Podcasts about cpu gpu

Build 2026 is underway in San Francisco this week, and it started with a big, overly-long keynote as always. And Computex is this week, too. There's a lot going on, and some of it is fascinating. Plus, WWDC is next week because you cannot relax. Also, Microsoft GA's WinApp CLI, announces the Windows Platform Skills plug-in for native app creation, and you're not going to believe what Paul did next. OK, you will believe itBuild + Computex = OHMYGODOHMYGODOHMYGOD NVIDIA finally announces Arm-based N1X as the RTX Spark RTX Spark is an Arm-based portable workstation chip for Windows 11 Microsoft announces Surface Laptop Ultra - It and other RTX Spark-based PCs will appear in late 2026 Some of this leaked earlier, including a lower-end N1 chipset Microsoft continues to optimize and evolve Windows 11 for developers Windows Developer Configuration, Windows Developer Skills + WinApp CLI, Terminal, more Linux, and more on-device ("unmetered") AI - Tied to this, Copilot+ PC features are coming to more PCs, with CPU/GPU support - this, plus the RTX Spark stuff hints at answers to some obvious questions but there's nothing concrete from Microsoft Microsoft Edge is getting three new on-AI features Scout is a personal work agent powered by OpenClaw GitHub Copilot app arrives on desktop for your agentic coding and management needs Microsoft AI announces seven new foundation models Stevie Bathiche is back, baby! And he's talking about those AI app structures and how they've led to Project Solara Windows Microsoft discusses the progress it's made on Windows 11 pain points You can now test the new Start menu in Experimental - Paul did so along with the new Taskbar Qualcomm announces low-cost Snapdragon C for $300+ PCs to take on MacBook Neo And Acer is the first to announce a Snapdragon C laptop New Surface Pro with Snapdragon X2 leaks for June release (!) Dell XPS 13 is coming soon with Intel Wildcat (also to take on MacBook Neo) Dell revenues are through the roof, but not because of PCs HP revenues are up, and it is because of PCs AI and dev Anthropic gets a new valuation exceeding OpenAI and then it files for an IPO OpenAI adjusts GPT5.5-Instant for less sucking-up and releases computer use in Codex on Windows Flutter takes the lead on Flutter desktop development XBOX and gaming Asha Sharma says you can't please everyone and then immediately jumps the shark trying to please everyone XBOX delays Fable reboot because of GTA VI New titles coming to Game Pass in early June across platforms XBOX starts early testing of new console features ASUS announces ROG Xbox Ally X20 with OLED display and XReal R1 glasses Intel announces Arc G-series for gaming handhelds Call of Duty Modern Warfare 4 is next and it's the COD we've been begging for Tips and picks Tip of the week: Now you can vibe code a native Windows app from the CLI App pick of the week: iA Writer RunAs Radio this week: Data API Builder and SQL MVP with Jerry Nixon Brown liquor pick of the week: Old Malt Casking of Longmorn 20 These show notes have been truncated due to length. For the full show notes, visit https://twit.tv/shows/windows-weekly/episodes/986 Hosts: Leo Laporte, Paul Thurrott, and Richard Campbell Sponsors: joindeleteme.com/twit promo code TWIT threatlocker.com/twit cachefly.com/twit

