Podcasts about graphs

In today's Cloud Wars AI Minute, I look at how Microsoft is using Work IQ to make AI agents faster, more efficient, and potentially less expensive to operate.Highlights00:10 — Today, the topic is going to be Work IQ, and we're actually starting to see Microsoft make very interesting moves inside of Work IQ. We've already seen AI agents doing a lot of really interesting things, specifically around the Graph API. But the big difference here is Microsoft is really looking to monetize the ability for them to be able to do preloaded context, and so this allows us to be able to optimize our token spend.00:47 — It's already pre-calculated and just provided to you. And what we're starting to see is that, with Work IQ, you can get about an 80% reduction in the amount of tokens that it would require for you to go derive the same sort of information out of your Graph implementation of your data.01:10 — This is really powerful for organizations who are looking to optimize their token spend. It also means that there's going to be a lot of power coming into Microsoft around the ability for them to monetize this in a way that could be quite effective.01:57 — This is going to be a very interesting world as Microsoft leans into the ability to monetize all of this context that it has from its Graph. The other thing that's a little interesting about it as well is that it didn't offer it as yet another SKU. Visit Cloud Wars for more.

This show has been flagged as Clean by the host. -------------------- 01 Introduction This is the fourth episode in an 8 part series. 02 In the previous episode we looked at the Allen Bradley PLC2, one of the very early successful Programmable Logic Controllers. In this episode we will look at another early PLC, the Siemens S5 series, and use it as an example of both how PLCs became more sophisticated, as well as an example of a different approach to the overall software architecture. Again, I will only touch on this topic lightly, this is not a course on how to program PLCs. However, I will give you enough technical detail to give you are rough idea of what they were like. -------------------- 03 S5 Historical Background Siemens are the largest PLC vendor on a global basis, and have been since the early days. Their first PLC was the S3, released in 1975, two years before Allen Bradley released the PLC2. There seems to be almost no information available on the S3, so I can't say much about it. In 1978 or 1979, sources are unclear about the exact date, Siemens released the S5 series. This was a huge step forward in capability and was the foundation of their PLC product line until replaced by the S7 series in the mid 1990s. In this episode I will focus on the U series, which were an upgrade to the original S5. -------------------- 04 A Complete Family of PLCs Fully developed, the Siemens S5 series was a complete family of products covering the entire size spectrum from smallest to largest. Siemens covered the entire industrial control market, and there was almost nothing they didn't have somewhere in their catalogue. You could spend an entire career using nothing but Siemens products for everything. 05 At the lowest end were the S5-100 series, consisting of the S5-100, 102, 103, 90, and 95. These covered everything from the small "shoebox" all in one form factor to ones which overlapped the mid range in terms of capability except for their compact I/O cards which were smaller and slower than the full size models. The S5-115 formed the mid range, coming in a "full size" form factor in terms of packaging and rack size. The S5-135 and S5-155 extended the S5-115 in terms of speed and memory. 06 In this episode I will focus on the lower end S5-100 series for the sake of simplicity, and will not make any reference to any additional capabilities of the larger models. It will be much too time consuming and confusing to try to cover everything, and there wouldn't be much point to it. -------------------- 07 Data Memory Types If you recall the previous episode, the Allen Bradley PLC2 had a data table, or memory system, that was a single linear range of memory with all I/O and internal capabilities mapped into fixed numerical addresses. Allen Bradley were to abandon this approach in their later PLC5 series introduced in the mid to late 1980s, but we won't cover that here. The Siemens S5 series however took the approach of having different types of memory addresses for different purposes. I will describe those for the S5-100 series now. 08 Input and Output Addresses The standard I/O modules for these compact PLCs had 8 inputs or 8 outputs per I/O module. These module plugged into a flat bus (as opposed to an enclosed rack) with slots numbered from 0 through 31 in decimal. Each of these slots had an associated byte number, and each I/O point on the individual modules had a bit number. 09 Input and output addresses had the following format. Letter , number, decimal point, number. The letter was either I or Q. I indicated inputs. Q indicated outputs. The first number was the byte number, which corresponded to the physical I/O module slot. The second number was the bit number within the byte, starting from zero. 10 So "I2.3" indicated an input module located in the third slot (numbered from zero) and addressed the fourth input (numbered from zero) on that module. Outputs worked the same way. For example "Q 5.1" 11 Analogue modules could be located slots 0 through 7. Analogue if you recall, refers to voltages which have varying levels rather than just on or off, and can represent things like temperature. Each analogue module was assigned 8 bytes per slot, starting at address 64 and going up to 127. In practice many of these so called "analogue" modules had nothing to do with actual analogue voltages, but rather any advanced module which needed more address space was used here. Even later high density digital I/O used these addresses. 12 Each of these addresses could be addressed as bytes or words. In this case input bytes were indicated by an "IB" prefix, and input words were indicated by a "IW" prefix. Output bytes used a "QB" prefix and output words used a "QW" prefix. 13 Flags Flags are individual bits of internal memory that are used for storing intermediate logic values. These began with an "F" prefix. An S5-103 had a total of 2048, numbered from F0.0 to F255.7. 14 Counters Counters started with a "C" prefix. An S5-103 had 128 of these numbered from C0 to C127. 15 Timers Timers started with a "T" prefix. An S5-103 had 128 of these numbered from T0 to T127. 16 Data Blocks Programs needing to deal with byte and word memory could create what were known as "Data Blocks". These were blocks of PLC memory that could be created as the discretion of the programmer. An S5-103 could have a maximum of 254 data blocks numbered from DB2 to DB255. Two additional data blocks were reserved for the PLC's internal operations. Data blocks could contain up to 255 16 bit words. 17 However, the actual number and size of data blocks would in practice be limited by available memory. Even 20 kilo bytes was considered to be a large memory, and this had to be shared with the program. A data block would be "called" to make it the current data block. Then individual words would be addressed with the "DW" prefix. For example "DW5". -------------------- 18 Program Blocks If you recall with the PLC2, there was a "main" program and a series of optional numbered subroutines. In the S5, programs were split into "blocks". These were Organization Blocks Program Blocks Function Blocks Sequence Blocks An S5-103 could have up to 256 of each of these blocks. 19 Organization Blocks When the PLC started up, it would look for Organization Block 1 and begin execution there. You could consider this to be the equivalent of the "main" function in a C program. Organization blocks had an OB prefix. For example OB1. With a very simple program in a very small S5 PLC, all of the program logic could be simply put in OB1. However, with larger programs , it was convention to put the logic in subroutines and simply have OB1 call each subroutine. 20 Program Blocks These had a PB prefix. For example PB10. Normal ladder logic style programming would go in these. It was convention to split up a program into pieces and put piece in its own program block. Normal practice was to have a program block correspond to a particular part of the machine. 21 Function Blocks These had an FB prefix. For example FB64. Function blocks differed from program blocks in that they could take parameters. This allowed you to write an FB that performed some sort of complex data transforms and re-use it by calling it with different parameters. Some instructions could only be used in FBs, not in other types of blocks. There were also built in FBs in some CPU models which performed certain operations such as converting BCD (or Binary Coded Decimal) to normal integer. 22 Sequence Blocks These has an SB prefix. For example, SB25. These were like PBs, but were intended for controlling sequences of operations where each SB performed as single step of the sequence and multiple blocks were used for a sequence. -------------------- 23 Programming Languages The S5 series offered four different programming languages. These were Ladder (abbreviated as LAD) Statement List (abbreviated as STL) Control System Flowchart (abbreviated as SCF) GRAPH 5 24 You should be familiar with ladder if you listened to previous episodes. Briefly however, it is a graphical programming language which followed the appearance of electrical ladder wiring diagrams which were familiar to engineers, technicians and electricians working with manufacturing equipment. 25 Statement List is a text based programming language which resembles an assembly language for an abstract processor architecture. Boolean logic results were calculated and saved on a stack, while integer values were operated on directly from memory or in two accumulators, ACCU1 and ACCU 2. 26 Control System Flowchart is a graphical language that takes the form of rectangular blocks with one or more inputs and an output. These blocks are laid out on a screen and then wired together rather like logic gates in a electronic schematic. 27 GRAPH 5 is Siemens' implementation of the GRAFCET (Graphe Fonctionnel de Commande Étapes Transitions) IEC 60848 flow chart standard. The IEC are the International Electrotechnical Commission, an international standards organization. 28 I will focus on Ladder and Statement List, as these were by far the most commonly used and we have limited time to discuss the S5. -------------------- 29 Statement List As mentioned above, Statement List can be thought of as the assembly language for an abstract CPU architecture. The firmware in the S5 PLC will run the instructions in an interpreter. 30 The RLO Boolean or bit oriented instructions operate on a logic stack known as the RLO, or Result of Logic Operation. Boolean operations typically take one operand. When a boolean operation is executed, if the RLO is empty it simply loads the operand onto the RLO. If the RLO is not empty, then the instruction is executed using the operand and the RLO, and the result replaces the value on the top of the RLO stack. Some boolean instructions do not take an operand, but rather work on the two topmost values on the RLO. If you are familiar with languages such as FORTH, this should sound very familiar. 31 The Accumulators Word oriented instructions work on two accumulators, ACCU 1 and ACCU 2. A load instruction will load a word or byte value from a memory location or a constant into ACCU 1. The value that was in ACCU 1 then gets shifted into ACCU 2. A transfer instruction will write the value in ACCU 1 to memory. The remaining word oriented instructions will operate on the values in ACCU 1 and ACCU 2, and place their result in ACCU 1, overwriting the existing contents. So for example, an add instruction will add the value in ACCU 2 to the value in ACCU 1, leaving the sum in ACCU 1. 32 Ladder As mentioned above, ladder is a graphical programming language which has the appearance of an electrical control ladder diagram as used with actual relays. The is by far the most common programming language used with PLCs. A ladder program however is just another way of presenting statement list. 33 To put it another way, a statement list program can be displayed on a computer screen as either lines of text, or a graphical symbols laid out in a meaningful pattern. All valid ladder programs can be represented as statement list. However, not all valid statement list sequences can be represented as ladder. Statement list is inherently more flexible than ladder, which means that you simply can't always translate both ways fully. 34 None the less, most typical control logic can be represented as ladder, and ladder is the preferred representation as it is quicker to understand at a glance when debugging a software or hardware problem. On the other hand, some statement list instructions have no ladder equivalent, or the algorithm may not lend itself to being represented in ladder. In these cases normal practice is to put the simple control logic in ladder in Program Blocks, and put the non-translatable statement list logic in Function Blocks. 35 Control System Flowchart and GRAPH 5 Like ladder, CSF and GRAPH 5 are actually implemented in statement list, and many of the considerations mentioned above apply to those as well. However, I won't go into any further detail on these here. -------------------- 36 Programming Software and Hardware There were two ways of entering a program into an S5 PLC. One way was the hand held programming terminal. The other was via the STEP 5 programming software. 37 Hand Held Terminal Like many if not most PLC vendors, Siemens sold a hand held programming terminal. There were a series of these, an example being the PG 605. This was a box like a large calculator which had buttons and a small display. The buttons corresponded to various instructions and a numeric keypad, and the display would show the current instruction being entered. To use it you would plug a cable attached to the hand held terminal into the programming port on the front of the PLC CPU module. You could then enter or view the program, one instruction at a time. 38 STEP 5 Programming Software The other way of programming was to use software running on a PC. Siemens called their software STEP 5. This was an integrated editing and debugging program. Originally it ran on CP/M. 39 Siemens sold their own portable PCs with the software preloaded, and with additional built in hardware features such as an EPROM burner. These computers were also known as "PGs". STEP 5 was later ported to MS-DOS by running it on a CP/M emulator which people would use to run it on an off the shelf PC such as a laptop. Later still a native MS-DOS port was produced. 40 This had modes for Statement List. Ladder, and Control System Flowchart. GRAPH 5 was a separate program. You could toggle between STL, LAD, and CSF modes, and if the software could translate what was there form one mode to another, it would. 41 The PC would be connected to the PLC by a cable which had an interface box in the middle. This could be used to download a program to the PLC, upload a program to take a backup, modify a running program, or debug a running program. -------------------- 42 Instructions I won't list all the instructions, as that would take too long. I will instead give a brief overview of them. I will deliberately skip over obscure instructions that would take too long to explain. 43 Basic Operations Boolean Operations There is a full set of boolean operations including and, and not, or, or not, set, reset, and assign. The assign operation would assign the RLO value to the address of the specified operand. 44 Load and Transfer The Load instruction would load the value of a memory address word or byte or a constant into accumulator 1. There are about 2 dozen variations on the load instruction. The transfer operation would write a word or byte from accumulator 1 into memory. There are about a dozen variations of the transfer operation. 45 Timers and Counters There are 5 different types of timers. Timers take the value in accumulator 1 as their preset when started. There are up counters and down counters, which count up and down respectively. Again, these take the value in accumulator 1 as the preset. 46 Arithmetic Operations The only native arithmetic operations in the S5-100U series are addition and subtraction. These add or subtract the values in accumulator 1 and 2. Another set of add instructions allows adding a constant. However, the 103 and 95 CPUs offered integer multiple and divide as integrated function blocks, FB242 and FB243. You would call these function blocks with the desired parameters and they would return their results in other parameters. 47 Comparison Operations There are a complete set of integer comparison operations which operate on the values in the two accumulators. These include greater than, greater than or equal to, etc. The boolean result is saved on the RLO. 48 Block Call and Return There is a complete set of unconditional and conditional call operations to call all of the various types of blocks. There are also unconditional and conditional return operations. Calling a data block makes it the current data block, for load and transfer operations on data words 49 NOP or No Op As implied these instructions have no result. 50 Stop This stops the program executing 51 Binary Word Operations These operate on words in the accumulators. These include AND, OR, XOR, Shift Left, and Shift Right. 52 Bit Operations These test, set, or reset individual bits in data words or in timers and counters. 53 Conversion Instructions There are instructions to perform one's complement or two's complement of the value in accumulator 1. 54 Jump Operations There are unconditional and conditional operations to jump forward or back to a label. This is equivalent to a GOTO operation. 55 Indirect Addressing or "DO" operations These instructions allow jumping to a block or selecting a data word or flag word according to a number stored in another address. Essentially, these instructions are like pointers. 56 Miscellaneous There are also various other instructions that do things like copy blocks of words, or swap the contents of the accumulators, 57 Other Instructions There are still more instructions which I have skipped over as it would take too much time to explain them. 58 Instruction Summary As you can see the S5 PLC offers a very comprehensive set of operations which go well beyond what most people would think of as what a PLC does. This is a very powerful instruction set. 59 On the one hand this is a good thing, as you can do nearly anything. On the other hand, this is a bad thing, as some people take this as license to do nearly anything. In the hands of the wrong person, this results in a complex mass of spaghetti code that not even the person who wrote it can figure out later. 60 This goes against the philosophy of PLC programming, which is to make things as simple and obvious as possible. With a PLC program, perfection is achieved not when it has as many advanced features as possible, but rather when it is not possible to make it any simpler. 61 While Siemens gave the programmer a range of powerful low level instructions from which he could construct nearly anything, many other vendors took a different approach. They instead studied what was needed in major market segments and provided the programmer with instructions that operated at a higher level of abstraction, allowing programs to be written in less time and with fewer instructions. -------------------- 62 Execution Speed Execution speed of instructions varied depending on the PLC CPU model. Here are some typical values for the low end S5-100 series CPUs. 63 A boolean AND instruction executed in 70 microseconds on a S5-100U, or in 4 microseconds on an S5-102U, or in 1.6 microseconds on an S5-103U. 64 Adding two words in the accumulators executed in 55 microseconds on a S5-100U, or in 23 microseconds on an S5-102U, or in 1.6 microseconds on an S5-103U. 65 The S5-103U had a custom logic coprocessor chip, known as the M5. Instructions that were executed on the M5 coprocessor tended to take a uniform 1.6 microseconds. However, complex instructions that were not executed on the M5 coprocessor were often slower than on CPUs that didn't have a coprocessor, such as the 100 or 102. I suspect this is due to some sort of overhead relating to coordination of the two processors. 66 However, most instructions in a normal program are boolean logic, so very fast boolean execution time is what matters. Since every instruction is executed every scan cycle, large programs require faster instruction execution in order to keep up with demand. -------------------- 67 The S7 Series Replaces the S5 In the mid 1990s, Siemens began introducing the S7 series which would ultimately replace the S5, although that took a number of years to happen. 68 The hardware was all new, but from a programmer's perspective, the S7 was an evolution of the S5. Some statement list instructions were dropped and others were added. Most of the basic concepts however remained the same. Porting an S5 program to an S7 was a complete re-write, but it was a fairly straightforward one. 69 The exception to this was the S7-200 series, which came into Siemens from the outside as part of their take over of of the PLC division of Texas Instruments. The S7-200 was an American knock off of a Japanese PLC and sold as a German product. It was none the less an excellent PLC and sold in very large numbers. The S7-200 was eventually replaced by something that was compatible with the rest of the S7 line 70 The S7 series has gone through a number of hardware revisions and remains Siemens' main PLC line today. It can via the S5 trace its lineage back to the 1970s and the early days of PLCs. -------------------- 71 Other Developments Japanese PLC Vendors I haven't covered any Japanese PLC vendors in the same way that I covered the PLC2 or S5. I had planned on doing another episode on one but I don't think it would really add much to what we have already discussed. However, PLCs from Japan are among the top sellers world wide and are generally very highly regarded. 72 Pace of Innovation In terms of general history across the industry, in the early days of PLCs evolution was fairly rapid. Later on, innovation slowed down as compatibility with the existing install base became a bigger factor. Technology has become fairly static, but that isn't necessarily a bad thing from a business perspective. I will return to this subject later however. -------------------- 73 Conclusion In this episode we covered the Siemens S5 series of PLCs. We covered the history of the S5, it memory and addressing, the different types of subroutines, programming languages, the instructions, and how it was succeeded by the S7 series. 74 We saw how PLCs had rapidly evolved from their very early relay-replacement origins to become powerful programmable devices. In the next episode we will look at I/O modules and how they allow a PLC CPU to interface to hardware and control machines. 75 This has been the fourth episode in an 8 part series. -------------------- Provide feedback on this episode.

