message? Why you went there, why you built it this way, what are you doing about it?” But no, it’s going to be a continued push – much like the supply chain story here – these dialogues just repeat as you walk into client after client. Robert Dutt: Last one for me – along the same lines as the first question, but a slightly different lens. What’s one thing from this week that you think will genuinely change what Microserve brings to customers in the next 12 months? Nigel Brown: I come back to where we started – the whole side of agentic AI. That was not on my radar, not in a serious way. “Let’s play around with this, let’s lab it out, see where it’s getting explored.” When you see a name like Dell behind what we’re doing, that got me more excited than I would have thought. I want to pilot inside our org. And if we can start building something that works here, then absolutely – taking that to clients and saying, “Okay, look at the GB10s, look at the GB300s, let’s move up the ladder.” There’s a tangible path that gives them more value than trying to build massive solutions right out of the gate. There are quick wins there, and that’s what excites me – showing a customer how there could be a quick win if we did this right. And it ties into the last thread we were pulling on – “Okay, you’re telling me I shouldn’t have all this stuff running on public cloud, so where’s it going to run?” And you’re not talking megawatts and massive data centers here. All I want to do is automate tasks and do some of this lower-level stuff. I think that’s going to be an interesting entry point for a lot of clients – making it more accessible. Everybody’s used ChatGPT, Claude, whatever their tool of choice is, so they’re into prompting. Nobody’s really understanding Copilot or understanding agentic – it’s a big buzzword. That’s our job. We can show them a slice of the possible, mock up these use cases, and those are quick wins. Then it is something deployable at scale – you just move it from the little box to a bigger box. The more people take advantage of it and keep moving up the scale, you don’t need to go spend millions upfront to play around with something like that. It’s going to open more doors. Robert Dutt: No shortage of interesting opportunities. Good luck getting out there and chasing those, and thanks again for making the time this week. Nigel Brown: You bet. Thanks for having me.

Amerika en China zijn verwikkeld in een race om suprematie in AI en Europa bungelt weer eens achteraan. De gevolgen daarvan kunnen ingrijpend zijn. Te gast: Marietje Schaake, fellow aan het Institute for Human-Centered Artificial Intelligence aan de Universiteit van Stanford. Boek: De Machtscode Marietje Schaake schetst hoe kunstmatige intelligentie de machtsbalans tussen Verenigde Staten, China en Europa verandert en waarom Europese afhankelijkheid van Amerikaanse AI-bedrijven gevaarlijk wordt. Zij wijst op de politieke verwevenheid tussen Silicon Valley en Washington, de Cloud Act en bedrijven als Palantir, die tegelijk in Oekraïne en Gaza actie zijn. Dat alles maakt Europese digitale soevereiniteit tot een veiligheidsdossier, niet alleen een marktkwestie. Rob de Wijk en Arend Jan Boekestijn benadrukken hoe langzaam Europa beweegt in wetgeving, kapitaalmarkt en defensiesamenwerking, terwijl AI al wordt ingezet voor autonome wapensystemen. De AI Act, een Europese chipsindustrie en een echte kapitaalmarkt moeten tegelijk snelheid en bescherming bieden, maar botsen op politieke terughoudendheid. De vraag wordt hoe lang Europa nog kan vertrouwen op Amerikaanse bescherming terwijl het zelf geen AI-supermacht is. De gesprekspartners bespreken ook de acute risico's van open source intelligence, hackende AI-modellen zoals Anthropic’s Mythos en het strategische gewicht van ASML in de wereldwijde chipoorlog. Schaake ziet een groeiend verzet in de VS tegen datacenters en AI-banenverlies, en tegelijk een kans voor Europa om talent aan te trekken en een eigen democratisch AI-ecosysteem te bouwen. Wie wil begrijpen waarom AI geen gadget maar machtsinstrument is, hoort hier hoe die strijd nu al over Europese hoofden wordt uitgevochten. [Samenvatting geschreven door AI en gecontroleerd door mens] Over de Podcast Arend Jan Boekestijn en Rob de Wijk gaan onder leiding van Hugo Reitsma op zoek naar de nieuwe wereldorde. Wat betekenen oorlog, machtspolitiek en economische verschuivingen voor Europa en Nederland? In elke aflevering duiken zij in de geopolitieke actualiteit. In 2022 werd Boekestijn en De Wijk uitgeroepen tot winnaar in de categorie Nieuws & Politiek tijdens de Dutch Podcast Awards Reageren? Op X: @ajboekestijn en @robdewijk Bluesky: @hugoreitsma.bsky.social Mail: boekestijnendewijk@bnr.nl Over de makers: Arend Jan Boekestijn is een Nederlands historicus en voormalig politicus. Hij studeerde geschiedenis en politieke wetenschappen aan de Vrije Universiteit in Amsterdam. Boekestijn is voormalig Tweede Kamerlid (tot 2009). Sinds 1989 is hij verbonden aan de vakgroep geschiedenis van de Universiteit Utrecht en sinds 2016 lid van commissie Vrede en Veiligheid van AIV. Rob de Wijk studeerde eigentijdse geschiedenis en internationale betrekkingen, promoveerde op kernwapenstrategieën, werd hoogleraar in Leiden en richtte in 2007 het Den Haag Centrum voor Strategische Studies op. Hugo Reitsma studeerde rechten en politicologie. Hij werkte eerder als politiek verslaggever en vanuit verschillende conflictgebieden. Hij is auteur van het boek ‘Boekestijn en De Wijk voorspellen de toekomst’ (november 2023).See omnystudio.com/listener for privacy information.

Tom Bartlett spent 22 years inside the NHS. For three and a half years, he led the 150-person engineering team at NHS England that built the £330M Federated Data Platform with Palantir.He a month ago and he's the only insider speaking publicly about what the platform actually does, what it costs, and whether the NHS has any credible alternative.We cover the architecture, the cost, the CLOUD Act, the BMJ exposé, the conflicts of interest at Chelsea & Westminster, and whether ministers should trigger the break clause.

Rüdiger hat Tobis Geburtstag nicht vergessen – aber DHL schon fast. Mit vertauschten Absender- und Empfängeradressen startet Folge 121 entspannt chaotisch, bevor die beiden tief in Technik, KI und Google-Bewertungs-Intrigen eintauchen. Leserpost RailWise (iOS): Zwei Masterstudenten tracken minütlich den gesamten Fernverkehr Deutschlands – 25.000 Fahrten/Tag, Millionen Datenpunkte/Woche. Hörer Markus korrigiert Tobis Wire-Einschätzung: Der CLOUD Act greift auch bei europäischen Servern amerikanischer Betreiber. Hörer Klaus schickt zwei Links zu Signal-Angriff und Wire-CEO-Interview – kommen in die Shownotes. Kurzmeldungen WhatsApp bekommt eine eigene Backup-Cloud – endlich Plattformwechsel zwischen iOS und Android möglich. Apple ermöglicht Jahresabo-Preise bei monatlicher Zahlung im App Store. E-Ink-Fingernägel von Eye Polish (CES 2026): Farbe per App ändern – die App hat Standortdaten geleakt. KI & Tech Spielearchiv Myriad (390 TB, 6.000 $/Monat) stirbt an explodierenden Hosting-Kosten durch KI-Rechenzentren – Community spiegelt, aber Langfristigkeit fraglich. BambuLab droht OrcaSlicer-Entwickler juristisch, weil er gesperrte Druckerfunktionen per Fork wieder freigeschaltet hat. Microsoft testet 19 KI-Modelle bei Dokumentenarbeit: nach 7 Schritten 50 % der Dateien korrupt. Chrome-Extension aus Kanada zeichnet jeden Tastendruck in Google Docs auf – Lehrer können Hausaufgaben als Zeitraffer abspielen, alle fünf Investoren in der Höhle der Löwen dabei. Apple verschippt versehentlich eine interne CLAUDE.md-Datei in einem App-Update. Ex-GitHub-CEO Ned Friedman gibt seiner KI Zugriff auf Kameras, Bluttest und Tesla – sie schickt ihn Wasser trinken und ändert die Route zur Apotheke. Samsung The Frame nimmt Bilder ohne Authentifizierung von jedem Gerät im gleichen WLAN entgegen – Rüdiger hat's per Claude Code genutzt, Rick Astley im Media Markt wäre der nächste logische Schritt. Google-Bewertungen Google zeigt jetzt in Maps an, wenn ein Lokal Bewertungen löschen ließ – in 50er-Blöcken. Transparenz oder Freifahrtschein für Anwaltskanzleien, die davon leben? Plus: Spare Ribs, ein besoffener Russe im Bayerischen Hof und warum Tobi trotzdem nie eine Rezension schreibt. -- Links zur Folge immer auf https://podcast.ichglaubeeshackt.de/ Wenn Euch unser Podcast gefallen hat, freuen wir uns über eine Bewertung! Feedback wie z.B. Themenwünsche könnt Ihr uns über sämtliche Kanäle zukommen lassen: Email: podcast@ichglaubeeshackt.de Web: podcast.ichglaubeeshackt.de Instagram: http://instagram.com/igehpodcast

In this episode of the Crazy Wisdom Podcast, host Stewart Alsop sits down with Cindy Cohn, Executive Director of the Electronic Frontier Foundation (EFF), for a wide-ranging conversation covering the EFF's origins and mission, the countercultural roots of Silicon Valley, the rise of surveillance-based business models, the challenges facing open source software and open-weight AI models, the legal landscape around intellectual property and privacy law, and the growing tension between government overreach and civil liberties in the digital age. Cindy also discusses her upcoming departure from EFF after 26 years, the transition to new leadership, and her recently published book Privacy's Defender, which chronicles key legal battles she fought to protect digital privacy rights.Links mentioned:- EFF website: eff.org- Privacy's Defender book: eff.org/privacysdefenderTimestamps00:00 - Stewart introduces Cindy Cohn, EFF Executive Director, who explains the organization's mission protecting digital rights since 1990.05:00 - Cindy connects counterculture roots to early internet idealism, describing how digital communication broke down physical barriers for organizing.10:00 - Cindy reveals surveillance becoming the dominant business model surprised her, blaming corporate consolidation over naive techno-optimism.15:00 - Discussion shifts to Silicon Valley's military contractor substrate and how corporate money co-opted hacker ethos.20:00 - Open source community faces existential threat from age verification legislation while open-weight AI models emerge as critical alternative.25:00 - Cindy outlines legal frameworks like compulsory licensing and easements that could democratize access to foundational AI models.30:00 - Privacy principles around secondary data use identified as core surveillance problem, with Anthropic's domestic surveillance red line praised.35:00 - Cloud Act, Five Eyes surveillance networks, and global jurisdictional complexity examined through individual threat modeling lens.40:00 - Constitutional rights and democratic participation framed as irreplaceable bulwarks against authoritarian surveillance tendencies.45:00 - Cindy announces departure from EFF after 26 years, naming successor Nicole Ozer while planning return to courtroom litigation.Key Insights1. The Electronic Frontier Foundation was founded in 1990, before the World Wide Web existed, by Mitch Kapoor, John Perry Barlow, and John Gilmore, with early support from Steve Wozniak. Its core mission is to ensure that civil rights and freedoms follow people into the digital world, using lawyers, technologists, and activists to keep the internet on the side of users.2. The early countercultural movement of the 1960s and 70s heavily influenced the founders of the internet and EFF. Figures like Barlow believed the digital world could reduce physical barriers like race, class, and geography, allowing people to be judged by the quality of their ideas rather than the circumstances of their birth.3. The dominant surveillance business model that emerged was not inevitable. Cohn argues it resulted from deliberate policy failures, particularly the abandonment of competition law, which allowed a handful of companies to consolidate control over the entire internet and adopt 360-degree data collection as their primary revenue strategy.4. Open source communities remain active and vital but are under serious threat from legislation like age verification laws that make it practically impossible to maintain fully open tools. Cohn sees this community as essential to reclaiming public control over computation, especially in the age of AI.5. The open weights question for AI models is fundamentally different from traditional open source software because of the enormous capital required to train foundation models. Cohn suggests legal mechanisms like compulsory licensing, similar to how cover songs work in copyright law, as one possible path toward broader public access.6. A core privacy principle Cohn advocates is that data collected for one purpose must not be used for others. This single rule, if enforced, would begin dismantling the infrastructure that enables mass individual surveillance, including the AI-powered profiling she sees as the next dangerous frontier.7. Cohn is stepping down from EFF after 26 years to allow new leadership and return to litigation, which is where she believes her impact is greatest. She also wrote a book called Privacy's Defender to preserve the history of digital rights fights from the 1990s onward and to help people understand how current threats emerged so they can work to reverse them.