Build 2026 is underway in San Francisco this week, and it started with a big, overly-long keynote as always. And Computex is this week, too. There's a lot going on, and some of it is fascinating. Plus, WWDC is next week because you cannot relax. Also, Microsoft GA's WinApp CLI, announces the Windows Platform Skills plug-in for native app creation, and you're not going to believe what Paul did next. OK, you will believe itBuild + Computex = OHMYGODOHMYGODOHMYGOD NVIDIA finally announces Arm-based N1X as the RTX Spark RTX Spark is an Arm-based portable workstation chip for Windows 11 Microsoft announces Surface Laptop Ultra - It and other RTX Spark-based PCs will appear in late 2026 Some of this leaked earlier, including a lower-end N1 chipset Microsoft continues to optimize and evolve Windows 11 for developers Windows Developer Configuration, Windows Developer Skills + WinApp CLI, Terminal, more Linux, and more on-device ("unmetered") AI - Tied to this, Copilot+ PC features are coming to more PCs, with CPU/GPU support - this, plus the RTX Spark stuff hints at answers to some obvious questions but there's nothing concrete from Microsoft Microsoft Edge is getting three new on-AI features Scout is a personal work agent powered by OpenClaw GitHub Copilot app arrives on desktop for your agentic coding and management needs Microsoft AI announces seven new foundation models Stevie Bathiche is back, baby! And he's talking about those AI app structures and how they've led to Project Solara Windows Microsoft discusses the progress it's made on Windows 11 pain points You can now test the new Start menu in Experimental - Paul did so along with the new Taskbar Qualcomm announces low-cost Snapdragon C for $300+ PCs to take on MacBook Neo And Acer is the first to announce a Snapdragon C laptop New Surface Pro with Snapdragon X2 leaks for June release (!) Dell XPS 13 is coming soon with Intel Wildcat (also to take on MacBook Neo) Dell revenues are through the roof, but not because of PCs HP revenues are up, and it is because of PCs AI and dev Anthropic gets a new valuation exceeding OpenAI and then it files for an IPO OpenAI adjusts GPT5.5-Instant for less sucking-up and releases computer use in Codex on Windows Flutter takes the lead on Flutter desktop development XBOX and gaming Asha Sharma says you can't please everyone and then immediately jumps the shark trying to please everyone XBOX delays Fable reboot because of GTA VI New titles coming to Game Pass in early June across platforms XBOX starts early testing of new console features ASUS announces ROG Xbox Ally X20 with OLED display and XReal R1 glasses Intel announces Arc G-series for gaming handhelds Call of Duty Modern Warfare 4 is next and it's the COD we've been begging for Tips and picks Tip of the week: Now you can vibe code a native Windows app from the CLI App pick of the week: iA Writer RunAs Radio this week: Data API Builder and SQL MVP with Jerry Nixon Brown liquor pick of the week: Old Malt Casking of Longmorn 20 These show notes have been truncated due to length. For the full show notes, visit https://twit.tv/shows/windows-weekly/episodes/986 Hosts: Leo Laporte, Paul Thurrott, and Richard Campbell Sponsors: joindeleteme.com/twit promo code TWIT threatlocker.com/twit cachefly.com/twit

Build 2026 is underway in San Francisco this week, and it started with a big, overly-long keynote as always. And Computex is this week, too. There's a lot going on, and some of it is fascinating. Plus, WWDC is next week because you cannot relax. Also, Microsoft GA's WinApp CLI, announces the Windows Platform Skills plug-in for native app creation, and you're not going to believe what Paul did next. OK, you will believe itBuild + Computex = OHMYGODOHMYGODOHMYGOD NVIDIA finally announces Arm-based N1X as the RTX Spark RTX Spark is an Arm-based portable workstation chip for Windows 11 Microsoft announces Surface Laptop Ultra - It and other RTX Spark-based PCs will appear in late 2026 Some of this leaked earlier, including a lower-end N1 chipset Microsoft continues to optimize and evolve Windows 11 for developers Windows Developer Configuration, Windows Developer Skills + WinApp CLI, Terminal, more Linux, and more on-device ("unmetered") AI - Tied to this, Copilot+ PC features are coming to more PCs, with CPU/GPU support - this, plus the RTX Spark stuff hints at answers to some obvious questions but there's nothing concrete from Microsoft Microsoft Edge is getting three new on-AI features Scout is a personal work agent powered by OpenClaw GitHub Copilot app arrives on desktop for your agentic coding and management needs Microsoft AI announces seven new foundation models Stevie Bathiche is back, baby! And he's talking about those AI app structures and how they've led to Project Solara Windows Microsoft discusses the progress it's made on Windows 11 pain points You can now test the new Start menu in Experimental - Paul did so along with the new Taskbar Qualcomm announces low-cost Snapdragon C for $300+ PCs to take on MacBook Neo And Acer is the first to announce a Snapdragon C laptop New Surface Pro with Snapdragon X2 leaks for June release (!) Dell XPS 13 is coming soon with Intel Wildcat (also to take on MacBook Neo) Dell revenues are through the roof, but not because of PCs HP revenues are up, and it is because of PCs AI and dev Anthropic gets a new valuation exceeding OpenAI and then it files for an IPO OpenAI adjusts GPT5.5-Instant for less sucking-up and releases computer use in Codex on Windows Flutter takes the lead on Flutter desktop development XBOX and gaming Asha Sharma says you can't please everyone and then immediately jumps the shark trying to please everyone XBOX delays Fable reboot because of GTA VI New titles coming to Game Pass in early June across platforms XBOX starts early testing of new console features ASUS announces ROG Xbox Ally X20 with OLED display and XReal R1 glasses Intel announces Arc G-series for gaming handhelds Call of Duty Modern Warfare 4 is next and it's the COD we've been begging for Tips and picks Tip of the week: Now you can vibe code a native Windows app from the CLI App pick of the week: iA Writer RunAs Radio this week: Data API Builder and SQL MVP with Jerry Nixon Brown liquor pick of the week: Old Malt Casking of Longmorn 20 These show notes have been truncated due to length. For the full show notes, visit https://twit.tv/shows/windows-weekly/episodes/986 Hosts: Leo Laporte, Paul Thurrott, and Richard Campbell Sponsors: joindeleteme.com/twit promo code TWIT threatlocker.com/twit cachefly.com/twit