Send us Fan MailRecorded on 30 September 2026, this episode of "What's New in Cloud FinOps" brings together Frank Contrepois and Stephen Old for a comprehensive roundup of the latest developments across AWS, Google Cloud, Oracle Cloud Infrastructure (OCI), and Alibaba Cloud. With a significant focus on the rapidly evolving world of Generative AI cost management, sustainability, and cloud governance, this instalment was recorded live from the GreenIO Conference foyer in London. Expect a lively, candid flow with real-time analysis.Generative AI & Machine Learning Cost ControlThe episode kicks off with major updates for Amazon Bedrock, SageMaker, and other cloud AI offerings.Amazon Bedrock Agent Core Payments (GA): A key discussion point is that agents can now autonomously discover, access, and pay for paid APIs and content. This presents a major governance challenge, risking unmanaged spend and fragmented cost visibility. The hosts stress the need for spending caps, approval workflows, per-agent budgets, and unified cost aggregation.Amazon Bedrock Cost Allocation & Anomaly Detection: Bedrock now supports IAM principal cost allocation, allowing organisations to track usage back to specific users and roles in the CUR. Additionally, AWS Cost Anomaly Detection now covers third-party Foundation Models like Anthropic's Claude.Bedrock Pricing (OpenAI GPT-5.6): Promotional pricing is available at $4/million input tokens and $20/million output tokens (a claimed 20-33% reduction), valid through at least 21 November 2026.Amazon SageMaker: A new guided, low-code/no-code path in SageMaker Studio simplifies finding cost-effective inference configurations for custom models.Google Cloud BigQuery: Introduced preview monitoring for data agent performance, latency, and conversation costs through Google Cloud observability.Oracle Cloud Infrastructure (OCI): August's AI update expanded hardware choices and on-demand inference for Cohere and Meta models.Calculating AI ROI: Frank highlights an AWS article by Adam Richter on presenting AI return on investment. Steve recalls a previous podcast with Adam discussing a Cornell "levelised cost of AI" study, emphasising the need for measurable value.From AI Pilots to Production: The hosts discuss how many AI pilots run too long without a clear "exit to value" plan. Frank's workshops focus on defining proof-of-value criteria and graduation paths to production. Research shows 59% of organisations report increased AI spend wastage, while only 31% have accurate visibility.Commitments, Pricing Models & The New "SAM" ChallengeA clear trend is the application of traditional cloud commitment models to AI services, creating new challenges where FinOps and Software Asset Management (SAM) collide.Google Cloud: Introduced flexible, spend-based Savings Plans for Gemini Enterprise (10% for one year, 20% for three years) and a forthcoming "deferred execution" option offering up to a 50% discount for non-urgent jobs.Alibaba Cloud: Launched AI Savings Plans for Model Studio on 28 August 2026, with discounts up to 47%. They also introduced batch inference pricing at a 50% discount.The FinOps/SAM Collision: Models like Gemini Enterprise and Azure Copilot Studio blend per-user subscriptions with pay-as-you-go agent workloads. This creates a challenge in determining who owns which piece of the cost and how credits and commitments apply across teams.Hidden Risks & Governance GapsThird-Party AI EULAs: Stephen highlights a critical risk: when a user accepts the End-User License Agreement (EULA) for a third-party model (e.g., Anthropic on Bedrock), they could be binding their entire organisation to the provider's standard terms without legal or procurement oversight.Intellectual Property Risk: Frank points to a recent HBR article discussing how the "how" of AI usage—the prompts and methods—may not be protected, potentially exposing valuable IP.Cloud Infrastructure & Cost Management UpdatesAWS Compute:R9g/R9gd instances (Graviton 4): AWS claims up to 25% better compute, but Frank's analysis shows an 8-25% price uplift. FinOps takeaway: benchmark workloads before migrating.ECS fractional GPU scheduling: ECS now supports fractional GPU allocations on EC2 g6f, ideal for right-sizing smaller ML/AI workloads and improving utilisation.ElastiCache on Graviton 4: Offers ~40% higher throughput and ~31% better price-performance, but with higher memory costs.AWS Database & Analytics:Aurora Serverless: Can now scale to 12 ACU in ~1 second, reducing the need for high steady baselines, but requiring guardrails to prevent cost spikes.RDS for SQL Server BYOM: "Bring Your Own Media" is extended to more regions, requiring careful audit documentation.Oracle AI Database on AWS Exascale: Oracle Exadata DB Services now generally available via AWS, signalling evolving multi-cloud coexistence.AWS Glue 6.0 GA: A ~30% price reduction and runtime upgrades (Spark 4.1, Python 3.13) offer immediate savings opportunities.AWS Glue Data Quality: Anomaly detection is now free with an improved "observation mode".Amazon Kinesis: Can now deliver data directly to S3 in Iceberg format, potentially lowering delivery and query costs.AWS Billing & Lambda:AWS Data Exports: Now supports SQL-based row filtering at the source to create pre-filtered reports.AWS Billing Managed Dashboards: An official, pre-configured version of the popular open-source Cost Intelligence Dashboards.AWS Lambda: Recursive loop detection is now available in all commercial regions, reducing the risk of runaway bills.Google Cloud:BigQuery: Offers cost-effective cross-cloud connections (preview) but restricts Graph processing to Enterprise/Enterprise Plus editions.Billing Reports: A new "originating products" filter helps trace costs back to their source.Semantic Tags: A preview feature to automatically propagate application context to resources for better cost allocation.Storage Rapid Cache: Now allows selective ingest filtering to reduce expenditure.Oracle Cloud Infrastructure (OCI):Autoscaling for Red Hat OpenShift: New capability driven by licensing alignment with Red Hat/IBM.Service Limits: OCI has streamlined the service limit increase experience in the console.Sustainability UpdatesNew Availability Zones: New zones in London (EU-West-2D) and Las Vegas seem focused on AI training, raising questions about sustainability trade-offs.Google Cloud Carbon Methodology: A refresh offers more granular certificates and hourly electricity matching for better precision.Green Software Foundation: The "Beyond Watts" paper encourages evaluating environmental impact beyond just electricity consumption.Productivity AIAmazon Q for Microsoft 365: Now generally available for Excel, Word, etc., bringing AI into daily workflows but raising data governance questions.Referenced Resources & LinksFrank's instance price/performance comparison tool: https://aws.frankcontrepois.com/comparisons/

Yaniv Tal grew up in the Bay Area in the 90's with parents who were engineers. They were both in electrical engineering, specifically digital/signal processing. He really saw the internet sport up into a massive thing. He studied the hard sciences (electrical engineering, math and physics), and his first job out of college was at HP, doing board b ring up for printers. Outside of tech, he is recently married, having a lovely wedding on a beach in Mexico. When asked "what do you do for fun?", he can't answer it, cause he is a bit of a workaholic and loves what he does. But, he does make sure he spends quality time with his new bride. In 2017, Yaniv founded a protocol called The Graph, one of the core web3 protocols. After many years on this journey, he started to have a vision around the world's broader knowledge in a decentralized, verifiable way. Eventually, he and his team set out to launch a a solution around this vision, designed to solve fragmentation of knowledge, vulnerable to centralization, silos and misinformation. This is the creation story of Geo. Linkshttps://geobrowser.iohttps://thegraph.com/https://www.linkedin.com/in/yanivtal9/ Current Sponsors: Tiger Data Protected Harbor Render Fitnexa Perplexity Entelligence Checkout our Stacklist! https://stacks.codestory.co/ Hosted by Noah Labhart | Technical Founder & Startup Mentor. Our Sponsors:* Check out Granola and use my code granola.ai/CODESTORY for a great deal: https://granola.ai* Check out Perplexity and use my code CODESTORY for a great deal: https://www.perplexity.aiAdvertising Inquiries: https://redcircle.com/brandsPrivacy & Opt-Out: https://redcircle.com/privacy

a16z's Erik Torenberg sits down with Josh Elman, Olivia Moore, and David Booth to introduce Cosign, a new product built around professional reputation and the people, companies, and products you're willing to put your name behind.They unpack a simple idea at the heart of Silicon Valley: some of the most valuable professional signals aren't credentials, but who believes in you. From the mentor who shaped your career to the colleague you'd work with anywhere or the young builder you think everyone should be watching, Cosign is an attempt to make those signals more visible and durable.They also discuss why human endorsements may become more valuable as AI makes outreach and information abundant, what existing professional networks get right and miss, and how making reputation more legible could help talented people get discovered earlier, find collaborators, and carry the work they've done behind the scenes into whatever they do next.Resources:Follow Josh Elman on X: https://x.com/joshelmanFollow Olivia Moore on X: https://x.com/omooretweetsFollow David Booth on X: https://x.com/david__boothFollow Cosign: https://x.com/cosign_buildRead more about Cosign: https://www.a16z.news/how-silicon-valley-knows-its-people Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

What if you could ask your AI one lazy, one-line question—and get an answer that understands your projects, clients, decisions, tasks, mistakes, and business context?That's what becomes possible when you stop treating AI as a series of disconnected chats and start giving it persistent access to the context of your business. Instead of repeatedly explaining who you are, what you're working on, and what happened last time, your AI can find that context itself.And you don't necessarily need a sophisticated database or expensive knowledge-management platform to make it happen. In this episode, I break down the knowledge graph I built using primarily files, folders, tagging, indexes, and AI—and how you can start building your own.The goal is simple: give AI enough organized context that you can spend less time prompting, searching, explaining, and repeating yourself—and more time getting useful work done.In this session, you'll discover:Why context—not increasingly complicated prompts—is critical to getting better results from AI.How to move from prompt engineering toward context, harness, and process engineering.Why built-in AI memory alone may not provide the granular business context needed for specific projects and workflows.The five components behind my knowledge system.How I organize projects into folders and maintain evergreen project documents and task registries.How tagging and “front matter” help AI understand what files contain without reading everything.How shared drives, recorded meetings, CRM, ERP, task-management systems, email, and other sources can become accessible parts of the broader knowledge environment.Where Obsidian can be useful—and why it isn't necessary to make this system work.The result is an AI that behaves less like a blank chat window and more like a librarian for your business—able to locate relevant knowledge, connect the dots, and use that context when completing work.If you want AI to become more useful across your business, don't just improve what you ask it.Improve what it already knows when you ask.Listen to the full episode to learn how to build the system.About Leveraging AIMulti-Agent Orchestration Course: https://multiplai.ai/multi-agent-orchestration-course/YouTube Full Episodes: https://www.youtube.com/@Multiplai_AI/Connect with Isar Meitis: https://www.linkedin.com/in/isarmeitis/ Join our Live Sessions, AI Hangouts and newsletter: https://services.multiplai.ai/eventsIf you've enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