Tim Cook quitte la direction d'Apple et passe le relais à John Ternus • OpenAI dégaine un nouveau modèle d'images surpuissant • Une boutique 100 % gérée par une IA ouvre à San Francisco • La Chine impressionne avec un semi-marathon de robots humanoïdes • Google injecte de l'IA dans sa suite bureautique.⭐️ [Annonce] : découvrez Frogans : l'innovation française qui réinvente le Web===============Sommaire détaillé : ===============Apple : Tim Cook passe la main à John Ternus (03:11)C'est une page majeure qui se tourne chez Apple. Après quinze ans à la tête du groupe, Tim Cook quitte son poste de CEO et devient président exécutif, laissant les rênes à John Ternus, actuel responsable produits. Une transition en douceur pour l'entreprise valorisée près de 4 000 milliards de dollars, qui devra désormais relever les défis de l'intelligence artificielle et préparer l'après-iPhone. Au-delà du symbole, l'héritage de Tim Cook est considérable : montée en puissance des services, succès de l'Apple Watch et surtout virage stratégique vers les puces maison Apple Silicon, qui placent aujourd'hui la firme en position favorable dans la course à l'IA. Reste à savoir comment John Ternus imprimera sa marque.OpenAI muscle son jeu avec GPT-5.5 et ChatGPT Image 2.0 (05:44)OpenAI frappe fort avec GPT-5.5, nouvelle version de son grand modèle de langage, plus rapide, plus performant mais aussi plus cher, destiné aux abonnés payants et aux entreprises via API. L'objectif est clair : reprendre l'avantage face à Google et Anthropic dans une compétition devenue féroce. En parallèle, le nouveau modèle de génération d'images ChatGPT Image 2.0 impressionne par son réalisme et sa capacité à produire du texte fiable dans de nombreuses langues. Intégré à Codex pour les développeurs, il ouvre des perspectives créatives considérables… tout en soulevant des risques accrus d'usages frauduleux.DeepSeek V4 et le réveil chinois de l'IA (08:06)La Chine n'est pas en reste avec la sortie de DeepSeek V4, modèle open source décliné en version Pro et Flash, aux capacités agentiques renforcées. Depuis son irruption en 2025, DeepSeek bouscule le marché en affichant des performances comparables aux leaders américains avec des ressources optimisées. Selon le baromètre annuel de l'université Stanford, la Chine talonne désormais les États-Unis tandis que la France ne place qu'un seul modèle dans le haut du classement, signé Mistral AI. Les écarts d'investissement restent abyssaux, illustrant un décrochage européen préoccupant.Andon Market : la boutique créée par une IA (10:25)À San Francisco, sur Union Street, une petite échoppe baptisée Endowment Market intrigue : concept, bail commercial, commandes fournisseurs, site web… tout a été orchestré par une IA nommée Luna, dotée d'un budget initial de 100 000 dollars. Derrière l'expérience, la start-up Andon Labs teste une idée radicale : une intelligence artificielle peut-elle créer et gérer un commerce rentable dans le monde réel ? Si des humains assurent la vente en magasin, l'initiative pose une question vertigineuse sur l'autonomie économique des machines.Meta surveille ses employés pour entraîner ses IA (12:36)Chez Meta, un programme baptisé “Model Capability Initiative” installe des outils de suivi sur les postes de travail afin de collecter des données comportementales destinées à l'entraînement des modèles d'IA. Officiellement conçue pour améliorer les performances des systèmes, la démarche suscite des inquiétudes en interne, sur fond de licenciements. Jusqu'où peut-on aller dans la captation des données des salariés au nom de l'innovation ?Cyberattaques : l'ANTS piratée, un hacker arrêté (14:02)Nouvelle alerte en France avec le piratage de l'Agence nationale des titres sécurisés (ANTS). Un hacker affirme détenir 19 millions d'enregistrements comprenant des données d'état civil, désormais proposées à la vente sur le darknet. Une enquête est ouverte et la CNIL a été saisie. Dans le même temps, un pirate présumé de 21 ans, connu sous le pseudonyme Hexdec, a été interpellé en Vendée. Soupçonné d'être lié à plusieurs attaques majeures, il avait récemment revendiqué ses actes dans une interview, assumant agir uniquement pour l'argent.Health Data Hub : cap sur un hébergeur français (16:41)Le Health Data Hub ne sera finalement pas hébergé par Microsoft. Après polémique autour des risques liés au Cloud Act américain, le gouvernement confie l'infrastructure à Scaleway, filiale du groupe Iliad. La migration prévue fin 2026 marque un tournant stratégique vers une souveraineté numérique renforcée pour cette plateforme destinée à soutenir la recherche en santé grâce à l'IA.Google et l'entreprise agentique (43:04)À Las Vegas, lors de Cloud Next 2026, Google a présenté sa vision de “l'entreprise agentique”. Objectif : déployer des agents IA capables d'automatiser tâches répétitives, réponses à appels d'offres, analyses de données ou gestion RH. Dans Google Workspace, l'IA Gemini devient transversale et proactive, capable de synthétiser agenda, mails et documents pour assister l'utilisateur. Pour les entreprises, une marketplace d'agents et des outils de gouvernance promettent d'encadrer cette nouvelle génération d'assistants intelligents

The company you hired to protect you just got hacked. That is not a hypothetical, it is the defining threat pattern of the past 18 months. In this episode Darnley breaks down why cybersecurity vendors, including some of the most recognized names in the space, have become the highest-value targets for threat actors, how a single vendor compromise translates directly into a supply chain breach affecting hundreds or thousands of downstream clients, and what every business needs to do before signing another security contract. Featuring real-world case vendors including SolarWinds, Okta, CrowdStrike, Sisense, and the 2026 eScan compromise, plus a practical vendor vetting playbook and a hard look at why infrastructure-level privacy matters more.Listen hereClick here to send future episode recommendationSupport the showSubscribe now to Darnley's Cyber Cafe and stay informed on the latest developments in the ever-evolving digital landscape.

Em 23 de julho de 2025, Donald Trump assinou ordem para exportar “pacotes completos” de inteligência artificial, colocando o Brasil entre destinos prioritários ao lado de Egito e Indonésia. A medida intensifica a disputa com a China por influência tecnológica global. No mesmo período, o Brasil firmou memorando com Pequim e negocia com Washington, enquanto amplia dependência de infraestrutura digital estrangeira. Thiago Aragão, analista de geopolítica O Brasil está nominalmente na lista de destinos prioritários. Ao lado do Egito e da Indonésia, o país figura entre os mercados emergentes onde a presença americana precisa ser consolidada, antes que a influência chinesa se torne irreversível. Para entender o que isso representa na prática, vale olhar o que aconteceu com o Japão. Em outubro de 2025, durante a visita de Trump a Tóquio, os dois países assinaram um “Technology Prosperity Deal”, um acordo de alinhamento em política de IA que vai muito além da compra e venda de hardware. O documento inclui compromissos sobre padrões técnicos, frameworks de governança, fluxo de dados e cooperação em segurança digital. O Japão passou a integrar estruturalmente a órbita tecnológica americana, não apenas como parceiro comercial, mas como parceiro normativo. Washington quer replicar esse modelo em escala. Leia tambémPossível uso de inteligência artificial em ataques no Oriente Médio levanta questões, diz especialista A lógica americana é clara e, num certo sentido, legítima. A China exporta tecnologia de IA num modelo que analistas descrevem como “full-stack com condições embutidas”: hardware subsidiado, software com lógica de caixa preta e frameworks de governança que replicam o modelo regulatório de Pequim. Washington entendeu que competir chip a chip não é suficiente. É preciso exportar o ecossistema inteiro e, com ele, a arquitetura normativa que o acompanha. Para o Brasil, o problema é que os dois modelos chegam com política externa no rodapé do contrato. Leia tambémNa Índia, Lula diz que IA sem regulação ameaça a democracia e aprofunda desigualdades Em 2025, o governo Lula assinou um memorando de entendimento com a China para aprofundar a colaboração em inteligência artificial. No mesmo período, Brasília avançava nas conversas com Washington sobre o programa de exportação de IA e recebia anúncios de bilhões de dólares em data centers da Microsoft, Amazon e Oracle. Do ponto de vista diplomático, é um malabarismo admirável. Do ponto de vista tecnológico, é uma contradição estrutural que vai cobrar seu preço mais cedo do que se imagina. A questão central não é quem vende o chip. É quem treinou o modelo. Os grandes sistemas de linguagem e tomada de decisão que o setor público e privado brasileiro já usa, na análise de crédito, na triagem de políticas, na recomendação de conteúdo e na gestão de contratos foram desenvolvidos majoritariamente por empresas americanas, segundo padrões americanos, com dados que refletem realidades americanas. O viés não é necessariamente malicioso. Mas é estrutural. E tende a se aprofundar na medida em que o Brasil sustenta sua infraestrutura cognitiva sobre servidores sujeitos ao CLOUD Act americano, a lei que autoriza o governo federal dos EUA a requisitar dados armazenados por provedores americanos em qualquer jurisdição do mundo, independentemente de onde o servidor esteja fisicamente localizado. Leia tambémInteligência artificial: China alerta para cenário apocalíptico ao estilo 'O Exterminador do Futuro' O próprio debate regulatório revela a ambiguidade. O Senado brasileiro tem acompanhado de perto o AI Act europeu como referência normativa para sua legislação nacional, e o projeto em discussão cria um sistema de governança de IA sob a responsabilidade da Autoridade Nacional de Proteção de Dados. Na teoria, é soberania. Na prática, o discurso regulatório aponta para autonomia, mas a implementação é operada por corporações multinacionais americanas. O Brasil faz a lei, mas quem comanda a infraestrutura sobre a qual essa lei incide são outros. Isso não é uma acusação. É uma descrição de como o poder funciona no século 21. Brasil é maior mercado de dados da América Latina A boa notícia é que o Brasil tem cartas genuínas nessa mesa. É o maior mercado de dados da América Latina, tem uma das matrizes energéticas mais limpas do mundo, o que importa imensamente para data centers, e produziu o Pix, um dos sistemas de pagamentos digitais mais sofisticados em operação no planeta. Quando Washington e Pequim disputam o Brasil como parceiro de IA, não o fazem por generosidade. É porque o país tem o que ambos precisam: escala, energia e população conectada. A dependência, se vier, será escolhida, não imposta. A pergunta que o debate público brasileiro ainda não fez com a seriedade necessária é esta: ao aceitar o pacote completo de IA americano, com seus chips, seus modelos, seus padrões de governança e suas obrigações de compliance, o que o Brasil está abrindo mão em troca? Não em termos comerciais, mas em termos de autonomia sobre decisões que, daqui a dez anos, serão tomadas por sistemas que alguém, em algum lugar, já programou.

Synopsis Épisode spécial de La French Connection avec Jean-François Courteau, président et fondateur de 4DS Technologie, venu présenter son rapport sur la souveraineté des données au Québec. Patrick, Francis et Steve décortiquent avec lui les résultats d'une étude qui a fait jaser jusqu'à Radio-Canada : analyse géographique de l'hébergement Web et courriel des ministères, des 40 plus grandes villes, des 30 plus grandes entreprises et des 40 plus grandes PME du Québec. Le constat est sans équivoque : le Québec n'est pas maître de ses données. Plus de 90 % des organisations gouvernementales ont leurs courriels chez Microsoft, plus de 82 % de l'ensemble des organisations analysées ont leurs courriels filtrés ou hébergés par des entreprises américaines. Le trio revient sur le CLOUD Act américain, cette loi adoptée en 2018 qui donne aux agences fédérales le droit d'exiger les données de tout client d'une entreprise américaine, sans mandat canadien, et avec une clause bâillon qui empêche le fournisseur d'aviser son client. Jean-François explique pourquoi l'emplacement physique des serveurs ne veut rien dire si la compagnie qui les opère est soumise à une juridiction étrangère - et pourquoi la nationalité du fournisseur pèse trois fois plus lourd dans son système de pointage. Francis apporte une perspective critique : le rapport est un bon point de départ mais ne couvre pas toutes les dépendances cachées - Cloudflare, AWS, GitHub, les tenants Microsoft des clients et fournisseurs. La panne Cloudflare du 18 novembre 2025 a démontré à quel point l'économie numérique repose sur une poignée de joueurs. Patrick ramène le palmarès gênant des partis politiques québécois et canadiens : même le PQ et Québec Solidaire, ardents défenseurs de l'achat local, hébergent leurs services chez Cloudflare, Google et Microsoft. L'épisode se termine sur des pistes concrètes : l'annonce du ministre Gilles Bélanger de février 2026 qui marque un virage à 180 degrés vers le logiciel libre, l'exemple du Schleswig-Holstein en Allemagne qui économise 15 M€ par an en sortant de Microsoft 365, et l'idée d'un package “PME souveraine” basé sur Nextcloud, Postfix, Dovecot et LibreOffice pour donner aux entreprises une alternative clé en main aux géants américains. Invité Spécial Jean-François Courteau Crew Patrick Mathieu Francis Coats Steve Waterhouse Liens et ressources Rapport sur la souveraineté des données au Québec (2025) PowerOn Energy - Solutions de souveraineté numérique Le CyberPanier - Marché numérique souverain AugureAI - IA canadienne (tarification) 8x8 Video Conferencing Nextcloud Talk High Performance Backend avec Docker (Arno Welzel) Outil “Vérifier où est hébergé votre site Web” 4DS Technologie Décret 38-2019 (PCCTI) Annonce du ministre Bélanger - Politique de souveraineté numérique (février 2026) CLOUD Act - Quelle souveraineté sur les données numériques? Panne Cloudflare du 18 novembre 2025 Schleswig-Holstein sort de Microsoft 365 - 15 M€ d'économies Danemark abandonne Microsoft Office pour LibreOffice Lyon abandonne Microsoft Office 4,1 M$ par semaine envoyés à Microsoft par le Québec Nextcloud - alternative libre à Microsoft 365 Micrologic - “Projet Cirrus” LibreOffice Shamelessplug Join Discord securite.fm Hackfest iHack POLAR conference La French Connection sur YouTube Crédits Montage audio par Hackfest Communication Locaux virtuels par Streamyard