Build 2026 is underway in San Francisco this week, and it started with a big, overly-long keynote as always. And Computex is this week, too. There's a lot going on, and some of it is fascinating. Plus, WWDC is next week because you cannot relax. Also, Microsoft GA's WinApp CLI, announces the Windows Platform Skills plug-in for native app creation, and you're not going to believe what Paul did next. OK, you will believe itBuild + Computex = OHMYGODOHMYGODOHMYGOD NVIDIA finally announces Arm-based N1X as the RTX Spark RTX Spark is an Arm-based portable workstation chip for Windows 11 Microsoft announces Surface Laptop Ultra - It and other RTX Spark-based PCs will appear in late 2026 Some of this leaked earlier, including a lower-end N1 chipset Microsoft continues to optimize and evolve Windows 11 for developers Windows Developer Configuration, Windows Developer Skills + WinApp CLI, Terminal, more Linux, and more on-device ("unmetered") AI - Tied to this, Copilot+ PC features are coming to more PCs, with CPU/GPU support - this, plus the RTX Spark stuff hints at answers to some obvious questions but there's nothing concrete from Microsoft Microsoft Edge is getting three new on-AI features Scout is a personal work agent powered by OpenClaw GitHub Copilot app arrives on desktop for your agentic coding and management needs Microsoft AI announces seven new foundation models Stevie Bathiche is back, baby! And he's talking about those AI app structures and how they've led to Project Solara Windows Microsoft discusses the progress it's made on Windows 11 pain points You can now test the new Start menu in Experimental - Paul did so along with the new Taskbar Qualcomm announces low-cost Snapdragon C for $300+ PCs to take on MacBook Neo And Acer is the first to announce a Snapdragon C laptop New Surface Pro with Snapdragon X2 leaks for June release (!) Dell XPS 13 is coming soon with Intel Wildcat (also to take on MacBook Neo) Dell revenues are through the roof, but not because of PCs HP revenues are up, and it is because of PCs AI and dev Anthropic gets a new valuation exceeding OpenAI and then it files for an IPO OpenAI adjusts GPT5.5-Instant for less sucking-up and releases computer use in Codex on Windows Flutter takes the lead on Flutter desktop development XBOX and gaming Asha Sharma says you can't please everyone and then immediately jumps the shark trying to please everyone XBOX delays Fable reboot because of GTA VI New titles coming to Game Pass in early June across platforms XBOX starts early testing of new console features ASUS announces ROG Xbox Ally X20 with OLED display and XReal R1 glasses Intel announces Arc G-series for gaming handhelds Call of Duty Modern Warfare 4 is next and it's the COD we've been begging for Tips and picks Tip of the week: Now you can vibe code a native Windows app from the CLI App pick of the week: iA Writer RunAs Radio this week: Data API Builder and SQL MVP with Jerry Nixon Brown liquor pick of the week: Old Malt Casking of Longmorn 20 These show notes have been truncated due to length. For the full show notes, visit https://twit.tv/shows/windows-weekly/episodes/986 Hosts: Leo Laporte, Paul Thurrott, and Richard Campbell Sponsors: joindeleteme.com/twit promo code TWIT threatlocker.com/twit cachefly.com/twit