There is no pyramid anywhere in the ISA 95 standard. David Schultz explains what is actually inside it and how to turn it into a working data model.Most engineers meet ISA 95 as a layered triangle on LinkedIn and never get past it. David Schultz of Amárach StackWorks takes that image apart. The pyramid appears nowhere in the standard, which exists to define how business systems such as ERP exchange information with manufacturing and control systems. Eight parts are published today, Part 1 refreshed in 2025, with Part 9 still in committee. Start with Parts 1 and 2. Part 1 sets the scope, Part 2 delivers the resource, information and process models. Parts 3 and 4 add the activity models and the work master. Parts 5 and 6 supply the verbs and exchange patterns.From there it collapses into a frame you can hold in your head: four resources of equipment, physical asset, personnel and material, four domains of production, maintenance, quality and inventory, and four exchanges covering definition, capability, schedule and performance. Use what works and leave the rest. On adoption David is candid. The full set runs roughly $1,400, and the paywall keeps it out of most LLM workflows because ISA would forfeit copyright by opening it up.Where it breaks is data governance. When every site builds bespoke models, stitching them together means inventing a fourth abstraction. Vlad offers a first hand version from Procter and Gamble, where nobody at plant level owned the data structure moving from Rockwell controls into Proficy. For brownfield sites the advice is the Strangler fig pattern: leave what runs and replace functionality over time.The second half demystifies knowledge graphs using a family tree. A graph stores a subject, a predicate and an object, and the relationship is a first class citizen, not a foreign key. RDF defines the things, RDFS adds schema, OWL adds inference so a system can derive siblings from a shared parent, and SHACL constrains the shapes. Applied to a tank, that means volume and level as properties plus the resource relationship network to map which pump feeds which vessel. Rhize implemented the entire 2018 standard as a graph. David stays cool on AI. A shared ontology means a model never guesses whether two names match, but he expects it to underwhelm. His book pick is The Semantic Web for the Working Ontologist.About David SchultzDavid Schultz is a principal consultant and co founder of Amárach StackWorks, a consultancy describing itself as standards based manufacturing operations management built on a modern technology stack. He has spent close to 30 years in process control and factory automation, starting at the device layer before moving through SCADA into MES and ERP integration. He was previously a senior consultant at Rhize Manufacturing Data Hub, where ISA 95 was implemented as a knowledge graph. He last appeared on Manufacturing Hub for Episode 85.https://www.amarach.ioTimestamps0:00 Introduction and ICC preview2:05 David Schultz and Amárach StackWorks4:10 What ISA 95 actually is and why the pyramid is not in it7:20 Where to start with the standard9:40 The ROI of a shared data model15:00 Which parts of the standard to read first20:40 The adoption gap: cost and misconceptions24:20 Who inside a plant should own the model32:50 Knowledge graphs: RDF, RDFS, OWL and SHACL43:30 The standard is not prescriptive about protocols47:40 Knowledge graphs, ontologies and AI53:40 Future of standards, career advice and book pickReferencesISA 95 Standard and Training: https://www.isa.orgRhize ISA 95 Learning Resources: https://docs.rhize.com/isa-95/About Your HostsVladimir Romanov is a co-host of The Manufacturing Hub Podcast and the founder of Joltek, an independent manufacturing and industrial automation consulting firm specializing in modernization strategy, digital transformation, and workforce development. Joltek works with manufacturers and investors to de-risk modernization and build the internal capability to sustain results.Connect with Vlad: https://www.linkedin.com/in/vladromanov/Want to go deeper? Vlad and the team at Joltek have covered related topics here:Manufacturing Execution Systems: https://www.joltek.com/blog/manufacturing-execution-systems-mesUnlocking Industrial Data in Manufacturing: https://www.joltek.com/blog/unlocking-industrial-data-in-manufacturingDave Griffith is a co-host of The Manufacturing Hub Podcast and founder of Capelin Solutions, an industrial automation firm helping manufacturers adopt smart manufacturing technology. He brings 15 years of experience in industrial automation and digital transformation.Connect with Dave: https://www.linkedin.com/in/davegriffith23/Subscribe to Manufacturing Hub: https://www.manufacturinghub.liveLinkedIn: https://www.linkedin.com/company/manufacturing-hub-networkYouTube: https://www.youtube.com/@ManufacturingHub

This week I kick off a new series on the show: My Favorite Graphs. I'm going to walk through the projects that shaped how I think about data visualization, data communication, and data storytelling. I'm starting with the Guardian's “Bussed Out: How America Moves Its Homeless,” published in December 2017, an 18-month investigation into the one-way bus tickets that US cities hand out to homeless people. There's a special treat at the end too!My Favorite Graph: Bussed OutSubscribe to this podcast. Follow this show and everything dataviz on Instagram, LinkedIn, Substack, X, and YouTube. Check out the PolicyViz websiteto learn more about data and data visualization.Questions? Comments? Pitch ideas? Email me at: jon@policyviz.comHosted by Zencastr where you can record, transcribe, edit, and automatically publish your meetings, podcasts, and more.

What does a three decade-long path of working with AI and asking what separates human cognition from a machine's look like? The answer says as much about the machines as it does about us. At eleven years old, Carlos Perez was already choosing between two possible futures. One was first contact with an alien civilization. The other one was building AI. He picked Inner space over outer space. The route there wasn't direct. After deep learning broke through around 2012, Perez eventually concluded that These are systems that you actually grow - more similar to biology than to physics. Perez co-founded The Intuition Machine in 2015, and he has published a number of books and essays on AI. He also works with Fannie Mae as a a Consultant, providing AI strategy and architecture guidance for the Enterprise Architecture organisation.  We explore his path of discovery and tinkering: from AI's uneven capabilities to what's actually changing between model generations, enterprise architecture, AI harnesses, loop engineering and graph engineering for AI agents. Article published on Orchestrate all the Things: https://linkeddataorchestration.com/2026/09/15/mapping-the-jagged-frontier-of-ai-libraries-ladders-loops-and-graphs/

Målstyrning med Planner Planner har fått en Mål-flik, och den dyker upp i dina planer just nu. Frågan är om det är riktig målstyrning eller en att-göra-lista med en finare rubrik ovanför. För att kunna svara på det behöver vi först få ordning på begreppen: Vad som skiljer en vision från ett mål, en KPI från ett nyckelresultat, och varför OKR kräver ett lager som Planner helt saknar. Vi tar också upp historien om Microsofts föregående verktyg för målstyrning: Viva Goals lades ner vid årsskiftet, och några månader senare kommer något betydligt enklare i Planner. Slutsatsen är ärlig, och den landar i en enda fråga du kan ställa nästa gång någon säger "vi kör målstyrning i Planner". AI-skolan del 8 – Microsoft Graph och WorkIQ Varje gång Copilot svarar på något om din organisation har den först frågat Microsoft Graph, och det gör den med funktionen WorkIQ. Ändå är Graph något de flesta bara hört nämnas i förbifarten,  ett namn på en arkitekturbild och inget man kan peka på. I den här delen av AI-skolan reder vi ut vad Graph och WorkIQ faktiskt är, varför Copilot inte skulle fungera utan dem, och varför den förklarar både det Copilot hittar och det Copilot inte får se. Förstår du Graph och WorkIQ så förstår du varför behörigheter är den kanske viktigaste AI-frågan. Ordlistan bjuder på ett ord du säkert använt hundra gånger utan att riktigt kunna förklara det: API. Nyheter Word, Excel och PowerPoint får ett skrivfält direkt i Copilot-knappen, så du slipper öppna panelen alls. Brand Kits får två inställningar som gör att Copilot äntligen håller sig till mallen, och lyder instruktioner du skrivit in i mallens anteckningsfält. Teams låter dig sätta en påminnelse direkt på ett meddelande, och Purview får en retentionsregel som raderar filer ingen öppnat på länge, med den uttalade motiveringen att Copilot då svarar bättre. 

Check out this link to buy DB's Books[link] # Story Author Brief description 01 The 4D Doodler Graph Waldeyer A researcher's idle doodling attracts the attention of a fourth-dimensional intelligence, producing increasingly strange and dangerous consequences. 02 Bread Overhead Fritz Leiber An experiment in producing lighter bread goes spectacularly wrong, filling the skies with floating loaves and creating a satirical technological crisis. 03 Image of the Gods Alan E. Nourse Human colonists resist authoritarian interference from Earth and discover unexpected allies among their planet's indigenous inhabitants. 04 Martian V.F.W. G. L. Vandenburg Giant Martian ants appear at a New York parade claiming to represent extraterrestrial war veterans, but their participation conceals a darker purpose. 05 One Shot James B. Blish Intelligence officials investigating a mysterious object in New York Harbor turn to an exceptional gambler when millions of lives depend on one decision. 06 Out Around Rigel Robert H. Wilson Two travelers undertake an ambitious interstellar voyage, confronting jealousy, alien dangers, and the devastating consequences of time dilation. 07 Pygmalion's Spectacles Stanley G. Weinbaum An experimental pair of spectacles transports its wearer into an immersive artificial world where simulated experiences become intensely personal. 08 The Repairman Harry Harrison An interstellar mechanic must repair a malfunctioning hyperspace beacon concealed within a sacred alien structure. 09 Toy Shop Harry Harrison A seemingly impossible toy rocket catches the attention of military researchers and hints at an unexpected scientific breakthrough. 10 Warning from the Stars Ron Cocking A message from a supposedly dead scientist reveals extraterrestrial intervention aimed at preventing humanity's nuclear self-destruction.

Check out this link to buy DB's Books[link] # Story Author Brief description 01 The 4D Doodler Graph Waldeyer A researcher's idle doodling attracts the attention of a fourth-dimensional intelligence, producing increasingly strange and dangerous consequences. 02 Bread Overhead Fritz Leiber An experiment in producing lighter bread goes spectacularly wrong, filling the skies with floating loaves and creating a satirical technological crisis. 03 Image of the Gods Alan E. Nourse Human colonists resist authoritarian interference from Earth and discover unexpected allies among their planet's indigenous inhabitants. 04 Martian V.F.W. G. L. Vandenburg Giant Martian ants appear at a New York parade claiming to represent extraterrestrial war veterans, but their participation conceals a darker purpose. 05 One Shot James B. Blish Intelligence officials investigating a mysterious object in New York Harbor turn to an exceptional gambler when millions of lives depend on one decision. 06 Out Around Rigel Robert H. Wilson Two travelers undertake an ambitious interstellar voyage, confronting jealousy, alien dangers, and the devastating consequences of time dilation. 07 Pygmalion's Spectacles Stanley G. Weinbaum An experimental pair of spectacles transports its wearer into an immersive artificial world where simulated experiences become intensely personal. 08 The Repairman Harry Harrison An interstellar mechanic must repair a malfunctioning hyperspace beacon concealed within a sacred alien structure. 09 Toy Shop Harry Harrison A seemingly impossible toy rocket catches the attention of military researchers and hints at an unexpected scientific breakthrough. 10 Warning from the Stars Ron Cocking A message from a supposedly dead scientist reveals extraterrestrial intervention aimed at preventing humanity's nuclear self-destruction.

Thursday 9/3 at 12:00 PM CDT, Walker Reynolds is going LIVE to bring the Atlas build full circle. In Tuesday's “Is This What's Missing?” video, Walker broke down the relationship between the Unified Namespace and knowledge graphs: the UNS as the real-time data backbone, and the knowledge graph as the layer that gives those connections meaning. Tomorrow, he's opening it up live. Walker will walk through what he built with Flow Software's Atlas, dig deeper into where knowledge graphs fit in the industrial architecture, and answer your questions. If you're trying to understand how the UNS, knowledge graphs, and agentic AI fit together, bring your questions. Thursday, September 3 12:00 PM CDT LIVE with Walker Reynolds The UNS gives you the flow of information. Atlas gives you the map of meaning across that flow.