Rob Falzon, Head of Engineering in the Office of the CTO at Check Point Software Technologies Canada’s data sovereignty landscape is shifting faster than most organizations realize – and according to Rob Falzon, Head of Engineering in the Office of the CTO at Check Point Software Technologies, the conversation isn’t happening early enough. In this episode, Falzon breaks down the regulatory pressure building around Canadian data – including Quebec’s Law 25, Bill C-8, and new federal PIPEDA reform expected this spring that is expected to include data sovereignty provisions. He draws a sharp distinction between data residency (where data sits at rest) and data sovereignty (control over the entire processing chain) that many partners and their customers are still conflating – and explains why contracts alone can’t solve the problem. Falzon unpacks the CLOUD Act dimension: if data lives in the U.S., it is accessible to the U.S. government regardless of where your company is headquartered or what your service agreement says. For MSPs, the conversation turns to opportunity. Recent research from Kiteworks found that 23% of Canadian organizations experienced a data sovereignty incident last year, and mid-market firms lag enterprise by 15 to 25 percentage points in sovereignty maturity – despite facing the same penalties. Falzon’s advice: lead with risk, not product. He also raises a recent U.S. legal judgment holding that all data entered into ChatGPT belongs to OpenAI – and asks whether organizations using AI services even know where that data is going. Check Point launched a dedicated Canadian data region for CloudGuard WAF in March, opening doors to government and regulated-sector contracts that were previously unavailable to partners. But Falzon’s bigger point is this: the regulatory picture is still coming into focus, and MSPs who get educated now – before the legislation fully lands – have a real chance to stake out expertise and become the trusted voice in the room when urgency hits. Read Full Transcript Robert Dutt: Hello and welcome to In The Channel from ChannelBuzz.ca, bringing news and information to the Canadian IT channel community for the last 16 years. I’m Robert Dutt, editor of ChannelBuzz.ca and your host for the show. There’s a phrase you’re probably hearing more and more in customer conversations: data sovereignty. And if you’re not hearing it yet, you probably will soon. Canada’s regulatory landscape around data is shifting fast. Quebec’s Law 25 is already in force with real financial penalties. Bill C-8, the Critical Cyber Systems Protection Act, is working its way through committee. New federal privacy reform is expected this spring, and underneath all of that, there’s a growing realization that the old assumption—that if it’s okay for the U.S., it’s okay for us—may not hold up much longer. My guest today is Rob Falzon, Head of Engineering in the Office of the CTO at Check Point Software Technologies. Rob has spent over 30 years in large-scale security architecture, including government work, and he’s been with Check Point for over two decades. He’s based here in Canada and has a front-row seat to how this market handles security and compliance differently from the rest of the world. We’re going to talk about what’s driving the urgency around data sovereignty in Canada right now, the distinction between data residency and data sovereignty that a lot of partners are still conflating, and what it all means practically for MSPs serving the Canadian mid-market. Let’s get right into it—my chat with Rob Falzon. Rob, thanks for taking the time. I appreciate it. Robert Falzon: No trouble. Robert Dutt: You’ve been in the industry a long time, with Check Point for two decades, and you’ve had a front-row seat to how the Canadian market specifically handles security and compliance. For an audience of Canadian VARs and MSPs, how has the data conversation in Canada changed over, say, the last 18 months or so? It feels like something’s shifted in that discussion. Robert Falzon: Yeah, there’s been a significant shift. In the past, obviously, we’ve seen the changes that have happened with our neighbors to the south and how the climate and atmosphere have changed. It’s caused folks in Canada to have a closer look at what their various different arrangements are from a trust perspective, and what their comfort level might be in where they store their data and how they manage that data—and where their customers are based as well. I think that’s been the primary change in the last few months specifically. For a long time, we’ve had this feeling that Canada and the U.S. have been sort of the same. There wasn’t really a big concern because we have agreements back and forth. A lot of the recent changes have forced us to really revisit those arrangements and see: are we actually making sure that the information is safe and protected? As a result of that, we’ve been getting those questions at Check Point, and it’s incumbent upon us to manage it in such a way that our customers get the security and safety they need while meeting their business requirements. Robert Dutt: From the regulatory side of things, there’s a lot going on. We have Quebec’s Law 25 in place with real penalties behind it. We have Bill C-8 working its way through committee. There’s going to be PIPEDA reform coming up sometime fairly soon, which is rumored to include data sovereignty provisions. Back in November, the government introduced the Digital Sovereignty Framework. For a Canadian MSP who hasn’t been tracking all of this closely, what’s the picture they need to have in their head right now of the regulatory scene? Robert Falzon: Well, like you pointed out, there’s no comprehensive federal law just yet. As you mentioned, there are a number of things on the table and we have some direct focus now from the federal government. There’s a minister assigned specifically for AI that’s taking a very close look at how Canada is managing that. We also have this provincial patchwork. Ontario probably has the most established AI-specific roles so far. Alberta’s Privacy Commissioner also has a report they released last year talking about Alberta creating its own AI law and updating its privacy legislation. All of these changes are happening fairly quickly right now, and it’s incumbent upon MSPs to make sure they’re aware of what these changes are and where they are operating their businesses. There are two aspects to this. The first is the business side: if you have customers that want to consume your services, you need to make sure your services are consumable by them—that you are meeting their data regulation requirements and that the residency and sovereignty requirements these new pieces of legislation introduce are met by whatever services you’re providing. The challenge is that there’s not a lot of clarity right now around what these actual services are. Maybe AI is touching it, or some security component is touching it, but maybe it’s a different type of service related to marketing. This is going to be a challenge for MSPs to make sure they understand their compliance obligations and to closely look at their service offerings. They need to start to decouple what we used to think was an accepted understanding—that if it was okay for the U.S., it was okay for us. It’s not going to be the same anymore. Robert Dutt: There’s another piece of legislation, not necessarily on our side, but the CLOUD Act hanging over all this. Can you walk us through how the CLOUD Act changes the calculus for Canadian organizations using a U.S.-headquartered cloud or security provider? Robert Falzon: There are a few things here to unpack. First of all, it’s not finalized; there are still a lot of negotiations underway. This started back in 2021 or 2022, and obviously, when that started, we were in a completely different geopolitical context than we have today. That’s transformed things into a more complex policy debate and even, to some degree, a national security debate. For us, we’re going to have to start looking very carefully about what regulations we put in place at the federal level that impact us from a legal compliance perspective. Is your CISO well aware of what your obligations are under this? I think if I look at what’s going to change, we’re still going to have to start hosting much of the information we work with in Canada. Anything related to security rule sets, business transaction information—all of this is going to have to be stored in Canada. If you are still leveraging contracts that you might have in the U.S., you’re going to have to look at how you separate out those specific types of data that are protected by law and have them processed and stored in Canada. You may not be able to get out of some of these hosting contracts in the U.S., but the fact is, if that data is in the U.S., it’s going to be available to the U.S. government. If that availability contravenes any legislation we have here, it’s something you’re going to be liable for. Robert Dutt: A lot of times, maybe at the customer level and the partner level, there’s some conflation between data residency and data sovereignty. Can you break that apart? I think when a lot of people hear, “We have a Canadian data center,” they assume the compliance checkbox is checked. Robert Falzon: Yeah. The difference fundamentally is essentially data at rest versus data in motion. If you are storing databases or static information about customers, that data must be resident in Canada. Data sovereignty is essentially the entire chain. Any processing has to be done in Canada, storage has to be done in Canada—the data cannot leave the country or its control sphere the entire time it’s in your possession. I think that’s a critical differentiation because they are often, as you say, conflated to be the same thing. Robert Dutt: What does a sovereignty-defensible architecture actually look like? What are the non-negotiables to make sure you’re covered off there, especially as a service provider? Robert Falzon: You have to look at all of your vendors. You have to make sure that not only are you managing your data effectively yourself, but that all of the vendors you interact with are also following the same guidelines. The challenge here is that we are so integrated with U.S. providers—cloud providers, data center providers. All of those things need to come together, and we need to be aware at all times where this information is stored. Our understanding of where that data is has to improve, so we need better tools to manage that visibility. But we also need to start making actual changes in our infrastructure to make sure it physically resides in Canada. And then we need to look at the rule sets you’re using to manage that data. Do you have the proper security context to store and manipulate that information strictly in Canada as per data sovereignty regulations? Robert Dutt: Let’s bring this to the partner level. There’s a recent survey from Kiteworks that shows 23% of Canadian organizations experienced a data sovereignty incident last year. Mid-market firms lag enterprise by 15 to 25 percentage points in maturity, but they face the same penalties. For an MSP serving that mid-market space, where’s the actual opportunity in terms of educating and compliance? Robert Falzon: Well, if MSPs are at the stage where they’re concerned and trying to get information, imagine where many of our customers are standing. Customers are trusting their partners to provide them with guidance and leadership. If we think about verticals like healthcare, financial services, or the public sector—these are not organizations that typically have heavy internal services or the skill sets to make these decisions about where their cloud data is processed. They’re relying on partners for that. If there are issues, the buck stops with the customer themselves. By helping to educate their customers—making them aware of coming changes, understanding the differences between sovereignty and residency, and looking at their other vendors—partners can take a leadership position. There’s a bit of a vacuum right now in speaking with both partners and customers, where everybody’s just going, “I wonder what’s going to happen next? Am I even ready for this?” It’s a great opportunity to improve their business. Robert Dutt: Is the first question to that customer the general, “Do you know where your data is living and who has access to it?” Or what’s the first concrete question an MSP can take to their customers? Robert Falzon: Well, there are a whole lot of things. First, partners are going to have a better understanding of their customer profile. If they have customers with significant multi-cloud complexity or exposure to the CLOUD Act, they’ll want to start by talking to them about their immediate risk. The challenge we often have is that we want to go in and talk about how a product or service is going to make a difference. Ultimately, what we really need to do is share the conversation about risk. The risk conversation is often overlooked in favor of saying, “I’d like this customer to buy some more Check Point.” But at the end of the day, all of that comes back to their understanding of what the risk is. I would start with risk: talk about what’s in the CLOUD Act, talk about complexity, and talk to them about AI data exfiltration and how that impacts leakage from a legal perspective. Stay away from conversations about specific products and focus on the business outcomes for the customers. That’s what’s going to get you the traction. Robert Dutt: Check Point launched a dedicated Canadian region for CloudGuard Web Application Firewall in March at the Victoria Privacy Summit. What’s driving security vendors specifically to put in infrastructure in Canada right now? Robert Falzon: This is an interesting question because it’s really not a “right now” thing. This is something we’ve been actively looking at for some time. It’s not as easy as just saying, “I’m going to do this in Canada only.” There’s a lot of backend stuff that has to happen. Five years ago, the technology and infrastructure available were somewhat limited. You have to be able to trust the infrastructure you’re placed in. It’s taken years to get here, and we’re quite confident in our ability to deliver the exact same level of quality as we did when it was solely based in the U.S. Countries around the world are starting to take a close look at their most important assets—data and intellectual property—and seeing how easily technology is being used to gain access to private information. Companies would be well-served to understand that this has been a long cycle; it’s not something that just happened overnight. Robert Dutt: For a partner who’s already selling Check Point solutions, what practically changes for them now that this Canadian data region is in place? What deals or conversations does it unlock? Robert Falzon: Certainly anywhere where privacy is paramount, it’s going to have a huge impact because you can start the conversation with the understanding that anything we’re talking about today is going to be data resident and data sovereign to your Canadian customers. That immediately sets you apart from many other vendors who cannot make that claim. If you can address the concern of privacy legislation right out of the gate, then you can focus on the actual business outcomes. It’s going to open doors with agencies very sensitive to this—government entities at the municipal, federal, and provincial levels that might have been off-the-table to a partner that didn’t have solutions meeting those criteria. Robert Dutt: For the MSP who’s a little earlier in the process, what’s the first practical step internally to make sure you’re building this out as an opportunity? Robert Falzon: You have to be extremely well-educated in the legal aspects because you’re going to want to make sure you have a compliance story and accountability you can speak to with your customers. But looking at all the uncertainty relating to AI and machine learning, being able to tie data residency and sovereignty into how that impacts their ability to utilize these new technologies would be a real door opener. There’s a tremendous amount of misunderstanding and lack of information available to customers currently running these solutions. If I were a partner today, I’d be looking at how I have the conversation about security, privacy, and data sovereignty in terms of their ability to be more competitive in the future by leveraging these advanced technologies in a secure way. Robert Dutt: What’s the risk of doing nothing? If I’m a partner and I decide to just keep selling the same way and assume data sovereignty is someone else’s problem, what does that look like 12 months from now? Robert Falzon: Hopefully your customers are already taking a zero-trust approach, so it might be easy to say, “I’ll wait until this settles a bit.” It’s not crazy to think that could still be effective. But if one waits too long and it becomes legislation, now you’re playing catch-up. You won’t be perceived as a leader in the space, and as we know, it’s much harder to win business away from someone else than it is to keep business you already have. Robert Dutt: Last question: what’s the thing about data sovereignty in Canada right now that you think isn’t getting enough attention? Robert Falzon: I think honestly, the conversation about data sovereignty and residency itself is not mentioned enough. It seems to be addressed after the fact. I’m starting to see it come to the forefront, but I still don’t have conversations on a daily basis about this. Even though this announcement was made, I’m still not getting a lot of phone calls about what this means for me, and I would have expected to get a lot more. If we look forward five years, we’ll look back at this and go, “Wow, I can’t believe we only just got that then.” Things are moving so rapidly. If we look at the adoption of AI internal to large corporations—I’ll ask them if they are using AI services, where those services are based, and what the legal ramifications are. Nobody is talking about where the data from ChatGPT lives. There was a legal judgment in the U.S. a couple of weeks back where it was agreed that all data entered into ChatGPT belongs to them—it belongs to OpenAI. Imagine if that’s your company’s data, and you don’t even know it’s leaving because the services you’ve invested in are hosting data all over the world and not in Canada. That’s a risk that’s really not being discussed in an appropriate way. Robert Dutt: It’s an interesting indicator. If the conversation isn’t happening early, it suggests we’re still early in the cycle, and that’s an opportunity for an MSP to stake out a brand in this space. Robert Falzon: Exactly. At this very moment, anyone in the partner ecosystem should be looking at their internal systems and processes and finding out how compliant they are personally. If you don’t understand your internal architecture and what partnerships you have in your own pipeline, you’re going to be well behind when it actually comes to implementation. Robert Dutt: Great insights. Thank you very much for your time, Rob. Robert Falzon: Thank you so much. Robert Dutt: There you have it, Rob Falzon from Check Point Software Technologies. I’d like to thank Rob for his time and for a conversation that I think went well beyond the usual talking points. Thank you for listening. Here’s a few things that stood out for me from this conversation. First, there’s a really important distinction between data residency and data sovereignty that Rob laid out cleanly. Residency is about where the data sits at rest. Sovereignty is about the entire chain—processing, storage, the works—and making sure none of it leaves the country’s control sphere. If your customers think having a Canadian data center checks the compliance box, that’s a conversation worth having with them. Second, there was that striking point about AI data exfiltration. A recent U.S. legal judgment held that all data entered into ChatGPT belongs to OpenAI. If your customers are using AI services and don’t know where that data is going and who owns it once it gets there, that’s a risk that most people simply aren’t talking about yet. And that brings me to what I think was the most telling moment: Rob’s candid admission that even after Check Point’s Canada data region announcement, he’s not getting a lot of calls about data residency. That tells me we’re still early. The regulatory picture is coming into focus, but it’s not fully formed yet, and a lot of partners and customers are in wait-and-see mode. That’s actually an opportunity. If you’re an MSP who moves now—gets educated on the regulatory landscape, audits your own internal compliance, and starts leading the sovereignty conversation with your customers—you have a chance to stake out real expertise and become the trusted voice before this becomes urgent and everyone’s scrambling. Follow or subscribe to the show. You can find In The Channel on Apple Podcasts, Spotify, YouTube, and most podcast directories. Ratings and reviews are always appreciated—they help other folks in the channel find us. Until next time, I’m Robert Dutt for ChannelBuzz.ca, and I’ll see you in the channel.