Build 2026 is underway in San Francisco this week, and it started with a big, overly-long keynote as always. And Computex is this week, too. There's a lot going on, and some of it is fascinating. Plus, WWDC is next week because you cannot relax. Also, Microsoft GA's WinApp CLI, announces the Windows Platform Skills plug-in for native app creation, and you're not going to believe what Paul did next. OK, you will believe itBuild + Computex = OHMYGODOHMYGODOHMYGOD NVIDIA finally announces Arm-based N1X as the RTX Spark RTX Spark is an Arm-based portable workstation chip for Windows 11 Microsoft announces Surface Laptop Ultra - It and other RTX Spark-based PCs will appear in late 2026 Some of this leaked earlier, including a lower-end N1 chipset Microsoft continues to optimize and evolve Windows 11 for developers Windows Developer Configuration, Windows Developer Skills + WinApp CLI, Terminal, more Linux, and more on-device ("unmetered") AI - Tied to this, Copilot+ PC features are coming to more PCs, with CPU/GPU support - this, plus the RTX Spark stuff hints at answers to some obvious questions but there's nothing concrete from Microsoft Microsoft Edge is getting three new on-AI features Scout is a personal work agent powered by OpenClaw GitHub Copilot app arrives on desktop for your agentic coding and management needs Microsoft AI announces seven new foundation models Stevie Bathiche is back, baby! And he's talking about those AI app structures and how they've led to Project Solara Windows Microsoft discusses the progress it's made on Windows 11 pain points You can now test the new Start menu in Experimental - Paul did so along with the new Taskbar Qualcomm announces low-cost Snapdragon C for $300+ PCs to take on MacBook Neo And Acer is the first to announce a Snapdragon C laptop New Surface Pro with Snapdragon X2 leaks for June release (!) Dell XPS 13 is coming soon with Intel Wildcat (also to take on MacBook Neo) Dell revenues are through the roof, but not because of PCs HP revenues are up, and it is because of PCs AI and dev Anthropic gets a new valuation exceeding OpenAI and then it files for an IPO OpenAI adjusts GPT5.5-Instant for less sucking-up and releases computer use in Codex on Windows Flutter takes the lead on Flutter desktop development XBOX and gaming Asha Sharma says you can't please everyone and then immediately jumps the shark trying to please everyone XBOX delays Fable reboot because of GTA VI New titles coming to Game Pass in early June across platforms XBOX starts early testing of new console features ASUS announces ROG Xbox Ally X20 with OLED display and XReal R1 glasses Intel announces Arc G-series for gaming handhelds Call of Duty Modern Warfare 4 is next and it's the COD we've been begging for Tips and picks Tip of the week: Now you can vibe code a native Windows app from the CLI App pick of the week: iA Writer RunAs Radio this week: Data API Builder and SQL MVP with Jerry Nixon Brown liquor pick of the week: Old Malt Casking of Longmorn 20 These show notes have been truncated due to length. For the full show notes, visit https://twit.tv/shows/windows-weekly/episodes/986 Hosts: Leo Laporte, Paul Thurrott, and Richard Campbell Sponsors: joindeleteme.com/twit promo code TWIT threatlocker.com/twit cachefly.com/twit