Today we are talking about Drupal Performance, Rapid Development, and Drupal Canvas Maturity with our hosts. We'll also cover Microsoft 365 FullCalendar as our module of the week. For show notes visit: https://www.talkingDrupal.com/568 Topics Deprecating Module Theme Files Migrating Hooks to Classes Why This Change Matters Drupal Performance Gains Performance Audits and Lighthouse Automating Checks and Spreadsheet Rant AI Spreadsheet Cautionary Tale Privacy Concerns with AI Freelancer Pressure Rapid Change Reality Canvas Release Risks Community Support Needed AI For Documentation Canvas Production Readiness Canvas Architecture Debate AI For Voting Research LLM Bias And Sources Resources Rebrickable Webpagetest Lighthouse Tugboat Drupal canvas Guests Martin Anderson-Clutz - mandclu.com mandclu Hosts Nic Laflin - nLighteneddevelopment.com nicxvan John Picozzi - epam.com johnpicozzi Amber Matz - tugboatqa.com [amber himes matz](https://www.drupal.org/u/amber himes matz) MOTW Correspondent Martin Anderson-Clutz - mandclu.com mandclu Brief description: Have you ever wanted your users' own Outlook calendars to show up right alongside your Drupal content in a calendar view? There's a module for that. Module name/project name: Microsoft 365 FullCalendar Brief history How old: created just last month, August 19 2026, by fabianderijk of Finalist Versions available: 1.0.0, which works with Drupal 11 Maintainership Brand new — the first and only release is from last month, and the whole commit history is basically launch day Security coverage: brand new, so not yet Test coverage: yes, both unit tests and kernel tests Documentation: a genuinely thorough README — it walks through privacy, the config guard rails, and three different ways to customize event output Open issues: none yet, it's less than two weeks old Usage stats: Too new for a site count Module features and usage With this installed, it adds the signed-in user's Microsoft 365, or Outlook, calendar as an extra event source on a FullCalendar view — so their personal appointments sit right next to the Drupal content the view already renders It leans on the Microsoft 365 Connector module and its SSO submodule, plus the FullCalendar module. Each user must have signed in through Microsoft 365 SSO: anyone who hasn't just sees no events, which is a clean fallback It uses lazy loading, so it only fetches events in the date range the calendar is currently showing, not your whole calendar Privacy is baked in: anything marked private or confidential in Outlook is masked, so it shows up as just "Busy", with no title, location, or meeting link, unless the site builder deliberately turns masking off The response itself is per-user and marked private, no-store, so it never lands in a shared or CDN cache There's a clever server-side cache too: it stores the raw Graph response before masking, so a single fetch can serve several displays that each have different masking settings You get guard rails you can tune with Drush or an admin form: max events, max date range, cache lifetime, and a separate, shorter failure cache That failure cache is a nice touch — if there's no active Microsoft session, or Graph errors out, it caches the empty result briefly so a broken connection doesn't get re-polled on every single calendar click Under the hood it calls Graph's calendarView endpoint rather than /me/events, which means recurring meetings get expanded into their individual occurrences — exactly what a calendar grid needs Every event carries CSS classes for its status — busy, free, tentative, out-of-office, working elsewhere, cancelled — so you can style them however you want And if CSS isn't enough, there's a server-side alter hook and a JavaScript pre-build event for fully custom rendering. Nice detail: the hook is explicitly guarded so you can't use it to put back a title or location that masking just stripped out Clearly this will be more useful for edge cases, for example an intranet, but I think this is a really interesting example of the power of Drupal as an integration layer, or as some like to put it, the "glass" through which a user can interact with multiple systems

Andrew welcomes Morten Mynster back to the PowerShell Podcast to dig into the projects he's been building around Microsoft Graph, Entra, least privilege, and authentication. Morten walks through Least Privileged Entra, a module that uses activity logs to identify users who may have more permissions than they actually need, and MS Graph Proxy, which lets developers work with mocked Microsoft Graph data locally without connecting to a live tenant. They also get into managed identities, the rougher corners of Microsoft 365 APIs, testing Graph-based projects in CI/CD pipelines, and how contributing to open source can solve real problems while creating unexpected career opportunities.   Key Takeaways: · Least privilege is easier when you can see what people actually use. Morten's Least Privileged Entra module compares assigned Entra roles with activity data to identify permissions that may be unnecessary and suggest more limited alternatives. The goal is to give admins something actionable rather than simply reporting that a configuration passed or failed. · You don't always need a live Microsoft 365 tenant to develop against Microsoft Graph. MS Graph Proxy intercepts Graph requests and responds with mocked data, allowing developers to test scripts and modules locally, offline, or inside CI/CD pipelines. It can also identify the minimum Graph permissions associated with the endpoints an application uses. · Sharing your work can have benefits far beyond the project itself. Morten credits his Least Privileged MS Graph project with helping him land his current job. His approach is simple: solve a real problem, share the solution, contribute where you can, and let other people build on what you've learned.   Guest Bio: Morten Mynster is an IT professional and open source contributor focused on Microsoft Entra, Microsoft Graph, security, and least privilege. His projects include LeastPrivilegedMSGraph, LeastPrivilegedEntra, and MSGraphProxy, and he regularly contributes to community projects and discussions around Microsoft 365 security and PowerShell. Resource Links: https://github.com/Mynster9361/Least_Privileged_MSGraph https://github.com/Mynster9361/LeastPrivilegedEntra https://github.com/Mynster9361/msgraphProxy https://github.com/FriedrichWeinmann/EntraAuth The PowerShell Podcast on YouTube: https://youtu.be/cB_QE1HvAyI

In this episode of Being Human, Steve and Lisa Cuss explore how politics, social media, and the stories we tell ourselves shape who we become. Pastor and political analyst Ryan Burge explains how social media amplifies political extremes, distorts public discourse, and increasingly allows partisanship to shape religious identity. Jessie Cruickshank and Julia Schmaltz, co-authors of Becoming Good News, unpack how internalized narratives, legalism, perfectionism, and the Christian “Übermensch” ideal—the pressure to become a superhuman believer—can hinder spiritual growth. Steve and Lisa use these conversations to offer a path toward self-awareness, grace, presence, and embracing weakness as essential to genuine faith and transformation. Original Episodes: Being Human: How Memory, Identity, and Story Shape Discipleship with Jessie Cruickshank & Julia Schmaltz Being Human: How Politics Hijacked the Pulpit with Ryan Burge Ryan Burge's The Vanishing Church: How the Hollowing Out of Moderate Congregations Is Hurting Democracy, Faith, and Us Jessie Cruickshank & Julia Schmaltz's Becoming Good News: Reimagining Discipleship Through Identity, Story, and Science Nietzsche's concept of the Übermensch More from Ryan Burge: Ryan's Graphs about Religion Ryan on X (formerly Twitter) More from Jessie Cruickshank and Julia Schmaltz: Jessie Cruickshank's Ordinary Discipleship podcast Julia Schmaltz's Sermons Learn more about your ad choices. Visit podcastchoices.com/adchoices

Using Microsoft Graph and Powershell to Mine for Information - Stale Accounts and Licenses https://isc.sans.edu/diary/Using%20Microsoft%20Graph%20and%20Powershell%20to%20Mine%20for%20Information%20-%20Stale%20Accounts%20and%20Licenses/33264 Using Microsoft Graph and Powershell - Risk Detection Commands https://isc.sans.edu/diary/Using%20Microsoft%20Graph%20and%20Powershell%20-%20Risk%20Detection%20Commands/33266 Keycloak Vulnerability https://github.com/keycloak/keycloak/issues/51833 https://www.keycloak.org/2026/08/keycloak-2672-released CRYPTOGRAPHIC CONTEXT INJECTION ATTACK https://adversa.ai/blog/cryptographic-context-injection-grok-data-theft/ N-able password manager https://amibeingpwned.com/blog/solar-winds-part-2-avoided?_sp=75fd154a-e34f-41d0-8624-7c285776c13d.1787263544340 My Upcoming Classes https://www.sans.org/profiles/dr-johannes-ullrich

Every CDO is being told to make their data AI-ready. What that actually requires is a harder conversation — and this episode gets into it.

Alex talks about what’s on his mind when it comes to tech, enjoy!Find everything Alex at AlexMerced.com

I'm joined by Cobus Greyling, AI evangelist at Kore.ai, for a wide-ranging conversation on where enterprise AI actually stands today.We explore why many of the ideas now being presented as breakthroughs, from graph engineering and orchestration to intent detection and state management, have deep roots in conversational AI. We also get into one of the biggest barriers to enterprise AI adoption: data. We highlight why customer experience could become the entry point for much wider organisational transformation. We touch on how forward-deployed engineers could help businesses navigate that change and why security will become increasingly important as AI agents gain access to more systems and information.I close with a buzzword bingo round for Cobus covering fleet engineering, agent swarms, orchestration and universal agents, and Cobus explains why he thinks Open Claw was the biggest red herring of the year.Show notes
Follow Cobus on LinkedIn: https://www.linkedin.com/in/cobusgreylingCobus's website: https://cobusgreyling.meFollow Cobus on Substack: https://cobusgreyling.substack.com/Discover more about Kore.aiArticle - The Untrainable by Sarah Guo: https://saranormous.substack.com/p/the-untrainableVideo - Andrej Karpathy: Software Is Changing (Again): https://www.youtube.com/watch?v=LCEmiRjPEtQFollow Kane on LinkedIn:⁠⁠⁠⁠⁠⁠https://www.linkedin.com/in/kanesimms⁠⁠⁠⁠⁠⁠Find out more about VUX:⁠⁠https://vux.ai⁠⁠Subscribe to VUX World: ⁠⁠⁠⁠⁠⁠⁠https://vuxworld.typeform.com/to/Qlo5aaeW?utm_source=podcast&utm_medium=audio⁠⁠⁠⁠⁠Subscribe to The AI Ultimatum Substack: ⁠⁠⁠⁠⁠⁠⁠https://open.substack.com/pub/kanesimms⁠

Salk created some graphs that show the improvement of the Seahawks running game over the course of the season last year. Brock and Salk discuss how the Offensive Line got better and how it affected the Hawks running game and offense as a whole. They talk about how the Mariners get the train back on the tracks, tonight’s HBO Hard Knocks, and more in Need To Know. They discuss why this year is different from last year for the Mariners, why the expectations have changed, and why the team isn’t meeting those expectations. And Brock talks about what Dan Wilson could learn from Mike Macdonald, Quinnen Williams’ extension in Dallas, and more in Blue-88.

Graph engineering is AI's latest buzzy term—but it offers a useful framework for organizing agents, tools, knowledge and humans into working systems. NLW explains the evolution from prompts to graphs. In the headlines: OpenAI delays Astra, ByteDance trains a massive model, open-weight AI tests revenue sharing and Claude Code embraces Auto Mode.AIDB's AI Summer Adventure: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://summeradventure.ai/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Brought to you by:KPMG – Research from KPMG and the University of Texas at Austin shows the highest-impact AI users treat AI like a reasoning partner — and those skills can be taught at scale. Learn more at ⁠⁠⁠⁠https://kpmg.com/us/Sophisticated⁠⁠⁠⁠Hyperagent - Hire a fleet of always-on agents. New users get $1,000 in inference. ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠hyperagent.com/aidailybrief⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Rackspace Technology- One accountable partner to build, operate and run your full enterprise AI stack ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.rackspace.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Section - Section turns AI investment into workforce transformation and ROI - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.sectionai.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Blitzy - Want to accelerate enterprise software development velocity by 5x? ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://blitzy.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠AssemblyAI - The best way to build Voice AI apps - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.assemblyai.com/brief⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Robots & Pencils - Cloud-native AI solutions that power results ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://robotsandpencils.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://pod.link/1680633614⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Our Newsletter is BACK: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://aidailybrief.beehiiv.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Interested in sponsoring the show? sponsors@aidailybrief.ai

In this episode of Being Human, host Steve Cuss speaks with political scientist and pastor Ryan Burge about his book The Vanishing Church. They explore how American Christianity has become increasingly radicalized by political partisanship, the role of social media algorithms in amplifying extreme voices, and the disappearance of moderate congregations. Burge discusses how partisanship now shapes religious identity rather than the reverse, the impact of COVID had on churches, and how the church has become a "hospital for the healthy."  Episode Resources: Ryan Burge's The Vanishing Church: How the Hollowing Out of Moderate Congregations Is Hurting Democracy, Faith, and Us Michael Wear's CT review of Ryan's work on the “big church sort” Pew research: How the Pandemic Has Affected Attendance at U.S. Religious Services A Look at The Donatist controversy Who was Count Zinzendorf? Scripture Referenced: 1 John 1:5 (ESV)  Matthew 20:16; Luke 13:30; Mark 10:31 (ESV) Mark 2:17 (ESV) More from Ryan Burge: Ryan's Graphs about Religion Ryan on X (formerly Twitter) Sign up for Steve's Newsletter & Podcast Reminders Capable Life Newsletter Get the Assets & Liabilities pdf and the Life Giving List Join Steve at an Upcoming Intensive Capable Life Intensives Learn more about your ad choices. Visit podcastchoices.com/adchoices

AI agents can write code for hours, but ask them to do real work in the real economy, and they break. Mitch Troyanovsky is co-founder of Basis, a unicorn AI company whose agents run autonomously for hours — sometimes days — completing complex tax returns end to end. His answer to the reliability problem: stop grading outcomes, and start supervising the process.This is a definitive, reference-style conversation on building long-horizon AI agents. Mitch walks through the full history — from ReAct and the AutoGPT crash to reasoning models and RLVR — and explains why the industry abandoned process supervision in 2023, and why it's now coming back at a completely different scale. We go deep on behavior specs, the open standard Basis just released with Braintrust for defining and evaluating how agents behave across entire trajectories, with no ground truth required.Along the way: why context is really runtime training data, why your documentation must be treated like a codebase, ontologies as "worlds for agents to live in," the judge-as-agent architecture, why Basis hires philosophy majors as Language Architects, deploying agents as "onboarding 300 brilliant alien employees," and Mitch's prediction for when the bitter lesson swallows the harness.(01:09) Why Basis Engineers Whisper to Their Agents(04:12) Accounting as Compression: an Intelligence Layer Over the Economy(06:11) Defining Long-Horizon: When You Exceed the Context Window(08:24) Anatomy of a Multi-Day Autonomous Trajectory(10:19) Handoff Design: Optimizing Output for the Reviewer(11:17) ReAct and Why Reasoning Must Regulate Its Own State(12:33) Large Working Memory, No Long-Term Memory(14:13) Compounding Errors: Why AutoGPT and BabyAGI Broke(15:51) Opus 3, o1, o3: the Three Real Paradigm Shifts(17:07) Titrating Inference Compute Across Easy and Hard Steps(18:23) Process Reward vs. Outcome Reward: "Let's Verify Step by Step"(20:32) RLVR and Why the METR Curve Overstates Reliability(22:09) Verifiable at Runtime: the Real Reason Coding Won(25:14) No Ground Truth, No Cheap Verification, No Data(26:55) Encoding Deterministic Checks From Human Review Process(29:18) Synthetic Data Limits: Generating Artifacts, Not Text(33:16) 100 Evals Pass — Does It Generalize to Production?(35:53) Primary Sources vs. Pre-Training Knowledge(36:37) Behavior Specs: Markdown, Judges, and True/False/N.A.(39:58) Specificity vs. Brittleness in Spec Authoring(42:18) Context as Runtime Training Data(44:21) Judge-as-Agent: Trajectory Maps and Sub-Agent Attribution(46:45) The Move 37 Objection: Reliability Over Optimality(50:02) The Magic Box Model: Building Without Weights Access(52:41) "Nothing Paradigm-Shifting Has Changed Since o3"(54:56) Open-Sourcing the Behavior Spec Standard With Braintrust(01:02:54) Ontology Design: Virtual Filesystems, Graphs, Embeddings(01:04:20) Canonical vs. Non-Canonical: Docs as Codebase(01:06:33) Language Architects and Writing for Runtime Interpretation(01:09:05) Deployed Intelligence: 300 Alien Employees With No Context(01:11:10) Closing the Loop: Signal → Context, Tools, Harness(01:12:50) Context Slop: the Mistake Most Agent Builders Make(01:14:29) Reward Function Design and Credit Assignment Over Trajectories(01:17:01) Will the Bitter Lesson Swallow the Harness?(01:18:46) Business Moats vs. Technical Moats(01:21:03) Paradigm Thinking Over Timeline ADHD

Why You Should Be Using Graphs in Your Appraisal Reports

AI is moving fast in life sciences, but a confident answer is not the same thing as a correct, reproducible, auditable answer. Knowledge graphs, ontologies, and FAIR data practices are the keys to bridging this gap. Knowledge3's Tom Plasterer, CEO and co-founder, and Eric Little, chief data officer, join host Allison Proffitt to discuss what it takes to turn semantics into something practical: a knowledge product mindset, modular delivery, and “semantic ops” that can better manage data systems. Their conversation challenges the hype cycle around context graphs, explains why life sciences are a natural starting point for regulated, high-stakes AI, and shows how layered models can capture the right level of context without building a giant monolith that never takes off. Links from this episode:  Bio-IT World BioTeam Bio-IT World Europe Knowledge3 Bio-IT World's Trends from the Trenches podcast delivers your insider's look at the science, technology, and executive trends driving the life sciences through conversations with industry leaders. 