Francia quiere dejar Windows: el plan que sacude a Europa y genera dudasEl gobierno apuesta por Linux y software libre, pero mantiene contratos millonarios con Microsoft Por Félix Riaño @LocutorCo  Francia está tomando una decisión que podría cambiar el rumbo tecnológico de Europa. El gobierno anunció que empezará a dejar atrás el sistema operativo Windows para adoptar Linux en su administración pública. La idea es sencilla de explicar: depender menos de empresas tecnológicas de Estados Unidos y tener mayor control sobre sus propios datos y sistemas.  Pero cuando uno mira con más detalle, la situación es mucho más compleja de lo que parece.  El Estado francés ya tiene herramientas propias funcionando. Una de ellas es LaSuite, una especie de “oficina digital” que incluye chat, videollamadas y edición de documentos. Hoy está disponible en varios ministerios y ya cuenta con cientos de miles de usuarios activos cada mes. También existe Tchap, una aplicación de mensajería segura que usan funcionarios y que incluso es obligatoria para ministros.  A esto se suman plataformas como Onyxia, que permite hacer análisis de datos dentro de infraestructuras controladas por el propio Estado. Todo apunta a una estrategia clara: que la información pública no dependa de servicios extranjeros.  Pero aquí aparece el gran problema.   Mientras el gobierno promueve el uso de software libre, el sistema educativo firmó un contrato con Microsoft por unos 152 millones de euros, que cubre cerca de un millón de computadores hasta el año 2029. Esto genera una contradicción evidente. Por un lado, se habla de independencia tecnológica. Por el otro, se refuerza el uso de herramientas tradicionales en sectores clave como la educación.  Y ese detalle es más importante de lo que parece. Porque las herramientas que se usan en el colegio suelen marcar las que las personas usan en su vida profesional. En otras palabras, el cambio cultural se vuelve mucho más difícil.  También hay retos técnicos. Muchos documentos oficiales están en formatos propios de Microsoft, y trasladarlos a alternativas abiertas no siempre funciona bien. Archivos con macros, formularios complejos o diseños avanzados pueden perder funcionalidad.  Además, cambiar de sistema no es solo instalar otro programa. Implica formar a miles de personas que llevan años trabajando con las mismas herramientas. Y eso toma tiempo, dinero y paciencia.   Aun así, Francia decidió avanzar.   El gobierno ya pidió a todos los ministerios que presenten planes para reducir su dependencia tecnológica antes de que termine el año. También anunció una salida progresiva de Windows en favor de Linux, empezando por algunas áreas del Estado.   Este cambio no será inmediato. De hecho, el primer grupo afectado es pequeño. Pero la intención es escalarlo poco a poco hasta cubrir gran parte de la administración pública.  Francia no está sola en este camino. Alemania ya decidió que el formato de documentos ODF será obligatorio en su administración. Dinamarca y algunas regiones alemanas también están migrando hacia software libre.   Todo esto responde a una preocupación creciente en Europa: qué tan seguro es depender de empresas extranjeras para manejar información sensible.  Un caso reciente encendió las alarmas. Un juez de la Corte Penal Internacional perdió acceso a servicios digitales tras sanciones de Estados Unidos. Eso demostró que, incluso si los datos están en Europa, las decisiones pueden tomarse desde fuera.  Además, leyes como el Cloud Act permiten a las autoridades estadounidenses solicitar datos a empresas de su país, aunque esos datos estén almacenados en servidores europeos.   Por eso, la apuesta por Linux y el software libre no es solo técnica. Es política.  Pero tampoco es un camino fácil. En el pasado, ciudades como Múnich intentaron migrar a Linux y luego regresaron a Windows. Y aunque hoy la tecnología ha mejorado, los retos siguen ahí.  En contraste, algunas ciudades más pequeñas en Francia avanzan más rápido. Lugares como Lyon o Grenoble ya usan sistemas completamente basados en software libre. Su tamaño les permite cambiar con menos obstáculos.  Al final, lo que está pasando en Francia puede ser el inicio de algo más grande. Un movimiento europeo para recuperar el control de su tecnología.   La pregunta ahora es si ese cambio logrará avanzar al ritmo que el propio continente necesita.  

La vraie révolution n'est plus dans le Cloud, mais dans votre poche Par Régis BAUDOUIN « Désolé, je n’ai pas de connexion internet. » Cette phrase, qui rendait vos assistants “intelligents” totalement inutiles en zone blanche ou dans les transports appartient désormais au passé. En ce mois d’avril 2026, la tendance s’inverse radicalement : les géants de la tech et les champions de l’Open Source ne jurent plus que par les SLM (Small Language Models). Ces modèles réduits, capables de prouesses comparables aux géants de 2024, mais tournant à 100 % localement sur votre smartphone ou votre PC, marquent la fin de la dépendance aux data centers énergivores. Bienvenue dans l’ère de l’Intelligence artificielle souveraine par défaut. En clair fini l’exigence de disposer d’un carte graphique GPU pour héberger une Intelligence artificielle, un smartphone va suffire. C’est un mouvement de fond important. Pour le décideur, cette démesure des modèles LLM s’accompagne d’une “taxe sur l’inefficacité” de plus en plus lourde : factures cloud opaques, latence incompatible avec l’expérience utilisateur et risques persistants sur la souveraineté des données. Nous assistons bien aujourd’hui à un basculement stratégique : le passage vers l’Intelligence artificielle utilitaire. Les Small Language Models (SLM) ne sont pas de simples versions “dégradées” des géants ; ils représentent une réponse agile, souveraine et économiquement viable aux besoins réels des entreprises. L’intelligence ne se mesure plus au volume de données ingurgitées, mais à l’efficacité du résultat produit. Qu’est-ce qu’un SLM ? Depuis trois ans, la course à l'Intelligence artificielle s'est résumée à une démesure de paramètres. Plus le modèle était gros (LLM – Large Language Models), plus il semblait intelligent. Mais cette approche a atteint ses limites : coût écologique désastreux, latence et surtout, une opacité totale sur le traitement des données. La rupture de 2026 s’appelle le SLM. Grâce à des techniques de “quantification” (une compression mathématique de pointe) et à l’émergence des puces NPU (Neural Processing Unit) intégrées dans nos processeurs, un modèle de 3 à 7 milliards de paramètres peut désormais raisonner aussi bien que GPT-4. La différence ? Il pèse quelques gigas et réside physiquement dans le processeur de votre appareil. L’intelligence ne voyage plus sur le réseau ; elle habite votre machine. A propos des NPU, ils sont maintenant intégrés aux derniers processeurs Intel et la rumeur veut que le prochain Windows 12 exige des NPU pour fonctionner. L’idée qu’un modèle plus petit est nécessairement moins intelligent est un mythe qui s’effondre. En se concentrant sur moins de 10 milliards de paramètres, des modèles comme Phi-3.5 Mini (3,8B), Llama 3.1 8B ou Mistral Nemo 12B rivalisent avec les mastodontes sur des tâches métier précises. La force d’un modèle comme Phi-3.5 réside notamment dans sa capacité à gérer un contexte long de 128 000 tokens, le rendant idéal pour l’analyse de documents juridiques ou techniques complexes. La véritable rupture réside dans le duo gagnant : Petit moteur + Grande bibliothèque. Grâce au RAG (Retrieval Augmented Generation) et aux bases de données vectorielles, un SLM n’a plus besoin de “tout savoir”. Il utilise sa capacité de raisonnement pour interroger en temps réel les documents internes de l’entreprise. Cette approche élimine le besoin de réentraînements coûteux, réduit drastiquement les hallucinations et garantit que l’Intelligence artificielle parle la langue de votre expertise. Le point clé est de connecter ces SML sur vos données et pas sur la totalité de du contenu d’internet. Vous être plus pertinent et les réponses sont propres à vos données passées sasn risque des les faire fuiter. Le triple dividende : Souveraineté, Écologie, Vitesse Le passage au local n’est pas qu’une prouesse technique, c’est un acte politique et stratégique pour l’utilisateur. Quels sont les bénéfice de cette évolution de modèle d’Intelligence artificielle? Souveraineté et Vie Privée : C'est l'argument massue. Puisque la donnée ne quitte jamais votre appareil, le risque d’espionnage industriel via le Cloud Act américain est fortement réduit. C'est enfin une Intelligence artificielle compatible avec le secret professionnel, le secret médical et la vie privée la plus stricte. Ca c’est sur le papier. En réalité les services sont liées à leurs fournisseurs car certains traitements nécessitent d’interroger une base distante. Sobriété Énergétique : Selon les derniers rapports de l'Arcep (2025/2026), interroger une Intelligence artificielle locale consomme jusqu’à 100 fois moins d’énergie que de solliciter un serveur à l’autre bout du monde. La relocalisation du calcul est la clé de la décarbonation du numérique. Surtout que le puissance nécessaire est moindre. Latence Zéro : Plus besoin d’attendre que le serveur de la Silicon Valley vous réponde. La traduction instantanée ou la rédaction de mails deviennent aussi fluides qu’un simple correcteur orthographique, même en mode avion. Les maîtres du jeu local : Mistral, Apple et l’Open Source Sur ces solutions SLM qui sont les plus avancées ? Mistral AI : Le fleuron français confirme sa domination sur le segment de l’efficience. Leurs modèles “Small” et “Tiny” sont devenus la référence européenne pour les administrations soucieuses de leur indépendance. J’ai eu l’occasion de le tester en autohébergèrent. C’est plutôt efficace. L’écosystème Open Source : Des plateformes comme Hugging Face (via leur Open LLM Leaderboard) ont permis une accélération sans précédent. En quelques mois, la communauté a optimisé des modèles pour qu’ils tournent sur de simples Raspberry Pi ou des smartphones de milieu de gamme. Apple et Google : Avec Apple Intelligence 2.0 et Gemini Nano, les deux géants ont dû se résoudre à suivre la marche : le smartphone de 2026 est avant tout un terminal d’Intelligence artificiellelocale. Le combat faire rage pour imposer son Intelligence artificielle locale avant que les régulateurs interviennent (comme pour internet explorer) Le passage aux SLM transforme radicalement le plan d’investissement IA. L’argument massue ? L’accessibilité matérielle. Là où les LLM exigent des GPU haute performance (cartes Nvidia H100) rares et onéreux, les SLM peuvent fonctionner sur des CPU standard (Intel ou AMD) ou du matériel standard déjà présent dans l’entreprise. La rentabilité est sans appel : Investissement initial : 12 000 à 25 000 € pour un SLM spécialisé, contre 45 000 à 80 000 € pour un LLM. Coût mensuel d’infrastructure : 400 à 800 € contre 1 200 à 2 400 € pour les modèles géants. Time-to-ROI : Atteint en 3 à 6 mois seulement, contre près d’un an pour un LLM. Infrastructure : Utilisation possible de serveurs existants sans upgrade GPU massif. C’est vraiment un game changer à une moment ou le pris de la RAM et des processeurs s’envole. Comment un modèle de quelques milliards de paramètres peut-il être aussi pertinent ? L’ingénierie utilise trois leviers principaux : La Distillation : C’est le transfert de connaissances d’un modèle “professeur” (LLM) vers un “élève” (SLM). L’élève n’apprend pas seulement les réponses, mais reproduit les schémas de raisonnement de son mentor avec une fraction de sa taille. L’Élagage (Pruning) : Cette technique consiste à supprimer les neurones et paramètres redondants. On “taille” le modèle pour ne conserver que les connexions vitales, réduisant la taille sans sacrifier la précision. La Quantification : Il s’agit de réduire la précision numérique des données. Par exemple, passer de 32 bits à 8 bits permet de diviser par quatre l’empreinte mémoire. Pour vulgariser : si vous mesurez la température, stocker 20,5234°C (32 bits) est inutile là où 20,5°C (8 bits) suffit largement à l’application. On gagne en vitesse et en espace avec une perte de pertinence négligeable. Le retour à la propriété numérique Il y a dix ans, on nous expliquait que pour être “smart”, tout devait être dans le Cloud. En 2026, la boucle est bouclée : la véritable intelligence est celle que vous possédez, celle qui ne dépend d’aucun abonnement, d’aucun câble sous-marin et d’aucune juridiction étrangère. Le passage aux SLM est plus qu’une mise à jour logicielle, c’est une reprise de pouvoir. En rapatriant l’Intelligence artificielle dans nos poches, nous reprenons les clés de notre souveraineté numérique. Cela ne veut pas dire que tout sera gratuit. Comme souvent la réalité sera hybride. Le local pour les traitement courants et répétitif, le cloud en LLM pour la recherche exploratoire et le raisonnement de haut niveau. Pour en savoir plus sur l’IAThe post La vraie révolution de l'intelligence artificielle n'est plus dans le Cloud, mais dans votre poche first appeared on XY Magazine.

Er gehört zu den Urgesteinen der europäischen Open Source Bewegung und legt es mit seinem Unternehmen darauf an, Microsoft möglichst viel Geschäft in Europa wegzunehmen: Frank Karlitschek ist heute mit Nextcloud mit einem Team 165 Mitarbeitern profitabel unterwegs und treibt gleichzeitig mit anderen europäischen Firmen Open Source Software als Alternative zu proprietärer US-Software voran. Im Podcast spricht Karlitschek über:☁️ Was ist Nextcloud und warum setzen europäische Ministerien auf die Open-Source-Alternative zu Microsoft 365