In episode 297 of our SAP on Azure video podcast we talk about Fabric IQ access to SAP Graph. AI agents are becoming more and more powerful, and in the SAP world one of the big questions is still: how do we give these agents meaningful access to SAP business context without just copying data around? This is where things like SAP Graph, SAP ontologies, knowledge graphs, MCP servers and Fabric IQ become really interesting.Mario has been doing a lot of work in this space. The MCP Server for SAP Datasphere environments that he has developed has gained a lot of popularity and is used by several customers. So for today I am really happy to have Mario de Felipe joining us for the first time, to show us what he has built and to talk about how all of this can help bring SAP context into Microsoft Agents.Find all the links mentioned here: https://www.saponazurepodcast.de/episode297Reach out to us for any feedback / questions:* Goran Condric: https://www.linkedin.com/in/gorancondric/* Holger Bruchelt: https://www.linkedin.com/in/holger-bruchelt/ #Microsoft #SAP #Azure #SAPonAzure #FabricIQ #Fabric #Copilot

Parker Fleming (@Statsowar) has released some fun CFB graphics ahead of the season!

手机拍的 “图”、画布上绘制的 “图”、以及数学公式在坐标系中对应的 “图”……这些不同类型的 “图”，在英语里的说法一样吗？又该如何区分使用 “picture”、“image”、“photo” 和 “graph” 呢？听节目，跟主持人步理和 Phil 一起辨析这四个表示 "图，图像” 的单词之间的区别。

Bun quitte Zig pour Rust en 11 jours à coups de Claude Code, pour 165 000$ payés par Anthropic : la réaction du créateur de Zig ne se fait pas attendre. TypeScript 7 débarque, réécrit en Go, 8 à 12x plus rapide. Entre les deux, Vidocq réimplémente Jakarta EE en souverain, le COBOL met un uppercut aux microservices, et un CTO demande à son équipe combien de temps il lui faudrait pour revenir à sa vélocité antérieure sans Claude Code. De quoi réfléchir avant le prochain rewrite. Enregistré le 17 juillet 2026 Téléchargement de l'épisode LesCastCodeurs-Episode-342.mp3 ou en vidéo sur YouTube. News Langages Est-ce qu'on peut aussi utiliser des double, des longs, ou autre pour gérer les montants monétaires en Java ? https://blog.frankel.ch/bigdecimal-vs-double/ double (IEEE 754) Usage : Calculs scientifiques, métriques, statistiques. Avantages : Très performant (matériel), idéal pour l'approximatif. Risques : Erreurs d'accumulation, égalité (==) trompeuse, NaN / -0.0. Bonnes pratiques : Utiliser une tolérance (epsilon ou ULP) pour comparer ; utiliser des algorithmes de sommation compensée (Kahan/Neumaier) pour la précision. BigDecimal Usage : Finance, comptabilité, fiscalité (précision décimale stricte). Avantages : Contrôle total des arrondis et de l'échelle. Risques : Lent (allocations), immutabilité (risque de mauvaise réaffectation), confusion equals() vs compareTo(). Bonnes pratiques : Initialiser via String ou valueOf() ; utiliser compareTo pour l'égalité. Point fixe (long) Usage : Trading, systèmes haute performance, paiements. Avantages : Très rapide, déterministe, zéro allocation. Risques : Gestion manuelle de l'échelle et des débordements (Math.addExact). Points de vigilance en production Sérialisation (JSON) : Préférer les String pour BigDecimal pour éviter la perte d'échelle. Atomicité : double n'est pas atomique ; utiliser volatile ou DoubleAdder (pour les compteurs). Tests : Toujours définir un delta ou Offset pour les tests de flottants. Bibliothèques recommandées Moneta (JSR 354) : Standard bancaire complet. decimal4j : Optimisé pour le point fixe haute performance. Apache Commons Numbers : Outils robustes pour la précision et les sommations. Typescript 7 est de sortie devblogs.microsoft.com/typescript/announcing-typescript-7-0 Performance majeure : Portage natif en Go offrant des gains de vitesse de 8x à 12x et une consommation mémoire réduite. Architecture optimisée : Utilisation du multithreading (mémoire partagée) et parallélisation native (analyse, vérification de types,émission). Nouvelles options de contrôle : Introduction des flags –checkers, –builders (parallélisation) et –singleThreaded (mode mono-cœur). Nouvel observateur de fichiers : Passage à une solution basée sur @parcel/watcher pour une meilleure réactivité et stabilité du mode –watch. Compatibilité et transition : Compatible avec les bases de code TypeScript 6.0. Utilisation du package @typescript/typescript6 recommandée pour maintenir des outils dépendants de l'ancienne API. Changements de configuration : Durcissement des défauts (ex: strict activé par défaut) et suppression de nombreuses options obsolètes (target: es5, baseUrl, etc.). Amélioration de l'expérience éditeur : Serveur de langage (LSP) plus stable avec une réduction de 80 % des erreurs et 60 % des crashs. Limitations actuelles : Support incomplet pour les frameworks utilisant des plugins de langage (Vue, Svelte, Astro, Angular) en attendant une API stable. "Java, the documentary" est sur YouTube, retraçant l'histoire du langage youtube.com/watch?v=… La vidéo n'était pas encore disponible à l'heure de l'enregistrement. Sortie officielle le 17 juillet. Avec des interviews de James Gosling, Brian Goetz, Venkat Subramaniam, et bien d'autres. Librairies What's New in 8.0 - Hibernate docs.hibernate.org/orm/8.0/whats-new L'intégration de Jakarta Persistence 4.0 apporte des nouveautés majeures comme EntityAgent (qui standardise la StatelessSession), les mappings de result set en SQL natif, et de nouvelles options de configuration de session et de requêtes (Session Creation Options, Query Options). Le support de Jakarta Data 1.1 est ajouté pour les Hibernate Data Repositories, incluant l'intégration avec les requêtes statiques JPA4, les projections @Select, et les repositories asynchrones via Jakarta Concurrency ou Hibernate Reactive. L'introduction du Graph-based Flushing remplace l'ancienne approche basée sur des heuristiques par un modèle de dépendances utilisant les contraintes relationnelles, afin d'améliorer la fiabilité des tris, la gestion des batchs et les performances globales (bien que l'ancienne méthode reste temporairement disponible). L'API ProcedureCall a été améliorée pour faciliter le casting des résultats (asResultSetOutput) et permettre la déclaration paresseuse (lazy) du mapping des ResultSet. Hibernate supporte désormais la sécurité au niveau de la ligne (Row-Level Security) de manière native pour les bases de données compatibles (PostgreSQL, Db2, SQL Server, CockroachDB) afin de gérer la visibilité en contexte multi-tenant. Une nouvelle méthode getReference() permet dorénavant de récupérer la référence d'une entité directement à partir de son natural id. Le mode Safe Mode Validator (hibernate.query.safe_mode_enabled=true) fait son apparition pour bloquer les opérations risquées comme sql(), function() ou column() dans les requêtes HQL et Criteria, ce qui est particulièrement utile pour les applications exposées aux LLMs. La gestion des associations bidirectionnelles lors de la phase de flush peut maintenant être prise en charge automatiquement par Hibernate (hibernate.bidirectionality_management=true), synchronisant la référence côté inverse de l'association. Le Subselect Fetching est considérablement amélioré, supportant dorénavant les associations "to-one" pour le bulk select fetching (au lieu de se limiter aux collections) et devenant une option de premier ordre via FetchMethod.BY_SUBQUERY. Un des papas de Cucumber et Gherkin lance Var, une alternative pour le test et le BDD var.oselvar.com Lancement de Vár : Nouvel outil de test créé pour pallier les défauts de Cucumber. Limites de Cucumber : Syntaxe Gherkin trop rigide, intégration difficile avec les exécuteurs de tests et support éditeur limité. Usage avec l'IA : Conçu spécifiquement pour vérifier que les agents IA respectent les intentions et spécifications de l'utilisateur. Fonctionnement : Utilisation du Markdown plutôt que du Gherkin ; sert à la fois de guide et d'outil de vérification. Développement assisté : Code et documentation générés en grande partie par Claude sous supervision humaine. Appel aux retours : Projet ouvert aux tests et aux critiques de la communauté. Web Une nouvelle méthode HTTP : QUERY https://kreya.app/blog/new-http-query-method-explained/ Méthode HTTP QUERY (RFC 10008) pour les recherches complexes. Problème : GET (limité par l'URL) vs POST (sémantique inadaptée). Avantages : Permet un corps de requête, sûr, idempotent et cacheable. Limites : Support infrastructurel faible, non partageable par lien, cache complexe. Usage : À réserver aux requêtes complexes si l'environnement le permet. Comment je fais du design en tant que dev backend eventuallycoding.com/p/comment-je-fais-du-design-en-tant-que-dev-backend Hugo Lassiège retrace l'évolution de son workflow de création d'interfaces en tant que développeur backend, depuis ses débuts avec Bootstrap jusqu'à l'ère de l'intelligence artificielle. L'article explique comment la structuration des éléments visuels a progressé grâce à l'Atomic Design, l'émergence des design systems et l'adoption des design tokens via un framework comme Tailwind. L'auteur détaille son processus actuel qui s'appuie fortement sur Claude Design pour générer et itérer sur des maquettes à partir d'un brief, d'un screenshot ou d'un design system de référence. Il aborde également le risque de slopification et de standardisation extrême apporté par ces outils, rappelant que si l'IA simplifie la technique, il reste crucial d'injecter de l'identité et de l'originalité pour éviter un web trop aseptisé. Data et Intelligence Artificielle De l'utilisation de SKILL.md et de "loop engineering" pour augmenter sa productivité glaforge.dev/posts/…/of-skills-and-loops-with-ai-assistance Les skills permettent d'encoder une procédure de manière répétable et automatisable Le loop engineering enlève l'humain de la boucle afin que l'agent atteigne un objectif donné de façon plus autonome Pour écrire des Codelabs (sorte de tutoriel guidé pas à pas) Guillaume a transformé une séance de création de codelab avec son agent préféré (Antigravity) en skill réutilisable pour l'écriture de ses prochains codelabs Il a également utilisé l'approche de "loop engineering" à la mode en ce moment pour que son agent IA compile, exécute, teste les instructions et le code de son codelab, pour qu'il soit complètement fonctionnel Gain estimé : passer de 2 jours de travail à moins de 2 heures ! Redeploying Claude Fable 5 anthropic.com/news/redeploying-fable-5 Anthropic a annoncé le rétablissement de l'accès à ses modèles Claude Fable 5 et Mythos 5, qui avaient été suspendus suite à des restrictions d'exportation imposées par le gouvernement américain le 12 juin 2026. Cette suspension faisait suite à un rapport d'Amazon démontrant une méthode pour contourner les garde-fous de Fable 5, lui permettant d'identifier et d'exploiter une vulnérabilité logicielle (un jailbreak). Pour y remédier, Anthropic a renforcé ses mécanismes de sécurité en déployant un nouveau classifieur capable de bloquer cette technique spécifique dans plus de 99 % des cas, acceptant en contrepartie une augmentation des faux positifs sur des requêtes bénignes. Face à l'absence de consensus sur l'évaluation des jailbreaks, Anthropic s'associe à Amazon, Microsoft, Google et d'autres partenaires pour développer un standard industriel évaluant la sévérité de ces failles selon quatre critères : gain de capacité, étendue du gain, facilité d'arsenalisation et découvrabilité. L'entreprise s'engage également à approfondir sa collaboration avec le gouvernement américain, notamment via des évaluations pré-déploiement, un partage rapide d'informations sur les failles, et des ressources dédiées à la recherche conjointe sur la sécurité de l'IA. Outillage La réécriture de Bun en Rust et la réaction du créateur de Zig bun.com/blog/bun-in-rust et andrewkelley.me/post/my-thoughts-bun-rust-rewrite.html Bun, le runtime JavaScript et TypeScript écrit à l'origine en Zig, a été entièrement réécrit en Rust pour des raisons de stabilité et de gestion de la mémoire. Cette migration massive d'un demi-million de lignes de code a été bouclée en seulement 11 jours grâce à l'utilisation intensive de Claude Code fonctionnant en parallèle, pour un coût d'API estimé à 165 000 dollars financé par Anthropic. Andrew Kelley, le créateur de Zig, a réagi publiquement en qualifiant l'ancienne base de code de Bun de "slop" remplie de hacks et de fuites mémoire accumulées par une course aux fonctionnalités. Kelley exprime son soulagement face à ce départ, expliquant que les plantages incessants de Bun devenaient un passif réputationnel toxique pour le langage Zig et sa fondation. Le rachat de Bun par Anthropic fin 2025 avait déjà mis fin aux donations financières de Bun envers la Zig Software Foundation, facilitant cette séparation. La nouvelle version Rust de Bun passe désormais la quasi-totalité des tests, réduit la taille du binaire et est déjà déployée de manière transparente en production dans Claude Code. Nouveautés de Git 2.55 github.blog/open-source/git/highlights-from-git-2-55 Support natif de FSMonitor sous Linux via inotify pour accélérer les commandes comme git status sur les grands dépôts Intégration de la compaction incrémentale MIDX (multi-pack index) dans git repack pour optimiser la réécriture des métadonnées Amélioration drastique des performances de génération des bitmaps et des pseudo-merge bitmaps lors des tâches de maintenance Nouvelle commande expérimentale git history fixup pour intégrer facilement des modifications locales dans un commit antérieur Possibilité d'exécuter des hooks configurés en parallèle pour optimiser le temps de build et de validation Utilisation d'un autostash automatique lors d'un git checkout -m en cas de conflit de fusion pour éviter de bloquer l'espace de travail Nouvelle commande git format-rev permettant de formater rapidement des commits reçus via l'entrée standard (stdin) Support du push simultané vers un groupe de remotes configuré Protection contre l'exécution de séquences de contrôle de terminal malveillantes via les flux de progression distants Vidocq, une réimplémentation souveraine et sans dépendance de Jakarta EE et Microprofile vidocq.dev/posts/vidocq-a-sovereign-jakarta-ee-and-microprofile-runtime Lancement de Vidocq : Runtime Java open source complet, compatible Jakarta EE Core Profile et Souveraineté numérique : Projet européen hébergé sur Codeberg, sous licences EUPL 1.2, EPL 2 et GPL 2.0. Standardisation totale : Implémentation fidèle des spécifications (CDI, REST, JSON, etc.), validée par 5 650 tests TCK officiels. Sécurité radicale : Zéro dépendance externe et aucune bibliothèque tierce. Aucune manipulation de bytecode à l'exécution (« magie » générée à la compilation via JDK 25). Compatible JPMS, AOT, GraalVM et Leyden CDS. Disponibilité : Projet en phase alpha, code et documentation accessibles sur vidocq.dev. Article complémentaire qui revient sur la genèse de Vidocq, en utilisant l'IA et les TCKs pour driver l'aspect spec-driven development vidocq.dev/posts/the-story-of-vidocq Le "selfware" : Guillaume s'est fait plais' en vibe-codant son propre éditeur de texte glaforge.dev/posts/…/selfware-building-my-own-text-editor-without-knowing-swift Concept de « Selfware » : création de logiciels conçus exclusivement pour soi-même, sans monétisation ni contraintes liées aux utilisateurs tiers. Le rôle de l'IA : les agents de programmation (comme Antigravity) suppriment la barrière technique de l'apprentissage des langages (Swift, APIs) pour les non-développeurs. Développement minimaliste : privilégier la performance et l'utilité directe (démarrage instantané, interface native) au détriment des fonctionnalités complexes (plugins, télémétrie, gestion de comptes). Absence de pression : libération des contraintes liées à la compatibilité, à la maintenance logicielle et aux retours utilisateurs ; le logiciel n'a besoin d'être « assez bon » que pour ses propres besoins. Incitation à l'autonomie : encourager la création d'outils sur mesure pour résoudre les frictions quotidiennes plutôt que de subir les limitations des logiciels commerciaux. Architecture Le cobol a donné un uppercut au microservices https://freedium-mirror.cfd/@maahisoft20/your-microservices-lost-to-cobol-let-that-sink-in-8ce2e236d007 Retour d'expérience sur la migration d'un système COBOL vers des microservices cloud-native qui s'est soldée par un retour en arrière après avoir constaté que le traitement batch initial était plus rapide, moins cher et plus fiable Là où le batch COBOL traitait 2.4 millions d'enregistrements en 11 minutes, le système distribué modernisé à base de message queues, retries et Kubernetes prenait 47 minutes et tombait sous la charge COBOL brille par ses caractéristiques conçues spécifiquement pour la finance comme le calcul décimal précis sans floating point errors et l'absence totale d'overhead réseau, de conteneurs ou de cold starts Rappel que distribuer un système multiplie les points de défaillance silencieux et complexifie la gestion de la cohérence transactionnelle par rapport à une exécution locale séquentielle Une invitation à se demander si les projets de décomposition en microservices apportent réellement un gain de performance de bout en bout pour l'utilisateur final ou s'ils optimisent seulement le diagramme d'architecture Méthodologies Ma meilleure question d'entretien Spring beaufume.fr/articles/spring-interview Florian beaufumé partage sa question d'entretien favorite pour évaluer des développeurs Spring de niveau intermédiaire à avancé : "Que pouvez-vous me dire sur le paramètre spring.jpa.open-in-view ?". Ce paramètre détermine l'activation du pattern Open Session In View (OSIV) qui, lorsqu'il est à true (la valeur par défaut dans Spring Boot), maintient l'un EntityManager JPA ouvert durant toute la requête HTTP. Si l'OSIV facilite le développement en évitant les fameuses LazyInitializationException lors de la sérialisation des entités en JSON, il pose d'importants problèmes de performance en provoquant des requêtes SQL non maîtrisées (comme le problème du N+1 select) en dehors de la couche service. Maintenir l'OSIV actif augmente également le temps de rétention des connexions au sein du pool de la base de données, limitant la scalabilité de l'application. La recommandation est de désactiver ce comportement en le positionnant à false, et de gérer explicitement le chargement des données requises au sein des transactions (via des DTOs, des requêtes JOIN FETCH ou des Entity Graphs) pour garder le contrôle sur les accès à la base de données. 10 points à retenir du rapport AI Engineering 2026 : The Acceleration Whiplash faros.ai/blog/ai-acceleration-whiplash-takeaways L'IA a franchi un cap et est devenue l'auteur principal du code : le taux d'acceptation du code généré est passé de 20% à 60% dans les équipes étudiées par Faros AI. La vélocité métier est bien réelle, avec une augmentation de 66% des epics livrées et une hausse de 33,7% du throughput des tâches par développeur. Ce volume cache un code churn massif (+861%), ce qui signifie qu'une quantité énorme de code est supprimée ou remplacée peu après avoir été ajoutée. La qualité en aval se dégrade fortement : les bugs par développeur ont augmenté de 54% et le nombre d'incidents par pull request a explosé de 242,7%. Le processus de code review est complètement saturé, entraînant un temps médian de relecture multiplié par cinq et une augmentation de 31,3% des PRs mergées sans aucune revue. Le système repose de plus en plus sur les développeurs seniors qui subissent une "senior engineer tax", devant relire un volume insoutenable de code à l'apparence correcte mais structurellement fragile. Contrairement à certaines hypothèses récentes de DORA, une forte maturité DevOps ne protège pas les entreprises contre cette détérioration ; le "Acceleration Whiplash" frappe de la même manière les équipes très performantes. En résumé, les outils d'IA inondent les pipelines de livraison avec un volume de code pensé pour un rythme machine, alors que les systèmes de vérification reposent toujours sur un rythme de validation humain. Loi, société et organisation Le coût d'une equipe d'engineering qui ne sait plus ce qu'elle fait dans un contexte d'augmentation de coût des coding agents https://freedium-mirror.cfd/@developer_programmer/i-spent-47-000-on-claude-code-in-90-[…]-asked-me-one-question-and-i-couldnt-answer-it-af3b203f81bb Une équipe de 8 ingénieurs a vu sa vélocité de développement exploser en utilisant Claude Code de manière intensive, jusqu'à recevoir une facture d'API salée de 47 213 $ pour seulement trois mois d'utilisation. Face à cette dépense, la question piège du CTO n'était pas sur le montant, mais sur la dépendance : "Si nous arrêtions Claude Code demain, combien de temps faudrait-il pour que notre vélocité revienne à son niveau initial ?". L'auteur s'est rendu compte qu'il était incapable de répondre car son équipe, en particulier les profils juniors, avait commencé à perdre l'habitude de concevoir et d'implémenter des fonctionnalités complexes sans l'aide permanente d'un agent. Le deuxième risque stratégique soulevé est celui de la dépendance tarifaire et du vendor lock-in : si l'outil devient une infrastructure indispensable au quotidien, l'entreprise perd tout pouvoir de négociation face aux augmentations de prix de l'éditeur d'IA. Pour éviter que l'IA ne devienne une béquille qui atrophie les compétences de l'équipe, l'article suggère de poser des limites budgétaires strictes, d'organiser régulièrement des sprints sans IA ("AI-free sprints") et de concevoir des processus de développement portables. Retour de Nicolas Delsaux sur jqwik qui donne une perspective plus complète concernant jqwik, il me semble que vous oubliez (comme tous les gens qui parlent de LLM dans "l'industrie") que l'auteur n'a pas fait ça juste pour faire chier le monde, mais parce que ces outils ont des externalités incroyablement négatives, ce dont l'auteur s'explique dans son blog (blog.johanneslink.net/2026/06/09/the-jqwik-anti-ai-affair) Vous oubliez également de signaler que le ticket (github.com/jqwik-team/jqwik/issues/708) par lequel un utilisateur se plaint de cette fonctionnalité a été écrit par un agent. N'oubliez pas non plus que l'enthousiasme pour ces technologies n'est en fait pas universel, et que ces technologies sont loin d'être inévitables (les gains de vitesse ne sont, d'après circle CI - circleci.com/resources/2026-state-of-software-delivery, pas des gains de productivité ) OkHttp, Okio, Retrofit et SQLDelight rejoignent Commonhaus ! commonhaus.org/activity/315.html La fondation Commonhaus, via une publication de Andres Almiray, annonce l'arrivée de quatre projets majeurs de l'écosystème Java et Kotlin : OkHttp, Okio, Retrofit et SQLDelight. Ces projets, initialement créés chez Square (devenu Block), sont désormais regroupés et gérés sous la bannière lysine.dev au sein de la fondation. Jesse Wilson et Jake Wharton, créateurs et mainteneurs historiques de ces outils, rejoignent Commonhaus en tant que leaders de lysine.dev. Suite à leur départ de Block, ils expliquent avoir choisi Commonhaus pour offrir à leur immense communauté d'utilisateurs un cadre de gouvernance pérenne, stable et digne de confiance. Conférences La liste des conférences provenant de Developers Conferences Agenda/List par Aurélie Vache et contributeurs : 28-30 août 2026 : State of the Map - Champs-sur-Marne (France) 4 septembre 2026 : JUG Summer Camp 2026 - La Rochelle (France) 10-11 septembre 2026 : Nantes Craft - Nantes (France) 17 septembre 2026 : dotAI - Paris (France) 17-18 septembre 2026 : API Platform Conference 2026 - Lille (France) 18 septembre 2026 : WordCamp Bretagne - Rennes (France) 18 septembre 2026 : dotJS - Paris (France) 18 septembre 2026 : WordCamp Bretagne - Rennes (France) 22 septembre 2026 : Salon Data 2026 - Nantes (France) 22-23 septembre 2026 : Agile en Seine & IA 2026 - Paris (France) 24 septembre 2026 : aMP Day Montpellier 2026 - Montpellier (France) 24 septembre 2026 : OWASP AppSec Days France 2026 - Paris (France) 24 septembre 2026 : PlatformCon Paris - Paris (France) 24 septembre 2026 : React Native Connection 2026 - Paris (France) 24-26 septembre 2026 : Paris Web 2026 - Paris (France) 25 septembre 2026 : SAP Inside Track Paris 2026 - Paris (France) 28-29 septembre 2026 : 4th Tech Summit on AI & Robotics - Paris (France) & Online 1 octobre 2026 : WAX 2026 - Marseille (France) 1-2 octobre 2026 : Volcamp - Clermont-Ferrand (France) 2 octobre 2026 : DevFest Perros-Guirec 2026 - Perros-Guirec (France) 5-9 octobre 2026 : Devoxx Belgium - Antwerp (Belgium) 8-9 octobre 2026 : Forum PHP 2026 - Marne-la-Vallée (France) 12 octobre 2026 : Dev With AI - Paris (France) 22-23 octobre 2026 : Agile Tour Bordeaux 2026 - Bordeaux (France) 26 octobre 2026 : Agile Tour Montpellier - Montpellier (France) 27-29 octobre 2026 : Directions EMEA 2026 - Paris (France) 29-30 octobre 2026 : Campus Agile Grenoble - Grenoble (France) 29-30 octobre 2026 : BDX I/O 2026 - Bordeaux (France) 29-30 octobre 2026 : Agile Tour Nantais 2026 - Nantes (France) 29 octobre 2026-1 novembre 2026 : Pycon FR - Biarritz (France) 30 octobre 2026 : Cloud Nord 2026 - Lille (France) 4-5 novembre 2026 : Devoxx Morocco - Casablanca (Morocco) 14-15 novembre 2026 : Capitole du Libre - Toulouse (France) 19 novembre 2026 : DevFest Toulouse 2026 - Toulouse (France) 19 novembre 2026 : Agile Laval 2026 - Laval (France) 19 novembre 2026 : OVHcloud Summit - Paris (France) 19 novembre 2026 : Codeurs en Seine - Rouen (France) 27 novembre 2026 : DevFest Paris 2026 - Paris (France) 1-3 décembre 2026 : Apidays Paris - Paris (France) 2-3 décembre 2026 : Cloud Native AI Summit Europe - Paris (France) 4 décembre 2026 : DevFest Lyon 2026 - Lyon (France) 4 décembre 2026 : DevFest Dijon 2026 - Dijon (France) 9-10 décembre 2026 : OpenSource Expérience - Paris (France) 9-10 décembre 2026 : DevOps REX - Paris (France) 10 décembre 2026 : KCD Provence - Aix-en-Provence (France) 10 décembre 2026 : DevCon 28 : sécurité | post-quantique | hacking édition 2027 - Paris (France) 14-16 janvier 2027 : SnowCamp 2027 - Grenoble (France) 7-9 avril 2027 : Devoxx France 2027 - Paris (France) 3 juin 2027 : Cloud Native Days France 2027 - Paris (France) Nous contacter Pour réagir à cet épisode, venez discuter sur le groupe Google https://groups.google.com/group/lescastcodeurs Contactez-nous via X/twitter https://twitter.com/lescastcodeurs ou Bluesky https://bsky.app/profile/lescastcodeurs.com Faire un crowdcast ou une crowdquestion Soutenez Les Cast Codeurs sur Patreon https://www.patreon.com/LesCastCodeurs Tous les épisodes et toutes les infos sur https://lescastcodeurs.com/