Parce que… c'est l'épisode 0x736! Préambule Nous sommes à la Cage durant un match des Canadiens. Le bruit ambiant a fait que nous parlons en “criant”, pour nous entendre. Le lendemain, je n'avais plus de voix. Shameless plug 14 au 17 avril 2026 - Botconf 2026 20 au 22 avril 2026 - ITSec Code rabais de 15%: Seqcure15 28 et 29 avril 2026 - Cybereco Cyberconférence 2026 9 au 17 mai 2026 - NorthSec 2026 3 au 5 juin 2026 - SSTIC 2026 19 septembre 2026 - Bsides Montréal 1 au 3 décembre 2026 - Forum INCYBER - Canada 2026 24 et 25 février 2027 - SéQCure 2027 Description Introduction et mise en contexte Dans cet épisode spécial, l'animateur reçoit Nicolas Bédard, un expert en infonuagique qui a évolué chez Google avant de rejoindre Palo Alto Networks. Le prétexte de l'épisode est concret : Nicolas s'apprête à se rendre à Paris pour participer à la grande conférence de S3NS, une entreprise issue d'un partenariat franco-américain au cœur de la souveraineté numérique en France. C'est l'occasion idéale pour aborder un sujet devenu incontournable, alors qu'il était encore marginal il y a quelques années seulement. S3NS : la franchise numérique franco-américaine Pour expliquer le modèle de S3NS, Nicolas utilise une analogie parlante : celle de la franchise de restauration rapide. Comme un franchisé de McDonald's ou Tim Hortons qui investit dans son propre établissement tout en appliquant la recette du franchiseur, S3NS est une entreprise majoritairement détenue par le groupe français Thales, qui a obtenu de Google le « livre de recettes » de son infrastructure infonuagique — tant sur le plan matériel (serveurs, stockage, réseau) que logiciel (les services Google Cloud). S3NS propose deux offres distinctes : Cryptance : une couche de souveraineté appliquée directement sur l'infrastructure de Google Cloud. Les charges de travail (workloads) sont placées dans un dossier (folder) auquel sont attachées des règles de gouvernance strictes — notamment des clés de chiffrement gérées par S3NS — garantissant que les données demeurent sur le territoire français. Prémiance : un niveau supérieur, où Thales et Google ont co-construit des centres de données physiques en France. Ces infrastructures, propriété de S3NS, utilisent l'architecture de Google Cloud mais sont opérées de façon indépendante. Prémiance est certifiée SecNumCloud, la norme française de cybersécurité gérée par l'ANSSI, ce qui constitue une validation rigoureuse issue de plusieurs années d'expérimentation sur le terrain. La force des règles technologiques Un point central de la discussion porte sur la nature même des règles technologiques : contrairement aux règles entre humains, qui peuvent être interprétées, négociées ou contournées, les règles appliquées par une machine sont absolues. Nicolas illustre cela avec une anecdote personnelle : lors d'un projet pilote, des parties prenantes avaient exigé l'application de règles strictes, malgré ses mises en garde. En moins de 24 heures, le projet avait été abandonné parce que les participants ne pouvaient tout simplement pas « tolérer » des règles que la machine appliquait sans compromis. Cette rigidité est précisément ce qui rend la souveraineté numérique crédible. Le chiffrement, par exemple, ne se négocie pas : sans la clé détenue par l'entité française, même un acteur américain ne peut accéder aux données. Les craintes liées au Cloud Act américain, souvent agitées dans le débat public, se heurtent à cette réalité technique implacable. Le rôle de Palo Alto Networks Palo Alto Networks s'intègre dans cet écosystème parce que les clients qui migrent vers un cloud souverain s'attendent à retrouver les outils de sécurité qu'ils utilisaient dans leurs environnements traditionnels. Quatre produits de Palo Alto sont particulièrement concernés, tous basés sur des machines virtuelles (VM-based) : Les pare-feu de nouvelle génération (Next-Gen Firewalls) Prisma AIRS, le pare-feu dopé à l'IA pour contrer les injections de prompts et les fuites de données Panorama, la console centralisée de gestion des pare-feu XSIAM, la plateforme d'orchestration pour la réponse aux incidents Ces produits, parce qu'ils s'exécutent sur des machines virtuelles plutôt que sous forme de SaaS hébergé chez Palo Alto, peuvent être déployés aussi bien sur Cryptance que sur Prémiance, en conformité avec SecNumCloud. La vraie valeur du nuage et la modernisation des applications Nicolas insiste sur une vision souvent oubliée : le nuage, c'est avant tout du logiciel. Sa valeur réelle réside dans l'élasticité, la dématérialisation, les microservices et les conteneurs — pas dans la simple virtualisation de serveurs existants. Beaucoup d'entreprises ont mal abordé leur passage au nuage en y transférant des systèmes vieillissants sans les moderniser, ce qui les rend aujourd'hui incompatibles avec les exigences de la souveraineté numérique, qui impose une partition claire des services à l'échelle nationale. Rouler une application moderne — découpée en microservices indépendants, stateless, élastiques — sur un cloud souverain, c'est là que réside la vraie promesse de la prochaine génération d'infrastructures souveraines. Le retard canadien et les leçons européennes La France a commencé à réfléchir à la souveraineté numérique dès 2017 environ. Elle a construit ses normes, les a testées, les a fait évoluer. Le Canada, lui, s'est réveillé bien plus tard, emporté qu'il était par la « locomotive » américaine. Aujourd'hui, il n'existe pas d'équivalent canadien à SecNumCloud, et la maturité réglementaire nécessaire pour encadrer une véritable infrastructure souveraine est encore à construire. Les deux interlocuteurs s'entendent pour dire que le Canada devrait s'inspirer davantage de ce qui se fait en Europe — en France, en Allemagne, à travers des initiatives comme celle de la ville de Munich — plutôt que de se limiter à répéter le mot « souveraineté » sans se donner les outils concrets pour la réaliser. Cela implique une réarchitecture profonde des systèmes existants, un développement de compétences nationales et un cadre réglementaire sérieux. Conclusion La souveraineté numérique est un chantier de longue haleine, mais l'exemple de S3NS en France démontre qu'il est possible de concilier technologie américaine de pointe et contrôle national effectif. Nicolas conclut sur une note d'enthousiasme, impatient de voir la conférence parisienne et d'espérer que des modèles similaires finissent par voir le jour au Canada. Un suivi de l'épisode est d'ores et déjà prévu. Collaborateurs Nicolas-Loïc Fortin Nicolas Bédard Crédits Montage par Intrasecure inc Locaux réels par La Cage - Complexe Desjardins

In Episode 156 des c't-Datenschutz-Podcasts widmen sich Holger und Joerg einem Thema, das die Datenschutzwelt umtreibt wie kaum ein anderes: Microsoft 365. Zu Gast ist Kai Korte, Geschäftsführer der lexICT GmbH, Rechtsanwalt für IT- und Datenschutzrecht in Hannover und Lehrbeauftragter an der Hochschule Hannover. Zunächst beleuchten die drei eine auffällige aktuelle Entwicklung: Die Zahl der Datenschutzbeschwerden hat sich bei einigen Aufsichtsbehörden nahezu verdoppelt, wie aus den neuen Tätigkeitsberichten hervorgeht. Baden-Württemberg verzeichnete 2025 einen Anstieg von rund 4.000 im Jahr 2024 auf über 7.600, Hamburg von 2.600 auf 4.200. Mehrere Behörden führen das auch auf KI-generierte Beschwerden zurück. Betroffene lassen sich offenbar per Chatbot fertige Texte erstellen und schicken diese an die Aufsicht. Die Behörden warnen, dass sie an ihre Kapazitätsgrenzen stoßen. Ebenfalls Thema ist ein aktuelles Urteil des Europäischen Gerichtshofs (EuGH) zum Auskunftsrecht. Im Fall "Brillen Rottler" entschied das Gericht, dass bereits ein erster Auskunftsantrag als rechtsmissbräuchlich eingestuft werden kann. Ein Mann hatte sich zum Newsletter eines Optikers angemeldet und nur 13 Tage später ein umfassendes Auskunftsersuchen gestellt. Der EuGH setzt allerdings hohe Hürden: Der Verantwortliche muss objektiv nachweisen, dass der Antragsteller gar nicht die Ziele des Datenschutzrechts verfolgt, und er muss zusätzlich eine missbräuchliche Absicht belegen. Korte warnt davor, das Urteil vorschnell zu verallgemeinern, es bleibe ein Ausnahmefall. Dann steigen die drei ins Hauptthema ein: Microsoft 365. Holger macht zunächst deutlich, wie komplex das Produktuniversum ist. Hinter dem Namen verbergen sich zahlreiche Einzelprodukte – von Office-Anwendungen bis zu Teams und SharePoint. Die Preise reichen von knapp 11 Euro pro Nutzer und Monat für kleinere Unternehmen bis über 55 Euro im großen Enterprise-Paket. Viele Sicherheitsfunktionen sind nur in den teuren Varianten enthalten, und Microsoft verschiebt regelmäßig Features zwischen den Paketen. Datenschutzrechtlich bleibt Microsoft 365 ein Minenfeld. Zwar nimmt der US-Konzern am Transatlantic Data Privacy Framework teil, was den Datentransfer in die USA formal absichert. Kau weist aber darauf hin, dass der CLOUD Act und weitreichende Zugriffsbefugnisse US-amerikanischer Behörden weiterhin problematisch sind. Microsoft sichert vertraglich zu, Daten europäuscher Unternehmen innerhalb der EU zu speichern -– die sogenannte EU Data Boundary. Doch Wartungszugriffe aus den USA sind weiterhin möglich, und ein Microsoft-Manager räumte öffentlich ein, dass man sich einem US-Herausgabebeschluss kaum widersetzen könnte. Ein zentraler Kritikpunkt betrifft die Verarbeitung von Telemetrie- und Diagnosedaten. Microsoft sammelt im Hintergrund umfangreich Metadaten – Anmeldezeiten, Gesprächsdauern, Teilnehmerlisten, Nutzeraktivitäten. Welche Datenfelder genau erfasst werden und was damit geschieht, bleibt weitgehend intransparent. Der hessische Datenschutzbeauftragte hat Ende 2025 einen Bericht veröffentlicht, der den Einsatz von Microsoft 365 in öffentlichen Stellen unter bestimmten Bedingungen für möglich erklärt. Allerdings hat er sich dabei ausschließlich auf Microsofts eigene Angaben verlassen, ohne die Datenflüsse technisch zu überprüfen. Holger verweist zudem auf Microsofts Zusatzprodukte Sentinel und Purview, mit denen Unternehmen das Verhalten ihrer Mitarbeiter analysieren können – bis hin zur KI-gestützten Erkennung potenzieller Whistleblower oder Datendiebe. Kai räumt ein, dass Purview ein mächtiges Sicherheitswerkzeug sei, warnt aber vor dem schmalen Grat zwischen legitimer Sicherheitsüberwachung und unzulässigem Profiling von Beschäftigten. Ob Unternehmen vor dem Einsatz von Microsoft 365 eine Datenschutz-Folgenabschätzung durchführen müssen, beantwortet Kai differenziert: Die Aufsichtsbehörden fordern das zwar, doch bei kleineren Unternehmen ohne besondere Datenkategorien sei das nicht

Es ist ein Interessenkonflikt. Passwortmanager vereinfachen einem das Leben, weil man mit Ihnen mehr oder weniger komfortabel für jedes (Online-)Konto ein eigenes und sicheres Passwort vergeben kann. So kann man hunderte Passwörter einsetzen, ohne ein fotografisches Gedächtnis zu besitzen. Gleichzeitig aber bietet man eine sehr attraktive Angriffsfläche, gerade Online-Passwortmanager, die die Passwörter via Server zwischen mehreren Endgeräten synchronisieren. Dieser Datenschatz erweckt auch das Interesse von Behörden. Populäre Passwortmanager – Bitwarden, LastPass, Dashlane – aus den USA kommen oder von dortigen Firmen entwickelt werden. Und US-Behörden könnten mit Verweis auf Cloud Act und Foreign Intelligence Surveillance Act (FISA) Zugriff auf die Daten verlangen. Eine aktuelle Untersuchung der ETH Zürich zeigte zudem, dass trotz Ende-zu-Ende-Verschlüsselung unter bestimmten Bedingungen Passwörter abgreifbar sein können – etwa wenn der Server manipuliert wird. Manch einer wird sich daher fragen, ob man die eigenen Passwörter nicht vielleicht in souveränere Gefilde umzieht. Welche Alternativen es gibt und wie sinnvoll die sind, diskutieren die c't-Redakteure Jan Schüßler und Niklas Dierking in der neuen Folge von c't uplink mit Moderator Keywan Tonekaboni. Jan Schüßler hat fünf Passwortmanager getestet, die entweder aus Europa stammen und/oder Open Source sind – sowohl cloud-basierte Dienste als auch lokale Lösungen wie KeepassXC/KeepassDX. Niklas Dierking hat Passbolt auf einem eigenen Server installiert und ordnet die Erfahrung im Vergleich zu VaultWarden ein. Die drei c't Redakteure vergleichen Komfort, Kosten und Sicherheitskonzepte der verschiedenen Alternativen. Außerdem gibt das Team praktische Tipps für den Umstieg von einem Passwortmanager zum anderen, erklärt Synchronisierungswege über Syncthing oder Nextcloud und warnt vor typischen Stolperfallen bei der Migration.

Es ist ein Interessenkonflikt. Passwortmanager vereinfachen einem das Leben, weil man mit Ihnen mehr oder weniger komfortabel für jedes (Online-)Konto ein eigenes und sicheres Passwort vergeben kann. So kann man hunderte Passwörter einsetzen, ohne ein fotografisches Gedächtnis zu besitzen. Gleichzeitig aber bietet man eine sehr attraktive Angriffsfläche, gerade Online-Passwortmanager, die die Passwörter via Server zwischen mehreren Endgeräten synchronisieren. Dieser Datenschatz erweckt auch das Interesse von Behörden. Populäre Passwortmanager – Bitwarden, LastPass, Dashlane – aus den USA kommen oder von dortigen Firmen entwickelt werden. Und US-Behörden könnten mit Verweis auf Cloud Act und Foreign Intelligence Surveillance Act (FISA) Zugriff auf die Daten verlangen. Eine aktuelle Untersuchung der ETH Zürich zeigte zudem, dass trotz Ende-zu-Ende-Verschlüsselung unter bestimmten Bedingungen Passwörter abgreifbar sein können – etwa wenn der Server manipuliert wird. https://www.heise.de/news/Schwachstellen-in-Cloud-basierten-Passwort-Managern-11179212.html Manch einer wird sich daher fragen, ob man die eigenen Passwörter nicht vielleicht in souveränere Gefilde umzieht. Welche Alternativen es gibt und wie sinnvoll die sind, diskutieren die c't-Redakteure Jan Schüßler und Niklas Dierking in der neuen Folge von c't uplink mit Moderator Keywan Tonekaboni. Jan Schüßler hat fünf Passwortmanager getestet, die entweder aus der EU stammen oder Open-Source-Community-Projekte sind – sowohl cloud-basierte Dienste als auch lokale Lösungen wie KeepassXC/KeepassDX. Niklas Dierking hat Passbolt auf einem eigenen Server installiert und ordnet die Erfahrung im Vergleich zu VaultWarden ein. Die drei c't Redakteure vergleichen Komfort, Kosten und Sicherheitskonzepte der verschiedenen Alternativen. Lösungen – etwa fehlende biometrische Entsperrung am Desktop. Außerdem gibt das Team praktische Tipps für den Umstieg von einem Passwortmanager zum anderen, erklärt Synchronisierungswege über Syncthing oder Nextcloud und warnt vor typischen Stolperfallen bei der Migration. Zu Gast im Studio: Niklas Dierking und Jan Schüßler Host: Keywan Tonekaboni Produktion: Tobias Reimer Im Newsletter c't Open Source Spotlight ordnen Keywan und Niklas aktuelle Entwicklungen rund um freie Software ein und stellen innovative Open-Source-Anwendungen vor. Jetzt anmelden und an jedem zweiten Freitag eine neue Ausgabe erhalten. https://www.heise.de/newsletter/anmeldung.html?id=ct-opensource Passwortmanager: Gute Gründe für europäische Clouds oder Self Hosting: https://www.heise.de/ratgeber/Passwortmanager-Gute-Gruende-fuer-europaeische-Clouds-oder-Self-Hosting-11172904.html Fünf Open-Source-Passwortmanager im Vergleich: https://www.heise.de/ratgeber/Fuenf-Open-Source-Passwortmanager-im-Vergleich-11172914.html Passbolt: Den europäischen Open-Source-Passwortmanager selbst hosten: https://www.heise.de/ratgeber/Passbolt-Den-europaeischen-Open-Source-Passwortmanager-selbst-hosten-11172920.html Anleitung: Von LastPass zum Passwortmanager KeePassXC wechseln: https://www.heise.de/ratgeber/Anleitung-Von-LastPass-zum-Passwortmanager-KeePassXC-wechseln-5075363.html Raspberry Pi als zentralen Backup-Server mit Syncthing einrichten - https://www.heise.de/ratgeber/Raspi-Backup-Plattformunabhaengiges-Backup-mit-Syncthing-einrichten-6111168.html - https://www.heise.de/ratgeber/Raspberry-Pi-als-zentralen-Backup-Server-mit-Syncthing-einrichten-6109494.html Anleitung: Raspberry Pi als Passwort-Server einrichten: https://www.heise.de/ratgeber/Anleitung-Raspberry-Pi-als-Passwort-Server-einrichten-6005925.html