In this episode, Conor and Bryce chat about CityStrides, graph algorithms, GPT 5.6 Solver, and more!Link to Episode 295 on WebsiteDiscuss this episode, leave a comment, or ask a question (on GitHub)SocialsADSP: The Podcast: TwitterConor Hoekstra: LinkTree / BioBryce Adelstein Lelbach: TwitterShow NotesDate Recorded: 2026-07-13Date Released: 2026-07-17ADSP Episode 149: CityStrides.com, Graph Algorithms and More!Carlo de Lorenzi: Toronto runner with terminal brain cancer runs every street in the cityCarlo's fundraiser for Community Music Schools of Toronto FoundationCityStrides.comOpen Street Mapscity-strides-hacking GitHub RepoHamiltonian PathEulerian PathEpisode 220: Graph Algorithms & 7 Bridges of KönigsbergNVIDIA cuoptIntro Song InfoMiss You by Sarah Jansen https://soundcloud.com/sarahjansenmusicCreative Commons — Attribution 3.0 Unported — CC BY 3.0Free Download / Stream: http://bit.ly/l-miss-youMusic promoted by Audio Library https://youtu.be/iYYxnasvfx8

The Pirates have a very little chance to win the NL Central because the Brewers have created such a gap. But what about a wild card spot? Fan Graphs gives them the best chance of any team outside looking in to make the postseason. The site gives the Pirates just under a 42% chance to make it to October. Are they better than the Cardinals, Diamondbacks and Marlins? Steelers insider Ray Fittipaldo from the PG joined the show. Jon Gruden continues to praise the work of Will Howard as we have yet to see him in a pro game. Ray thinks it's more of the same and added what we have heard so many times from Mike McCarthy so far. Per usual, Ray said the offense in the preseason will be vanilla, so it could come down to what we see in practice. Ray reacted to the idea that 6 Steelers, including Keeanu Benton, would fetch a 1st round pick in return if traded. Ray expects a bounce back year from TJ Watt and thinks he will have a 12-sack season. When will the Joey Porter Jr. contract buzz start to pick up again and when will that deal get done? How long will it take for Max Iheanachor to become the starter at right tackle? Poni took us on a ride through some football preview books.

Read the article here: https://journals.sagepub.com/doi/full/10.1177/30494826261455199

Many corporate marketing strategies face dropping conversion tracking metrics because they rely on closed ad-tech vendors that introduce platform bias into campaign attribution. Mathieu Roche, co-founder and CEO of ID5, breaks down how to construct neutral identity infrastructure to protect cross-channel addressability and secure reliable measurement.He shares his operational playbook for navigating the shift toward agentic advertising, altering data graphs to track automated consumer assistants, and using internal team hackathons to automate routine sales and CRM processes.Key tactical themes covered:Building neutral identity channels to protect unbiased attribution tracking.Modifying data graphs to track verified virtual agents across the purchase funnel.Shifting identity tech priorities from baseline audience targeting to outcome measurement.Linking automated AI transactions directly back to programmatic ad investments.Utilizing structured team sprints to automate routine operational follow-up tasks.Mathieu Roche is the co-founder and CEO of ID5, directing un-siloed identity architecture and multi-platform validation systems for leading digital publishers.Connect with our guest:Follow Mathieu Roche on LinkedIn: https://www.linkedin.com/in/mathieuroche/Explore ID5 Solutions: http://id5.ioOptimize Performance Advertising with Strike Social: https://strikesocial.com/guaranteed-paid-social-media-ads-outcomes/Connect with Host Dylan Conroy: https://www.linkedin.com/in/dylanconroy/

In this episode, Jason and Jordan kick things off with post–Fourth of July stories, from eight-inch mortars and a professional fireworks barge to the unexpected lesson in staffing, leadership, and broken promises that came with a canceled show. They pivot into the green industry and dig into how Jason is onboarding virtual assistants (VAs) to help track lead flow, build out data-driven reporting, and improve budgeting and cash flow visibility. The conversation gets candid around switching to weekly payroll, the hidden cost of payroll advances, and where to draw the line between supporting employees and becoming "the bank." They compare live answering options like Ruby versus AI-backed tools such as RingCentral, and explore how VAs, phone systems, and SingleOps can work together to tighten operations. Jason shares how he's thinking about social media ROI, content strategy with Sky Palm Studios, and why long-term trust-building may beat pure ad spend. Jordan closes with two powerful branding and values stories: how 25 years of Empire Today ads led him to a no-hesitation buying decision, and how an HOA dust-up over flying the American flag at the community boat ramp turned into a real-world case study in leadership, perception, and standing your ground. Connect with Jason and Jordan:

MUKTZAH 5 GREAT GRAPHS

Dan Adams is a 13-year Microsoft 365 and SharePoint veteran who joined Andrew to talk about the often misunderstood world of SharePoint, the shift to Microsoft Graph, and his custom PowerShell module built which can assess and benchmark M365 environments. Dan breaks down why SharePoint gets such a bad reputation (spoiler: it's usually about who built it, not the platform), explains why "everything in M365 is SharePoint" isn't just a meme, and digs into how the Microsoft Graph API is changing the way admins interact with the platform. He also shares his experience going from longtime podcast listener to first-time guest, and closes with some honest advice about reaching out to people in the community. Key Takeaways: SharePoint's bad reputation often comes from poor initial architecture, not the platform itself. Getting the information structure right at the start has downstream effects on everything from Copilot to Purview to legal compliance, and PowerShell automation is only as useful as the metadata you've set up to work with. The Microsoft Graph API is consolidating how everything in M365 is accessed. Where older APIs like the SharePoint client-side object model might count a batch of 100 items as 100 separate API calls, Graph treats that same batch as a single call, which is a meaningful difference for performance and rate limiting. Dan's custom PowerShell module (SP'r Smash Bros Automation) automates M365 permission extraction to produce security assessments against CIS benchmarks and Microsoft Secure Score. Built as a practical tool for consultants and admins, it delivers a fast, standardized baseline of a tenant's security posture, featuring an optional AI sidecar to instantly generate personalized remediation plans. Guest Bio: Dan Adams is a Microsoft 365 and SharePoint Architect with 13 years of M365 consulting experience. Leveraging deep SharePoint and strategic information architecture expertise, he has led teams to deliver over 40 Fortune 500 intranets across industries ranging from healthcare to the NFL. He is the architect behind multiple award-winning portals, including two Ragan "Best Overall Intranet" winners. Dan is also an AI enthusiast, an advanced PowerShell expert, and the creator of the custom "SP'r Smash Bros Automation" module for Microsoft 365. Resource Links: Dan Adams on LinkedIn: https://www.linkedin.com/in/dan-adams-10887650/ Dan Adams on Github: https://github.com/sprsmashbrosautomation Connect with Andrew: https://andrewpla.tech/links PnP PowerShell: https://pnp.github.io/powershell/ PDQ Discord Community: https://discord.gg/pdq The PowerShell Podcast on YouTube: https://youtu.be/OLsTaKC9GmM PowerShell Wednesday Playlist: https://www.youtube.com/watch?v=XQT8lrn8hhU&list=PL1mL90yFExsix-L0havb8SbZXoYRPol0B

Join us as Du'An digs into the real mechanics of running AI locally and in production - from GPU memory math to multi-agent architectures, observability, and the economics of self-hosted inference. Du'An walks through how model weights and KV cache compete for GPU memory, why continuous batching matters when you have more than a handful of users, and how agent architectures like single-agent, workflow, graph, swarm, and supervisor patterns each solve different problems. You will learn how to instrument your agents with Langfuse for observability and cost tracking, when to use Ollama versus vLLM, how prompt caching can cut provider costs by up to 75%, and why GPUs should never sit idle. Episode two of three - the next episode covers deploying at scale. Timestamps 0:00 Welcome & Introduction 1:47 Du'An's New Role at Akamai Cloud 3:10 Data Privacy and the Case for Self-Hosted AI 7:21 Anthropic and OpenAI as the New Cloud Layer 12:48 Local Models for Specific Use Cases - Cancer Detection Example 15:02 GPU Memory Math - Weights, KV Cache, and Context Windows 19:32 Continuous Batching and GPU Time Slicing 20:03 Observability with Langfuse - Live Demo 27:44 Agent Architectures - Single Agent, Workflow, Graph, Swarm, Supervisor 36:36 Token Economics, Prompt Caching, and GPU Cost Planning 45:32 Ollama vs vLLM - Prototyping vs Production How to find Du'An: https://duanlightfoot.com https://www.linkedin.com/in/duanlightfoot/ Links from the show: https://langfuse.com/ https://github.com/akamai-developers/akamai-workshop-solution-architect-agent https://amzn.to/4bvHn1p https://vllm.ai/

What if the biggest barrier to successful AI isn't the model itself, but the lack of context behind every decision your teams make? As AI agents become more capable, how do organisations ensure they understand the people, projects, documentation, and history that shape real work? In this episode of Tech Talks Daily, recorded at Team '26, I'm joined by Taroon Mandhana, CTO of AI and Teamwork at Atlassian. His responsibilities span engineering for products including Jira, Confluence, Loom, and Trello, alongside the company's AI strategy and the development of Rovo. Our conversation explores why Atlassian believes AI should become a teammate rather than simply another chatbot. Taroon explains why enterprise context has become one of the most valuable assets in the AI era. While today's foundation models continue to improve at an incredible pace, they still lack the organisational knowledge that human teams naturally accumulate over time. Atlassian's Teamwork Graph aims to bridge that gap by connecting people, projects, documentation, code, goals, and conversations into a living knowledge network that AI agents can use to produce more accurate, relevant outcomes. We also discuss why Atlassian has chosen an open approach, making its Teamwork Graph available through technologies such as MCP rather than limiting it to its own AI products. Taroon shares why interoperability will become increasingly important as businesses adopt multiple AI platforms and why organisations should be free to use the agents that best suit their needs without losing access to valuable business context. Another fascinating part of our conversation focuses on how Atlassian's own engineering teams are changing the way they build software. Smaller teams, tighter collaboration, AI-assisted development, and faster iteration cycles are allowing products to move from concept to release in weeks rather than months. Taroon explains how AI is changing both software development and the structure of engineering teams themselves. We also examine where AI should take ownership of work inside platforms like Jira, where human judgement remains essential, and why successful organisations are treating AI adoption as an ongoing product journey rather than a one-time technology deployment. If your business is looking beyond isolated AI experiments and wondering how to build AI into everyday work, this conversation offers valuable insight into the role context, openness, and organisational change will play in the next generation of enterprise software. As AI becomes part of every workflow, what do you think will become the real competitive advantage: better models, or better organisational knowledge?

For law firms, artificial intelligence has often arrived as a choice between speed and control. Stephen Costigan, founder of Atlas AI, argues that choice deserves a rethink. In this episode of The Geek in Review, we speak with Costigan about private legal AI infrastructure, knowledge graphs, and why a firm's internal work product may become its most valuable long-term asset.Atlas AI focuses on turning documents, matter history, precedents, clauses, parties, and obligations into a curated legal knowledge graph inside a firm's own environment. Costigan contrasts this approach with standard vector search and retrieval systems, which find text with similar language but often lack context around clients, matters, entities, and relationships. A knowledge graph offers structure, linking people, documents, clauses, and legal concepts in ways closer to how lawyers understand their work.The conversation also explores data quality, a subject with enough baggage to fill a records room. Costigan argues firms no longer need year-long cleanup projects before seeing results. Agent-led curation, entity extraction, duplicate resolution, and ontology mapping reduce much of the manual sorting traditionally associated with knowledge management. Human judgment still matters, especially around practice-area vocabularies and lower-confidence results, but the machines get assigned more of the janitorial work.Security and governance sit at the center of Costigan's model. Rather than asking firms to trust a vendor's assurances around privileged data, Atlas AI runs within a firm's Azure environment, under firm-controlled keys and policies. Costigan frames this as a shift from confidentiality as a contractual promise to confidentiality as an architectural decision. For legal organizations handling sensitive client information, the location of data, embeddings, audit trails, and model interactions matters as much as the interface lawyers see on screen.Looking ahead, Costigan predicts a divide between firms renting generic AI tools and firms building durable knowledge infrastructure from their own experience. As routine drafting, diligence, and review work compress, firms with structured and reusable internal intelligence may productize expertise, offer new fixed-fee services, and rely less heavily on traditional leverage models. The future question, Costigan suggests, will not center on which AI tool sits on a lawyer's desktop. The bigger question will ask who owns the knowledge behind the work.Listen on mobile platforms:  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Apple Podcasts⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ |  ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Spotify⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ | ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠YouTube⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ | ⁠Substack⁠[Special Thanks to ⁠⁠Legal Technology Hub⁠⁠ for their sponsoring this episode.]⁠⁠⁠⁠⁠Email: geekinreviewpodcast@gmail.comMusic: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Jerry David DeCicca⁠⁠⁠⁠⁠⁠⁠⁠⁠Transcript:

Text us your thoughts!In this quick follow-up to last week's episode, we invited our guest (Deborah) to model a short, fun classroom debate. In just a few minutes, you can hear a sample debate that captures the spirit of productive mathematical argumentation. Tune in for a rapid-fire glimpse of what these debates can look like in action as we ask the question:Which is better: bar graphs or pie charts?You can find Dr. Deborah Peart Crayton on social media: @mathersgonnamath Check out her website: MathersGonnaMath.comListened to the episode? Now, it's your turn to share! Find us on Social Media: @DebateMath to share your thoughts.Don't forget to check out the video version of this podcast on our YouTube channel!Keep up with all the latest info by following @DebateMath or going to debatemath.com. Follow us @Rob_Baier & @cluzniak. And don't forget to rate and review us on Apple Podcasts!

Here are the top five insurance products you should be looking to sell during the Medicare lock-in period. The insurance industry never stops moving. Make sure you know the products to keep your business moving along with it.   Read the text version   Get Connected:

Aji and Joël join forces to discuss graph and tree structures, and their connection to the emergent properties, attributes and qualities you can find from a largely connected group of data. Joël dives into their recent graphs and tree work through a contracting system, whilst Aji looks back at when he previously tried to serialise a graph or tree to a database. — Watch Joël's Blue Ridge Ruby talk here, or Matheus' Ruby Internal talk from last year here. There's still time to secure your place at thoughtbot's upcoming UK meet ups over the next month. London Tech Leader Meetup - Tuesday June 23rd Brighton Tech Leader Meetup - Wednesday June 24th Brighton Ruby - Thursday June 25th Evolve - Friday June 26th Your hosts for this episode have been thoughtbot's own Joël Quenneville and Aji Slater. If you would like to support the show, head over to our GitHub page, or check out our website. Got a question or comment about the show? Write to our hosts: hosts@bikeshed.fm This has been a thoughtbot podcast. Stay up to date by following us on social media - YouTube - LinkedIn - Mastodon - BlueSky © 2026 thoughtbot, inc.

Thanks so much for listening! For the complete show notes, links, and comments, please visit The Grey NATO Show Notes for this episode:https://thegreynato.substack.com/p/380-tgraph2The Grey NATO is a listener-supported podcast. If you'd like to support the show, which includes a variety of possible benefits, including additional episodes, access to the TGN Crew Slack, and even a TGN edition grey NATO, please visit the link below.Support the show

After decades of decline, many church leaders believe that religious life is on the upswing as some younger Americans flock to Christianity — including Vice President JD Vance, whose new book on his Catholic conversion drops this week. But the fuller picture is more complicated. Coming up, we'll talk to religion reporters and a church leader about what may be driving this shift, and what its lasting impacts could be. Guests: Michael O'Loughlin, executive editor, National Catholic Reporter; O'Loughlin has covered the Catholic church for both the Boston Globe and Crux; author, "Hidden Mercy: AIDS, Catholics and the Untold Stories of Compassion in the Face of Fear" Lauren Jackson, deputy editorial director for newsletters and the host of “Believing," The New York Times Ryan Burge, professor of practice at the John C. Danforth Center, Washington University; author, “Graphs about Religion” Danté Stewart, author, “Shoutin' in the Fire: An American Epistle;” an ordained minister at Tabernacle Baptist Church in Augusta, Ga. Learn more about your ad choices. Visit megaphone.fm/adchoices

Media Watch 2026 Episode 19: Musk's army; You're havin' a graph

Jure Leskovec, Professor of Computer Science at Stanford University and Chief Scientist at Kumo.ai, speaks with host Sriram Panyam about relational and graph language models and their transformative impact on enterprise decision-making and predictive modeling. Jure begins by establishing the critical importance of predictive modeling across industries - from fraud detection in financial institutions to customer churn prediction, lifetime value estimation, product recommendations, and healthcare risk assessment. He notes that while AI has made remarkable advances in natural language understanding and computer vision, predictive modeling over enterprise operational data stored in relational databases has been largely left behind, still relying on 30-year-old machine learning approaches that are expensive, time-consuming, and require manual feature engineering. His proposed solution to the fundamental problem with current approaches is relational deep learning and relational transformers. The discussion explores how this approach differs from traditional graph neural networks (GNNs), which Jure pioneered and deployed successfully at Pinterest. Jure concludes with practical guidance for software engineers and data scientists interested in exploring this technology.

In this episode of the Crazy Wisdom Podcast, host Stewart Alsop sits down with Joshua Bate, founder of Bonfires.ai and DeciWorld, for a wide-ranging conversation covering knowledge management, graph technology, ontologies, decentralized science, and the future of how humans organize and share information. They break down the differences between personal and enterprise knowledge management, explore why flat ontological graphs may be the key to making diverse knowledge bases interoperable, and get into why traditional RAG systems break down at scale and how graph RAG offers a more principled solution. The conversation expands into the philosophy of categorization, the slow death of basic "gentleman science" under institutional pressures, and how decentralized protocols might restore a kind of mycelial knowledge network connecting small groups of researchers, enthusiasts, and communities — much like the original spirit of the encyclopedia before it was co-opted by institutions. You can learn more about Joshua's work at bonfires.ai and deci.world or follow him on X at @Bonfiresai and @DeSciWorld.Timestamps00:00 - Stewart introduces Joshua Bate, founder of Bonfires.ai, discussing personal versus enterprise knowledge management and their fundamental differences at scale.05:00 - Joshua explains ontologies as classifiers for knowledge structures, describing their two-year search for a perfect ontology and ultimately building a flat, ontology-less graph protocol.10:00 - Stewart connects categorization to shamanic practice and intercategorical theory, noting how major companies like Netflix and Yahoo built graph-based ontologies while the discipline remains underappreciated philosophically.15:00 - Joshua traces Bonfires origins through decentralized science, explaining how NFT community excitement inspired redirecting capital toward funding unconventional researchers locked out of institutional systems.20:00 - Joshua describes building federated knowledge networks through hackathons and conferences, comparing the vision to what Wikipedia could have been with decentralized incentive structures.25:00 - Discussion shifts toward inevitable collapse of rigid scientific institutions, debating patchwork age theory, nation-state fragmentation, and rhizomatic versus arboreal knowledge structures.30:00 - Joshua articulates the mycelial network vision, enabling direct cross-cultural information access where individuals control their own narrative lens, warning against collective we thinking and authoritarianism.Key Insights1. Knowledge management exists on a spectrum from personal to enterprise, but the founder of Bonfires argues this split is artificial. He believes knowledge itself does not respect those boundaries, and that small groups, researchers, hobbyists, and large institutions all possess knowledge that can and should interoperate with each other.2. After two and a half years of searching for the perfect ontology to structure their knowledge graph, the team concluded that no perfect ontology exists. Their solution was to build the flattest possible graph structure with only events, entities, and edges, creating a base layer others can build specialized ontologies on top of.3. Graph-based knowledge systems are more efficient than traditional databases for AI traversal because once a graph is computed, it is relatively free to query. Graph RAG combines the discovery power of vector search with the structured precision of graph traversal, solving many hallucination problems associated with standard retrieval augmented generation.4. Basic scientific research, the soil from which applied discoveries grow, is deteriorating because institutional funding structures only reward commercially viable outcomes. The founder built his platform partly to redirect community-driven capital toward researchers who are doing important work without institutional support.5. The institutionalization of science has historically blocked the open exchange of ideas that drove the original scientific revolution. The human spirit for open inquiry has not changed, but people cannot pursue it without financial support, and building decentralized infrastructure could restore that possibility.6. A federated knowledge network would allow individuals to access information from any contributor and filter it through their own preferred lens, rather than receiving information pre-filtered by centralized platforms. This represents a form of information symmetry similar to how mycelial networks distribute nutrients across a forest.7. The concern is not whether current scientific and governmental institutions will change but in what direction the rebuilding goes. Those capitalizing on the transition carry the same incentives as the previous era, which risks reproducing the same problems inside new structures.

This week on “Jesuitical,” Ashley and Zac speak with Ryan Burge, author of the “Graphs about Religion” Substack and the new book, The Vanishing Church: How the Hollowing Out of Moderate Congregations Is Hurting Democracy, Faith, and Us. They discuss the polarization of U.S. Christianity and the supposed Gen-Z “religious revival.” In Signs of the Times, Ashley and Zac discuss some highlights from Pope Leo's trip to Africa; what Pope Leo called the not-exactly-accurate media narrative around him and President Trump; and the first anniversary of Pope Francis' death. 00:00 A Gen-Z religious revival? 3:38 Highlights of Pope Leo's trip to Africa 10:05 VP Vance questions Pope Leo's theology 20:37 Remembering Pope Francis 22:50 Moderate Christianity is vanishing 25:49 U.S. religion is coded "conservative" 34:54 Catholic demographic trends 37:15 Political implications 40:53 Are young people going back to church? 48:18 Winner churches 52:56 Gen-Z religious trads 1:04:08 Faith Sharing: Pope Francis' humble tomb Links: ⁠Order Ryan's book, The Vanishing Church⁠ ⁠Graphs about Religion⁠ ⁠Pope Leo walks in the footsteps of St. Augustine in Hippo⁠ ⁠Pope Leo denounces those who use the name God for military gain⁠ ⁠Pope Leo named one of Time magazine's ‘100 Most Influential People of 2026'⁠ ⁠Pope Leo remembers ‘the great gift' of Pope Francis on the first anniversary of his death⁠ You can follow us on ⁠X⁠ and on ⁠Instagram⁠ @jesuiticalshow.   You can find us on Facebook at ⁠facebook.com/groups/jesuitical.⁠  Please consider supporting Jesuitical by becoming a digital subscriber to America magazine at ⁠americamagazine.org/subscribe⁠ Learn more about your ad choices. Visit megaphone.fm/adchoices

What I Did as a Child – Co robiłem jako dziecko "Co robiłem jako dziecko" means "what I did as a child," and in this nostalgic micro-lesson you'll say it like you're flipping through old photo albums with your Polish grandmother. First you hear the phrase at native speed, then slowed down so you can master the rolling "r" and the soft "dziecko." We drop it into three memory-lane-ready sentences: – "Kiedy byłem mały…" (When I was little…) – "Lubiłem się bawić." (I liked to play.) – "To było dawno temu." (That was a long time ago.) Repeat-along track included—perfect while you reminisce or share your own childhood stories. Challenge: Tell us in the comments what YOU did as a child—reply in Polish and Ania might sing your answer in the next episode     What we discussed:   0:00 Welcome & QR Code 0:45 "Dziecko" - The Polish Word for Child 1:30 Childhood Bedroom Memories 2:30 Sports & Experiments 3:30 School & Newspapers 4:30 University & Opinions 5:30 Wishes & Dashboards 6:30 Stars & Classrooms 7:30 Time & Dance 8:30 Building & Creating 9:30 Television & Media 10:30 Graphs & Grades 11:30 Comics & Parks 12:30 Photos & Memories 13:30 Games & Play 14:30 Suggestions & Ideas 15:30 Early Jobs & Studies 16:30 Army & Activities 17:30 Toys & Cars 18:30 Sunset & Evening Play 19:30 Mom's Voice & Family 20:30 Growing Up & Yesterday 21:30 Fitness & Sports 22:30 Comfort & Balm 23:30 National & Sharing 24:30 Adult Life & Planning 25:30 Vision & Yesterday's Story 26:30 Short Stories & Patches 27:30 Royal Games & Health 28:30 Memories & Reports 29:30 Calls & Moods 30:30 Sports & Chores 31:30 Swimming & Memories 32:30 Web & Changes 33:30 Books & People 34:30 School Initiatives 35:30 Travel & Problems 36:30 Chats & Opinions 37:30 Goals & School Sheets 38:30 Status & Movies 39:30 Platforms & QR Code 40:30 Your Turn to Practice!   English Polish Pronunciation Guide child dziecko dzyeh-tsoh childhood dzieciństwo dzyeh-cheen-stvo memory wspomnienie vspo-mnyeh-nyeh to remember pamiętać pah-myeh-tahch to forget zapomnieć zah-pom-nyehch to play bawić się / grać bah-veech sheh / grahch toy zabawka zah-bahf-kah game gra grah school szkoła shkoh-wah teacher nauczyciel / nauczycielka now-chi-tyel / now-chi-tyel-kah friend przyjaciel / przyjaciółka psi-ya-chyel / psi-ya-choow-kah family rodzina roh-jee-nah mom mama mah-mah dad tata tah-tah brother brat braht sister siostra syoh-strah home dom dohm room pokój poh-kooy bed łóżko woo-shkoh to sleep spać spahch to eat jeść yeshch favorite food ulubione jedzenie oo-loo-byoh-neh yeh-dzeh-nyeh sport sport sport football/soccer piłka nożna peew-kah nozh-nah ping pong ping pong ping pong swimming pływanie pwih-vah-nyeh bicycle rower roh-ver to read czytać chi-tahch book książka kyohnsh-kah comic book komiks koh-meeks TV telewizja teh-leh-veez-yah cartoon bajka bahy-kah video game gra wideo / gra komputerowa grah vyeh-deh-oh / grah kom-poo-teh-roh-vah to sing śpiewać shpyeh-vahch song piosenka pyoh-sen-kah to dance tańczyć tahyn-chich party impreza eem-preh-zah birthday urodziny oo-roh-jee-ni holiday wakacje / ferie vah-kah-tsyeh / feh-ryeh summer lato lah-toh winter zima zee-mah to grow up dorastać doh-rah-stahch adult dorosły / dorosła doh-roh-swi / doh-roh-swah young młody / młoda mwoh-di / mwoh-dah old stary / stara stah-ri / stah-rah yesterday wczoraj fchoh-rah-y today dziś dzeesh tomorrow jutro yoo-troh long ago dawno temu dahv-noh teh-moo always zawsze zahf-sheh never nigdy neeg-di sometimes czasami chah-sah-mee often często chen-stoh