Es ist ein Interessenkonflikt. Passwortmanager vereinfachen einem das Leben, weil man mit Ihnen mehr oder weniger komfortabel für jedes (Online-)Konto ein eigenes und sicheres Passwort vergeben kann. So kann man hunderte Passwörter einsetzen, ohne ein fotografisches Gedächtnis zu besitzen. Gleichzeitig aber bietet man eine sehr attraktive Angriffsfläche, gerade Online-Passwortmanager, die die Passwörter via Server zwischen mehreren Endgeräten synchronisieren. Dieser Datenschatz erweckt auch das Interesse von Behörden. Populäre Passwortmanager – Bitwarden, LastPass, Dashlane – aus den USA kommen oder von dortigen Firmen entwickelt werden. Und US-Behörden könnten mit Verweis auf Cloud Act und Foreign Intelligence Surveillance Act (FISA) Zugriff auf die Daten verlangen. Eine aktuelle Untersuchung der ETH Zürich zeigte zudem, dass trotz Ende-zu-Ende-Verschlüsselung unter bestimmten Bedingungen Passwörter abgreifbar sein können – etwa wenn der Server manipuliert wird. https://www.heise.de/news/Schwachstellen-in-Cloud-basierten-Passwort-Managern-11179212.html Manch einer wird sich daher fragen, ob man die eigenen Passwörter nicht vielleicht in souveränere Gefilde umzieht. Welche Alternativen es gibt und wie sinnvoll die sind, diskutieren die c't-Redakteure Jan Schüßler und Niklas Dierking in der neuen Folge von c't uplink mit Moderator Keywan Tonekaboni. Jan Schüßler hat fünf Passwortmanager getestet, die entweder aus der EU stammen oder Open-Source-Community-Projekte sind – sowohl cloud-basierte Dienste als auch lokale Lösungen wie KeepassXC/KeepassDX. Niklas Dierking hat Passbolt auf einem eigenen Server installiert und ordnet die Erfahrung im Vergleich zu VaultWarden ein. Die drei c't Redakteure vergleichen Komfort, Kosten und Sicherheitskonzepte der verschiedenen Alternativen. Lösungen – etwa fehlende biometrische Entsperrung am Desktop. Außerdem gibt das Team praktische Tipps für den Umstieg von einem Passwortmanager zum anderen, erklärt Synchronisierungswege über Syncthing oder Nextcloud und warnt vor typischen Stolperfallen bei der Migration. Zu Gast im Studio: Niklas Dierking und Jan Schüßler Host: Keywan Tonekaboni Produktion: Tobias Reimer Im Newsletter c't Open Source Spotlight ordnen Keywan und Niklas aktuelle Entwicklungen rund um freie Software ein und stellen innovative Open-Source-Anwendungen vor. Jetzt anmelden und an jedem zweiten Freitag eine neue Ausgabe erhalten. https://www.heise.de/newsletter/anmeldung.html?id=ct-opensource Passwortmanager: Gute Gründe für europäische Clouds oder Self Hosting: https://www.heise.de/ratgeber/Passwortmanager-Gute-Gruende-fuer-europaeische-Clouds-oder-Self-Hosting-11172904.html Fünf Open-Source-Passwortmanager im Vergleich: https://www.heise.de/ratgeber/Fuenf-Open-Source-Passwortmanager-im-Vergleich-11172914.html Passbolt: Den europäischen Open-Source-Passwortmanager selbst hosten: https://www.heise.de/ratgeber/Passbolt-Den-europaeischen-Open-Source-Passwortmanager-selbst-hosten-11172920.html Anleitung: Von LastPass zum Passwortmanager KeePassXC wechseln: https://www.heise.de/ratgeber/Anleitung-Von-LastPass-zum-Passwortmanager-KeePassXC-wechseln-5075363.html Raspberry Pi als zentralen Backup-Server mit Syncthing einrichten - https://www.heise.de/ratgeber/Raspi-Backup-Plattformunabhaengiges-Backup-mit-Syncthing-einrichten-6111168.html - https://www.heise.de/ratgeber/Raspberry-Pi-als-zentralen-Backup-Server-mit-Syncthing-einrichten-6109494.html Anleitung: Raspberry Pi als Passwort-Server einrichten: https://www.heise.de/ratgeber/Anleitung-Raspberry-Pi-als-Passwort-Server-einrichten-6005925.html

Benoît Martel, président et fondateur de R2I, explique que la souveraineté numérique n'est plus un débat théorique, mais une préoccupation très concrète pour les entreprises canadiennes. Dans un contexte où le Cloud Act américain inquiète davantage, il rappelle qu'une donnée hébergée par un fournisseur américain peut rester soumise à la juridiction des États-Unis, même si elle se trouve physiquement au Canada. R2I, qui offre depuis plus d'une décennie une solution infonuagique canadienne baptisée Edwin, dit constater une forte hausse de la demande pour des services locaux, souverains et redondants, notamment de la part d'organisations qui arrivent à l'échéance de leurs contrats avec les grands hyperscalers. Benoit Martel insiste aussi sur le fait que la souveraineté numérique ne concerne pas seulement l'hébergement, mais aussi la sécurité, la conformité et la capacité de reprise en cas de crise. R2i est partenaire de Mon Carnet. Merci à R2i de soutenir la production de ce podcast.

The episode's central development is the ongoing dispute between the U.S. Department of Defense and Anthropic regarding Pentagon demands for unrestricted access to Claude, Anthropic's AI model. According to Dave Sobel, the Pentagon has threatened to sever ties or invoke the Defense Production Act if the company does not comply, seeking capabilities that Anthropic argues may be illegal—specifically mass surveillance without warrants and autonomous weapons systems without human control. This move exposes Managed Service Providers (MSPs) serving defense contractors to unpredictable legal, operational, and compliance risks embedded in their AI workflows. The analysis highlights that a commercial AI provider's acceptable use policy now intersects directly with national security policy, and even partial vendor compliance can trigger regulatory or legal instability for dependent organizations. For MSPs, this means that building service offerings on AI infrastructures without clear fallback strategies or documented policy change clauses can lead to unmanageable risk and liability in the event of provider or legal regime shifts. Dave Sobel stresses that failing to address policy volatility as part of a managed service amounts to underwriting geopolitical risk without compensation. Other notable developments include the passage of the Small Business Artificial Intelligence Advancement Act, federal cybersecurity resource contraction as CISA operates with 38% staffing after layoffs, and heightened uncertainty around cloud infrastructure due to Microsoft's Azure Local “air-gapped” offering not wholly mitigating U.S. CLOUD Act exposure. Vendor news covered new AI-powered compliance features from Compliance Scorecard (version 10) and Beachhead Solutions (ComplianceEZ 2.0), Apple's accelerated retirement of Rosetta 2 translation technology, a Microsoft 365 Copilot DLP change, and continued fallout from VMware's acquisition by Broadcom, which has led to ongoing cost and trust challenges for cloud and infrastructure partners. The episode's clear implications for MSPs and IT providers are operational. Service catalogs and statements of work should actively address AI provider liability, dependency exit planning, and degraded federal cybersecurity support. Without scheduled and documented compatibility and risk reviews, MSPs absorb hidden exposure into their margins. Vendor stability can no longer be assumed, and proactive policy, renewal intelligence, and transparent advisory sessions are now required to avoid unplanned liability, budget crises, and damaged client trust. Four things to know today 00:00 Pentagon Threatens Anthropic Over Claude Access, Demands Autonomous Weapons Use 04:31 CISA Cuts, Azure Sovereignty Push Signal End of Federal MSP Safety Net 06:56 AI Compliance Tools Flood Market as MSPs Face Validation Gap 09:54 86% of Firms Cutting VMware Ties as Broadcom Renewal Costs Loom   This is the Business of Tech.    Supported by: Small Biz Thoughts Community

"It's not just about where your data lives - it's about who should, or shouldn't, have access to it."In this episode of Softcat's Explain IT podcast, host Helen Gidney, Head of Architecture at Softcat, is joined by Sabina Anja, Chief Technologist, VMware Cloud Foundation at Broadcom, and Gary Hawkins, Chief Technologist, Hybrid Platforms at Softcat, to demystify the complexities of Data Sovereignty.As organisations face increasing regulatory pressure and the rapid adoption of AI, understanding where your data lives - and who controls it - is critical. The discussion explores how governance, the Cloud Act, and GDPR are reshaping cloud strategies across Europe, driving a renewed interest in private cloud and sovereign cloud solutions.In this episode, Helen, Sabina and Gary discuss:• Defining Data Sovereignty: Why it is not just about location, but about jurisdiction, technical control, and operational access.• The Reality of Repatriation: Analysing the shift back to on-premise or Neo cloud environments to control data, without abandoning public cloud entirely.• Modern Infrastructure: How containers, Kubernetes, and AI demands are influencing infrastructure and data design.• The Power of Platforms: Meaningful insights on using VMware Cloud Foundation 9 (VCF9) to provide a unified control plane for policy-based data sovereignty.Thanks for listening to the Explain IT podcast from Softcat.This podcast is produced by The Podcast Coach. Hosted on Acast. See acast.com/privacy for more information.

Impact sur l'emploi, utilisation malveillante, perte de contrôle... L'intelligence artificielle soulève toujours autant d'interrogations et d'inquiétudes. Des démissions jettent le trouble et alimentent les questionnements. Et aussi : l'actu de la semaine.

In this episode of Shift, investor and innovation leader John Ruffolo offers a bold, forward‑looking take on the forces reshaping Canada's economic future and the massive opportunity ahead. From digital sovereignty and the evolving implications of the U.S. Cloud Act to strengthening industrial strategy and anchoring homegrown innovation, Ruffolo highlights why now is the moment for Canada to rethink how it builds wealth and scale its brightest ideas.Candid, optimistic, and deeply pragmatic, this conversation offers a roadmap for how Canada can seize this moment and build world‑leading companies on its own soil. A must‑listen for executives, investors, and builders shaping Canada's next chapter.

Robots, intelligence artificielle, dépendance aux géants américains, nouvelles lois sur Internet… Le mois de janvier a concentré toutes les fractures du numérique. Dans ce Grand Débrief, on prend le temps d'analyser ce que ces signaux disent vraiment de l'avenir de la tech.Le Grand Debrief vous est proposé en partenariat avec Free ProAvec François Sorel (Tech&Co) et Bruno Guglielminetti (Mon Carnet)CES 2026 : un salon moins spectaculaire, mais plus révélateurLe Consumer Electronics Show de Las Vegas a-t-il perdu de sa magie ? Moins d'annonces grand public, moins d'objets “wahou”, mais un salon qui confirme malgré tout plusieurs tendances lourdes : automatisation, robotique, intelligence artificielle omniprésente et montée en puissance des acteurs asiatiques. Bref, un CES 2026 plus sobre mais qui reflète mieux que jamais l'état réel de l'industrie technologique mondiale.- Voitures autonomes : la réalité derrière le fantasmeLes véhicules autonomes avancent vite… mais pas toujours là où on l'imagine. Waymo, Zoox ou Uber multiplient les expérimentations de niveau 4, capables de circuler sans conducteur dans des zones bien définies. En revanche, le niveau 5, celui d'une voiture autonome partout et en toutes circonstances, n'existe toujours pas.Contrairement au discours d'Elon Musk, le FSD de Tesla reste officiellement classé niveau 2, loin des critères d'autonomie totale.- Robots humanoïdes et “IA physique” : le vrai tournantLe CES 2026 a marqué une étape importante : le passage de l'IA logicielle à l'IA incarnée. Robots humanoïdes, machines domestiques intelligentes, automatisation du monde réel… la robotique entre dans un nouveau cycle. Si l'électromécanique et l'équilibre sont désormais maîtrisés, le véritable verrou reste l'intelligence elle-même.Les modèles d'IA actuels sont-ils capables de comprendre le monde physique, ou faudra-t-il changer de paradigme, comme le défend notamment Yann LeCun ?- La Chine, puissance technologique majeureTrès visible cette année à Las Vegas, la Chine n'est plus dans l'imitation mais dans l'exécution rapide et industrielle. Robots aspirateurs, robots humanoïdes, vidéoprojecteurs, électronique grand public : les innovations chinoises s'imposent par leur qualité et leur vitesse de développement. Un basculement stratégique majeur, qui redessine la concurrence mondiale — et interroge la place de l'Europe.Dépendance à la tech américaine : le réveil européen ?Pendant que les patrons de la tech défilaient au Forum économique mondial de Davos, le Parlement européen adoptait une résolution alertant sur la dépendance numérique de l'Europe. Cloud, logiciels, systèmes d'exploitation, IA : que se passerait-il en cas de tension politique majeure avec les États-Unis ? Faut-il craindre un "kill switch" (coupure totale) ou une dégradation des services ? La question n'est plus théorique, notamment après les menaces commerciales de Donald Trump et les débats autour du Cloud Act. Alors, peut-on réellement se passer de la tech américaine… si oui, à quel prix ?Cloud souverain : solution réelle ou illusion juridique ?AWS, Google et Microsoft multiplient les annonces de clouds souverains européens, comme le projet d'AWS European Sovereign Cloud. Mais une entité juridique locale suffit-elle à garantir une indépendance réelle ? Réseaux sociaux interdits aux mineurs : la fin de la récré ?Dernier grand sujet de ce Débrief : la loi française visant à interdire les réseaux sociaux aux moins de 15 ans. Après la loi sur la protection contre les contenus pornographiques, le RGPD, le DSA ou encore le projet Chat Control, la régulation numérique s'intensifie. Sommes-nous en train d'assister à la fin de l'Internet libre tel qu'on l'a connu ou à une tentative nécessaire de protection face à l'addiction, au temps d'écran et aux effets cognitifs sur les plus jeunes ?-----------♥️ Soutien : https://mondenumerique.info/don

First Topic - Podcast Content Plans for 2026 Every year, I like to sit down and consider what the podcast should be focusing on. Not doing so ensures every single episode will be about AI and nobody wants that. Least of all, me. If I have one more all-AI episode, my head is going to explode. With that said, most of what we talk about in this segment is AI (picard face palm.png). I think 2026 will be THE defining year for GenAI. Three years after the release of ChatGPT, I think we've hit peak GenAI hype and folks are ready for it to put up or shut up. We'll see winners grow and get acquired and losers pivot to something else. More than anything, I want to interview folks who have actually seen it work at scale, rather than just in a cool demo in a vendor sandbox. Also on the agenda for this year: The battle against infostealers and session hijacking: we didn't have a good answer in 2025. When is it coming? Will it include Macs, despite them not having a traditional TPM? The state of trust in outsourcing and third party use (Cloud, MSSPs, SaaS, contractors): 2025 was not a good year for third parties. Lots of them got breached and caused their customers a lot of pain. Also, there's the state of balkanization between the US and... the rest of the entire world. Everyone outside the US seems to be trying to derisk their companies and systems from the Cloud Act right now. Vulnerability management market disruption: there are half a dozen startups already plotting to disrupt the market, likely to come out of stealth in 2026 Future of the SOC: if it's not AI, what is it? What else??? What am I missing? What would you like to see us discuss? Please drop me a line and let me know: adrian.sanabria@cyberriskalliance.com Topic 2: The state of cybersecurity hiring This topic has been in the works for a while! Ayman had a whole podcast and book focused on all the paths people take to get into security. Jackie worked with WiSys on outlining pathways into a cybersecurity career. Whether you're already in cyber or looking for a way in, this segment crams a lot of great advice into just 15-20 minutes. Segment resources: Ayman's personal guide for getting into security https://www.wicys.org/wp-content/uploads/2025/10/WiCyS-Pathways-in-Cyber-PDF-9.24.25.pdf News Finally, in the enterprise security news, Fundings and acquisitions still strong in 2026! Santa might be done delivering gifts, but not protecting Macs! ClickFix attacks Weaponized Raspberry Pis MongoDB incidents for Christmas Top 10 Cyber attacks of 2025 US gets tough on nation state hackers? Brute force attacks on Banks An AI Vending Machine All that and more, on this episode of Enterprise Security Weekly. Visit https://www.securityweekly.com/esw for all the latest episodes! Show Notes: https://securityweekly.com/esw-441

First Topic - Podcast Content Plans for 2026 Every year, I like to sit down and consider what the podcast should be focusing on. Not doing so ensures every single episode will be about AI and nobody wants that. Least of all, me. If I have one more all-AI episode, my head is going to explode. With that said, most of what we talk about in this segment is AI (picard face palm.png). I think 2026 will be THE defining year for GenAI. Three years after the release of ChatGPT, I think we've hit peak GenAI hype and folks are ready for it to put up or shut up. We'll see winners grow and get acquired and losers pivot to something else. More than anything, I want to interview folks who have actually seen it work at scale, rather than just in a cool demo in a vendor sandbox. Also on the agenda for this year: The battle against infostealers and session hijacking: we didn't have a good answer in 2025. When is it coming? Will it include Macs, despite them not having a traditional TPM? The state of trust in outsourcing and third party use (Cloud, MSSPs, SaaS, contractors): 2025 was not a good year for third parties. Lots of them got breached and caused their customers a lot of pain. Also, there's the state of balkanization between the US and... the rest of the entire world. Everyone outside the US seems to be trying to derisk their companies and systems from the Cloud Act right now. Vulnerability management market disruption: there are half a dozen startups already plotting to disrupt the market, likely to come out of stealth in 2026 Future of the SOC: if it's not AI, what is it? What else??? What am I missing? What would you like to see us discuss? Please drop me a line and let me know: adrian.sanabria@cyberriskalliance.com Topic 2: The state of cybersecurity hiring This topic has been in the works for a while! Ayman had a whole podcast and book focused on all the paths people take to get into security. Jackie worked with WiSys on outlining pathways into a cybersecurity career. Whether you're already in cyber or looking for a way in, this segment crams a lot of great advice into just 15-20 minutes. Segment resources: Ayman's personal guide for getting into security https://www.wicys.org/wp-content/uploads/2025/10/WiCyS-Pathways-in-Cyber-PDF-9.24.25.pdf News Finally, in the enterprise security news, Fundings and acquisitions still strong in 2026! Santa might be done delivering gifts, but not protecting Macs! ClickFix attacks Weaponized Raspberry Pis MongoDB incidents for Christmas Top 10 Cyber attacks of 2025 US gets tough on nation state hackers? Brute force attacks on Banks An AI Vending Machine All that and more, on this episode of Enterprise Security Weekly. Visit https://www.securityweekly.com/esw for all the latest episodes! Show Notes: https://securityweekly.com/esw-441

First Topic - Podcast Content Plans for 2026 Every year, I like to sit down and consider what the podcast should be focusing on. Not doing so ensures every single episode will be about AI and nobody wants that. Least of all, me. If I have one more all-AI episode, my head is going to explode. With that said, most of what we talk about in this segment is AI (picard face palm.png). I think 2026 will be THE defining year for GenAI. Three years after the release of ChatGPT, I think we've hit peak GenAI hype and folks are ready for it to put up or shut up. We'll see winners grow and get acquired and losers pivot to something else. More than anything, I want to interview folks who have actually seen it work at scale, rather than just in a cool demo in a vendor sandbox. Also on the agenda for this year: The battle against infostealers and session hijacking: we didn't have a good answer in 2025. When is it coming? Will it include Macs, despite them not having a traditional TPM? The state of trust in outsourcing and third party use (Cloud, MSSPs, SaaS, contractors): 2025 was not a good year for third parties. Lots of them got breached and caused their customers a lot of pain. Also, there's the state of balkanization between the US and... the rest of the entire world. Everyone outside the US seems to be trying to derisk their companies and systems from the Cloud Act right now. Vulnerability management market disruption: there are half a dozen startups already plotting to disrupt the market, likely to come out of stealth in 2026 Future of the SOC: if it's not AI, what is it? What else??? What am I missing? What would you like to see us discuss? Please drop me a line and let me know: adrian.sanabria@cyberriskalliance.com Topic 2: The state of cybersecurity hiring This topic has been in the works for a while! Ayman had a whole podcast and book focused on all the paths people take to get into security. Jackie worked with WiSys on outlining pathways into a cybersecurity career. Whether you're already in cyber or looking for a way in, this segment crams a lot of great advice into just 15-20 minutes. Segment resources: Ayman's personal guide for getting into security https://www.wicys.org/wp-content/uploads/2025/10/WiCyS-Pathways-in-Cyber-PDF-9.24.25.pdf News Finally, in the enterprise security news, Fundings and acquisitions still strong in 2026! Santa might be done delivering gifts, but not protecting Macs! ClickFix attacks Weaponized Raspberry Pis MongoDB incidents for Christmas Top 10 Cyber attacks of 2025 US gets tough on nation state hackers? Brute force attacks on Banks An AI Vending Machine All that and more, on this episode of Enterprise Security Weekly. Show Notes: https://securityweekly.com/esw-441

First Topic - Podcast Content Plans for 2026 Every year, I like to sit down and consider what the podcast should be focusing on. Not doing so ensures every single episode will be about AI and nobody wants that. Least of all, me. If I have one more all-AI episode, my head is going to explode. With that said, most of what we talk about in this segment is AI (picard face palm.png). I think 2026 will be THE defining year for GenAI. Three years after the release of ChatGPT, I think we've hit peak GenAI hype and folks are ready for it to put up or shut up. We'll see winners grow and get acquired and losers pivot to something else. More than anything, I want to interview folks who have actually seen it work at scale, rather than just in a cool demo in a vendor sandbox. Also on the agenda for this year: The battle against infostealers and session hijacking: we didn't have a good answer in 2025. When is it coming? Will it include Macs, despite them not having a traditional TPM? The state of trust in outsourcing and third party use (Cloud, MSSPs, SaaS, contractors): 2025 was not a good year for third parties. Lots of them got breached and caused their customers a lot of pain. Also, there's the state of balkanization between the US and... the rest of the entire world. Everyone outside the US seems to be trying to derisk their companies and systems from the Cloud Act right now. Vulnerability management market disruption: there are half a dozen startups already plotting to disrupt the market, likely to come out of stealth in 2026 Future of the SOC: if it's not AI, what is it? What else??? What am I missing? What would you like to see us discuss? Please drop me a line and let me know: adrian.sanabria@cyberriskalliance.com Topic 2: The state of cybersecurity hiring This topic has been in the works for a while! Ayman had a whole podcast and book focused on all the paths people take to get into security. Jackie worked with WiSys on outlining pathways into a cybersecurity career. Whether you're already in cyber or looking for a way in, this segment crams a lot of great advice into just 15-20 minutes. Segment resources: Ayman's personal guide for getting into security https://www.wicys.org/wp-content/uploads/2025/10/WiCyS-Pathways-in-Cyber-PDF-9.24.25.pdf News Finally, in the enterprise security news, Fundings and acquisitions still strong in 2026! Santa might be done delivering gifts, but not protecting Macs! ClickFix attacks Weaponized Raspberry Pis MongoDB incidents for Christmas Top 10 Cyber attacks of 2025 US gets tough on nation state hackers? Brute force attacks on Banks An AI Vending Machine All that and more, on this episode of Enterprise Security Weekly. Show Notes: https://securityweekly.com/esw-441

In deze aflevering bespreken we de overname van DigiD-leverancier Solvinity door het Amerikaanse Kyndryl. Wat betekent het als een cruciale schakel in de Nederlandse digitale infrastructuur in Amerikaanse handen komt, en hoe verhouden technische maatregelen zich tot juridische verplichtingen onder de Cloud Act? Daarna duiken we in The Breachies 2025 van de Electronic Frontier Foundation, waarbij we de Mixpanel-breach gebruiken om het probleem van het surveillance capitalism ecosysteem uit te leggen. We sluiten af met een verontrustend Nederlands verhaal: een 17-jarige havo-scholier die geronseld wordt door de pro-Russische hackersgroep NoName057(16) en opgepakt wordt voor spionage. De eerste vervolging onder de nieuwe spionagewet. Bronnen Solvinity overname & DigiD - Follow the Money: https://archive.ph/20251214155809/https://www.ftm.nl/artikelen/amerikanen-hebben-straks-wel-toegang-tot-digid-ondanks-belofte-van-staats-secretaris - Computable: https://www.computable.nl/2025/11/13/verkoop-van-solvinity-aan-kyndryl-valt-slecht-bij-nederlandse-overheid/ - DigiD over Solvinity: https://www.digid.nl/solvinity/ - Solvinity/Kyndryl: https://www.solvinity.com/nl/kyndryl-nederland/ The Breachies 2025 & Mixpanel - EFF Breachies 2025: https://www.eff.org/deeplinks/2025/12/breachies-2025-worst-weirdest-most-impactful-data-breaches-year - TechCrunch Mixpanel breach: https://techcrunch.com/2025/12/02/a-data-breach-at-analytics-giant-mixpanel-leaves-a-lot-of-open-questions/ - Mixpanel statement: https://mixpanel.com/blog/sms-security-incident/ - OpenAI over Mixpanel incident: https://openai.com/index/mixpanel-incident/ Spionerende scholier & NoName057(16) - NRC hoofdartikel: https://archive.ph/oc6Jr - Operation Eastwood: https://www.europol.europa.eu/media-press/newsroom/news/global-operation-targets-noname05716-pro-russian-cybercrime-network

In this episode of Tank Talks, Matt Cohen and John Ruffolo break down a pivotal week for Canada's innovation economy. Microsoft's $7.5 billion investment in Canadian AI and cloud infrastructure sets the stage for a deeper discussion about whether foreign hyperscalers can genuinely support Canadian data and AI sovereignty under U.S. laws like the Cloud Act.John challenges the assumption that scale equals sovereignty, arguing for a more intentional strategy built through government procurement, layered infrastructure, and selective partnerships. The episode also examines Canada's new Quantum Champions program and the funding directed toward companies Anyon Systems, Xanadu, Photonic, and Nord Quantique, questioning whether current capital levels are enough to prevent Canadian breakthroughs from moving south.Layoffs across the consulting industry surface broader shifts in knowledge work, as information becomes increasingly commoditized in the age of AI. Matt and John discuss how trust, execution, and implementation are replacing traditional advisory models as the real sources of value. The episode closes with a collision of crypto and legacy power, as stablecoin issuer Tether pursues a controlling stake in Juventus, raising new questions about regulation, asset backing, and trust.As foreign capital pours in and domestic funding lags, how much control does Canada actually retain?Microsoft's $7.5B Canadian AI Investment & the Sovereignty Question (01:04)Microsoft announces a massive investment to expand AI and cloud infrastructure in Canada. Matt and John unpack why foreign capital is welcome, but claims of “sovereign AI” raise serious concerns under the U.S. Cloud Act and data jurisdiction realities.Sovereign Compute Strategy: Procurement Over Promises (04:39)John outlines how Canada could realistically build sovereign compute capacity by breaking the stack into layers, using government procurement to back domestic players, and making intentional choices about allies, chips, and infrastructure.Canada's Quantum Champions Program: A Signal or a Solution? (07:49)The federal government commits funding to four Canadian quantum startups, including Xanadu. The discussion explores whether milestone-based funding is enough or if Canada risks losing its quantum leaders to U.S. capital markets again.Why Canadian Capital Isn't Backing Its Winners (09:04)Xanadu's SPAC decision becomes a case study in Canada's capital formation problem. John explains why strong companies still struggle to raise meaningful domestic capital and what that means for long-term value creation.Consulting Firms Face Layoffs as Demand Shifts (11:36)McKinsey and other professional services firms prepare for significant job cuts. Matt and John discuss overhiring during COVID, slowing demand, and how AI is compressing the value of information-based consulting.The End of the Traditional Consulting Pyramid (14:07)AI-driven efficiency challenges the apprenticeship model. The conversation explores why implementation and trust now matter more than slide decks and why junior-heavy consulting structures may no longer work.Forward-Deployed Engineers & New Service Models (16:17)From Palantir's FDE approach to new AI-enabled services firms, Matt highlights how execution-first models are eroding traditional consulting margins and reshaping enterprise problem-solving.Crypto Meets European Dynasties: Tether & Juventus (19:00)Tether's attempted acquisition of Juventus sparks debate around stablecoin backing, asset quality, and trust. John questions whether a treasury-backed stablecoin should ever be tied to assets like football clubs.Trust as the Core Currency of the AI Era (21:03)The episode closes with a clear takeaway: information is cheap, execution is hard, and trust is everything, from sovereign infrastructure to consulting, investing, and crypto.Connect with John Ruffolo on LinkedIn: https://ca.linkedin.com/in/joruffoloConnect with Matt Cohen on LinkedIn: https://ca.linkedin.com/in/matt-cohen1Visit the Ripple Ventures website: https://www.rippleventures.com/ This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tanktalks.substack.com

C'est une révélation qui risque de refroidir les ardeurs des partisans du cloud « souverain ». En Allemagne, un rapport juridique commandé par le ministère de l'Intérieur, longtemps resté confidentiel, vient d'être rendu public grâce à une demande d'accès à l'information. Et son constat est sans appel : les lois américaines permettent bel et bien aux agences de renseignement des États-Unis d'accéder à des données hébergées en Europe. Pour y voir clair, Berlin avait missionné des juristes de l'Université de Cologne. Leur question était simple, mais explosive : jusqu'où s'étend réellement le pouvoir des autorités américaines sur les données stockées hors de leur territoire ? La réponse tient en quelques textes bien connus à Washington : le Stored Communications Act, renforcé par le Cloud Act, et surtout la section 702 du Foreign Intelligence Surveillance Act, prolongée par le Congrès jusqu'en 2026 au moins. Ensemble, ces lois offrent une portée extraterritoriale massive aux services américains. Le point clé est juridique, pas géographique. Peu importe que vos données soient stockées à Francfort, Dublin ou Paris. Ce qui compte, c'est qui contrôle l'infrastructure. Si la maison mère d'un fournisseur cloud est basée aux États-Unis, elle peut être contrainte de transmettre des données, même si celles-ci sont hébergées par une filiale européenne. Et la zone grise va plus loin encore : selon les experts cités par Heise Online, même certaines entreprises européennes peuvent être concernées dès lors qu'elles entretiennent des relations commerciales substantielles avec les États-Unis. On pourrait croire que le chiffrement règle le problème. Là encore, le rapport tempère. Le droit américain impose aux entreprises de préserver l'accès aux données jugées pertinentes dans le cadre d'enquêtes potentielles. Un fournisseur cloud qui se rendrait techniquement incapable d'y accéder s'exposerait à de lourdes sanctions. Résultat : un conflit frontal entre le RGPD européen, qui limite les transferts vers des pays tiers, et l'extraterritorialité revendiquée par Washington. Le Data Privacy Framework, censé servir de pont entre les deux blocs, apparaît plus fragile que jamais.Cette situation touche directement les géants américains du cloud, mais le cas de Microsoft 365, omniprésent dans les administrations et les entreprises européennes, cristallise les inquiétudes. Certains juristes estiment qu'un usage compatible avec le RGPD reste possible, à condition de mener des évaluations d'impact très poussées. D'autres jugent cette approche illusoire. Pour des acteurs comme Nextcloud, le diagnostic est clair : audits et clauses contractuelles ne suffisent plus. L'Europe doit investir massivement dans ses propres infrastructures, miser sur l'open source et développer des technologies réellement autonomes. Car une chose est désormais évidente : héberger des données en Europe ne garantit plus leur protection. Hébergé par Acast. Visitez acast.com/privacy pour plus d'informations.

La France accélère dans la course aux infrastructures numériques. Vendredi 28 novembre 2025, le ministère de l'Économie a dévoilé un véritable mode d'emploi pour implanter des centres de données sur le territoire. Objectif affiché : structurer, encadrer et surtout accélérer un secteur devenu stratégique. Aux côtés de Roland Lescure, la ministre déléguée au Numérique, Anne Le Hénanff, a voulu clarifier les règles du jeu. Le signal est fort : 63 sites sont déjà identifiés et plus de 109 milliards d'euros d'investissements annoncés. L'ambition est claire : faire de la France un hub européen majeur du data.Derrière chaque clic, chaque photo stockée, chaque requête d'intelligence artificielle, se cachent des hangars de serveurs fonctionnant jour et nuit. Le guide distingue désormais les petites salles informatiques, les centres de colocation géants, les infrastructures cloud ou encore les datacenters dédiés au calcul intensif pour l'IA. Si l'État s'en mêle autant, c'est pour une raison essentielle : la souveraineté. Héberger les données sensibles sur le sol national permet de se prémunir contre des lois étrangères à portée extraterritoriale, comme le Cloud Act, et de soutenir l'essor des start-up françaises de l'IA, fortement dépendantes de la puissance de calcul. Le déclic est venu du sommet parisien sur l'IA, en février. Une task force réunissant la Direction générale des Entreprises, Business France et RTE a passé le territoire au crible. Résultat : des sites repérés en Normandie, en Occitanie ou dans le Grand Est, afin de désengorger l'Île-de-France, déjà sous forte tension électrique.Car le nerf de la guerre reste l'électricité. Un datacenter moyen consomme entre 50 et 250 mégawatts, l'équivalent d'une petite ville. Le raccordement peut prendre jusqu'à neuf ans. À cela s'ajoute la connectivité, avec la nécessité d'un accès à la fibre et aux hubs de peering, comme Marseille, devenue un carrefour numérique grâce à ses câbles sous-marins. Le volet environnemental n'est pas oublié. Le texte fixe des exigences sur l'eau, le refroidissement et l'efficacité énergétique, avec un objectif de PUE à 1,20 d'ici 2030. La récupération de la chaleur devient également obligatoire pour les grandes installations. Enfin, l'argument économique reste décisif : un datacenter de 100 mégawatts crée une cinquantaine d'emplois directs et apporte des recettes fiscales significatives. Hébergé par Acast. Visitez acast.com/privacy pour plus d'informations.

Thierry Weber revient sur un virage majeur en Suisse : l'armée refuse désormais de stocker ses données sensibles chez Microsoft, invoquant les risques liés à l'infonuagique soumis au droit américain. Cette décision, motivée notamment par le Cloud Act, relance un débat qui dépasse le simple choix d'un fournisseur. Thierry Weber décrit une tension croissante entre efficacité technologique et indépendance nationale, alors que plusieurs pays repensent leurs infrastructures numériques à l'heure des cyberattaques et des ingérences potentielles.

En ,retard sur l'intelligence artificielle, Apple miserait sur Google Gemini pour muscler son assistant Siri. La Chine bannit les puces Nvidia. Microsoft promet un Copilot localisé pour rassurer sur la confidentialité des données.Avec Bruno Guglielminetti (Mon Carnet)Apple mise sur Google pour réinventer SiriApple aurait tranché : plutôt que de tout développer en interne, la firme californienne s'apprêterait à intégrer des modèles d'intelligence artificielle développés par Google dans son assistant vocal Siri. Selon plusieurs fuites concordantes, il s'agirait du modèle Gemini, avec ses 1 200 milliards de paramètres, le tout hébergé sur les serveurs Apple pour préserver la confidentialité des données. Un choix stratégique, signe d'un certain aveu de faiblesse sur l'IA, mais aussi d'un réalisme technologique.Ce partenariat inédit pose aussi la question de la différenciation : comment Siri saura-t-il se démarquer de l'expérience Pixel, propulsée par le même moteur IA ? Réponse attendue dans les prochaines versions d'iOS.Pékin boute Nvidia hors de Chine et crée ses propres microprocesseurs IALa Chine franchit un nouveau cap dans sa stratégie d'indépendance technologique. Pékin a officiellement interdit l'usage des puces IA étrangères dans ses centres de données publics. Nvidia — jusqu'ici très présent sur le marché chinois — est directement visé.Cette décision s'inscrit dans un mouvement entamé depuis plusieurs années : après avoir été privé des technologies américaines, Huawei a réussi à rebondir avec ses propres solutions. Le pays entend désormais faire de même avec les puces IA, en s'appuyant sur des acteurs comme Cambricon, Enflame ou encore Alibaba Cloud. Même le patron de Nvidia, Jensen Huang, reconnaît : « la Chine va gagner la course à l'IA ».Microsoft Copilot veut rassurer sur la souveraineté des donnéesMicrosoft promet que, d'ici fin 2026, son assistant Copilot de Microsoft 365 traitera les requêtes localement dans 15 pays, dont la France, le Canada et l'Allemagne. Une annonce destinée à rassurer les utilisateurs face aux enjeux de souveraineté numérique.Mais dans les faits, les données resteront soumises au Cloud Act, cette loi américaine qui autorise les autorités à accéder aux serveurs des entreprises US, même à l'étranger. En France, le sujet est particulièrement sensible, et le concept de “cloud souverain” a d'ailleurs été discrètement remplacé par celui de cloud de confiance.Au Canada, un budget national tourné vers l'infonuagiqueAu Canada, le nouveau budget fédéral 2025 prévoit d'importants investissements dans l'infrastructure numérique, avec des data centers locaux et une IA “made in Canada”. Une réponse directe aux enjeux géopolitiques et à la dépendance vis-à-vis des géants technologiques américains. Pendant ce temps, en France, l'IA reste (hélas) largement absente du débat budgétaire.-----------♥️ Soutien : https://mondenumerique.info/don

I remember the court case years ago when the US government wanted to access data in Azure that was physically stored in Ireland. I wrote lightly about this and linked to the article back in 2020. This has typically been more of a concern for the EU (and other countries) than the US, but I'm sure there are organizations in the US that use the cloud and don't want their data accessed by other countries' governments. Recently, a Microsoft executive was asked about this in the French Senate. The Microsoft response was that they  (Microsoft) cannot guarantee data sovereignty for French customers. If the US government served a warrant under the Cloud Act, a US corporation would have to turn over the data. Read the rest of Data Sovereignty in the Cloud

Arash Attar-Rezvani - M&A Partner, Skadden, Arps, Slate, Meagher & Flom LLP Arash Attar-Rezvani, M&A Partner at Skadden based in Paris, brings over two decades of cross-border M&A strategy experience to this in-depth conversation. From billion-dollar telecom deals across Latin America to luxury brand acquisitions spanning multiple jurisdictions, Arash reveals the hidden complexities that make international M&A uniquely challenging. M&A professionals will learn how to structure deals across incompatible legal systems, navigate emerging regulatory landscapes, and build the trust essential for successful cross-border transactions. Things you will learn: How to identify and manage multiple antitrust and national security clearances across jurisdictions with varying sophistication levels Why smaller transactions often require more innovation than billion-dollar deals, and how to build structures when no legal playbook exists The psychology behind cross-border deal-making and why trust trumps even the most ironclad contracts _________________ How One Small M&A Team is Closing 8 Deals This Year See how US Heart & Vascular is running faster, cleaner deals using Buyer-Led M&A™ and DealRoom. Join Kison in the live session on August 14 at 11am EST.

AWS Morning Brief for the week of July 28th, 2025, with Corey Quinn. Links:Launching Amazon CloudWatch generative AI observability (Preview) Amazon CloudWatch adds IPv6 supportBoost cold-start recommendations with vLLM on AWS Trainium AWS Private CA now supports issuing up to 100 million certificates per CA Amazon Connect announces per-day pricing for external voice connectors Amazon RDS for Db2 adds support for group-based authorization with self-managed Active Directory Manage multi-tenant Amazon Bedrock costs using application inference profilesSimplify serverless development with console to IDE and remote debugging for AWS Lambda | AWS News Blog AWS Generative AI for Developers Professional Certificate Simplify AWS Organization Tag Policies using new wildcard statement  Security Update for Amazon Q Developer Extension for Visual Studio Code (Version #1.84) Cost Optimization Hub now supports account names in optimization opportunities Year One of Valkey: Open-Source Innovations and ElastiCache version 8.1 for Valkey - go listen to the audio version of this newsletter specifically for this item. AWS Security Incident Response: The customer's journey to accelerating the incident response lifecycle AWS Service Reference Information now supports actions for last accessed services - Five facts about how the CLOUD Act actually works | AWS Security Blog Bob's Used Books: Build a .NET Serverless Application on AWS, Part 1: Deployment and Setup Amazon EC2 now supports skipping the operating system shutdown when stopping or terminating instances New whitepaper available: AICPA SOC 2 Compliance Guide on AWS Why 2025 is the Inflection Point for AWS Cloud Migration Beyond IAM access keys: Modern authentication approaches for AWS Introducing SRA Verify – an AWS Security Reference Architecture assessment tool Supercharging Ad Creative with Amazon Bedrock and Amazon Nova: How AI is Revolutionizing Content Generation for Advertising & Marketing Use-CasesBuilding resilient multi-tenant systems with Amazon SQS fair queues How Truth For Life transformed its viewer analytics while optimizing costs

The EU cloud exit movement is reshaping how European organizations think about data storage and sovereignty. Companies across Europe are moving away from US-based cloud providers like Microsoft 365, AWS, and Google Workspace due to concerns about the Cloud Act and data privacy regulations.In this episode, Curtis and Prasanna explore the backup implications of this major shift. They discuss the challenges of replacing comprehensive platforms like Microsoft 365 with multiple EU-based providers, the complexities of bringing services back in-house, and why the 3-2-1 backup rule becomes even more critical during these transitions.Whether organizations choose local providers or decide to self-host their infrastructure, data protection remains paramount. The hosts share real-world examples of failed backup strategies, including the Rackspace Exchange disaster and OVH's data center fire, to illustrate why third-party backup solutions are necessary regardless of your hosting choice.

LdN418 Staatsstreich in den USA?, Trump verhandelt mit Putin über Ukraine (Interview Claudia Major, Politikwissenschaftlerin), TV-Duell, Bundestagswahl nach neuen Regeln, Korruption in Deutschland, CLOUD Act und Datenschutz, Bürokratieabbau in der Praxis (Interview Heidrun Hausen, DELO)