Build 2026 is underway in San Francisco this week, and it started with a big, overly-long keynote as always. And Computex is this week, too. There's a lot going on, and some of it is fascinating. Plus, WWDC is next week because you cannot relax. Also, Microsoft GA's WinApp CLI, announces the Windows Platform Skills plug-in for native app creation, and you're not going to believe what Paul did next. OK, you will believe itBuild + Computex = OHMYGODOHMYGODOHMYGOD NVIDIA finally announces Arm-based N1X as the RTX Spark RTX Spark is an Arm-based portable workstation chip for Windows 11 Microsoft announces Surface Laptop Ultra - It and other RTX Spark-based PCs will appear in late 2026 Some of this leaked earlier, including a lower-end N1 chipset Microsoft continues to optimize and evolve Windows 11 for developers Windows Developer Configuration, Windows Developer Skills + WinApp CLI, Terminal, more Linux, and more on-device ("unmetered") AI - Tied to this, Copilot+ PC features are coming to more PCs, with CPU/GPU support - this, plus the RTX Spark stuff hints at answers to some obvious questions but there's nothing concrete from Microsoft Microsoft Edge is getting three new on-AI features Scout is a personal work agent powered by OpenClaw GitHub Copilot app arrives on desktop for your agentic coding and management needs Microsoft AI announces seven new foundation models Stevie Bathiche is back, baby! And he's talking about those AI app structures and how they've led to Project Solara Windows Microsoft discusses the progress it's made on Windows 11 pain points You can now test the new Start menu in Experimental - Paul did so along with the new Taskbar Qualcomm announces low-cost Snapdragon C for $300+ PCs to take on MacBook Neo And Acer is the first to announce a Snapdragon C laptop New Surface Pro with Snapdragon X2 leaks for June release (!) Dell XPS 13 is coming soon with Intel Wildcat (also to take on MacBook Neo) Dell revenues are through the roof, but not because of PCs HP revenues are up, and it is because of PCs AI and dev Anthropic gets a new valuation exceeding OpenAI and then it files for an IPO OpenAI adjusts GPT5.5-Instant for less sucking-up and releases computer use in Codex on Windows Flutter takes the lead on Flutter desktop development XBOX and gaming Asha Sharma says you can't please everyone and then immediately jumps the shark trying to please everyone XBOX delays Fable reboot because of GTA VI New titles coming to Game Pass in early June across platforms XBOX starts early testing of new console features ASUS announces ROG Xbox Ally X20 with OLED display and XReal R1 glasses Intel announces Arc G-series for gaming handhelds Call of Duty Modern Warfare 4 is next and it's the COD we've been begging for Tips and picks Tip of the week: Now you can vibe code a native Windows app from the CLI App pick of the week: iA Writer RunAs Radio this week: Data API Builder and SQL MVP with Jerry Nixon Brown liquor pick of the week: Old Malt Casking of Longmorn 20 These show notes have been truncated due to length. For the full show notes, visit https://twit.tv/shows/windows-weekly/episodes/986 Hosts: Leo Laporte, Paul Thurrott, and Richard Campbell Sponsors: joindeleteme.com/twit promo code TWIT threatlocker.com/twit cachefly.com/twit

無論是電動車或資料中心正經歷800V的革命… SIC碳化矽和傳統矽有什麼分別? SIC供應鏈有什麼美股? 為什麼最近才開始炒起?

AMD (AMD) shares hit record highs after a strong earnings beat driven by data center demand. Ben Bajarin and Matt Bryson discuss the resurgence of CPUs in AI workloads and how AMD's integrated CPU‑GPU strategy positions it against Nvidia (NVDA) and Intel (INTC) in a fast‑growing market.======== Schwab Network ========Options involve risks and are not suitable for all investors. Before trading, read the Options Disclosure Document. http://bit.ly/2v9tH6DEmpowering every investor and trader, every market day.Options involve risks and are not suitable for all investors. Before trading, read the Options Disclosure Document. http://bit.ly/2v9tH6DSubscribe to the Market Minute newsletter - https://schwabnetwork.com/subscribeDownload the iOS app - https://apps.apple.com/us/app/schwab-network/id1460719185Download the Amazon Fire Tv App - https://www.amazon.com/TD-Ameritrade-Network/dp/B07KRD76C7Watch on Sling - https://watch.sling.com/1/asset/191928615bd8d47686f94682aefaa007/watchWatch on Vizio - https://www.vizio.com/en/watchfreeplus-exploreWatch on DistroTV - https://www.distro.tv/live/schwab-network/Follow us on X – https://twitter.com/schwabnetworkFollow us on Facebook – https://www.facebook.com/schwabnetworkFollow us on LinkedIn - https://www.linkedin.com/company/schwab-network/ About Schwab Network - https://schwabnetwork.com/about

Two market trends are driving the need for a higher CPU:GPU ratio in AI infrastructure. First, the inferencing market is...

-Arm enters the chip business with Arm AGI CPU - Is AI inference increasing the CPU:GPU ratio? - Google's TurboQuant algorithm - Data center energy efficiency - Bringing back a 2,000-year-old cement battery - India's chip industry - TSMC's 2nm capacity crunch - $20B for a TeraFab ? [audio mp3="https://orionx.net/wp-content/uploads/2026/03/HPCNB_20260330.mp3"][/audio]   The post HPC News Bytes – 20260330 appeared first on OrionX.net.

AMDが組み込み機器用APU「Ryzen AI Embedded P100」のラインアップを拡充　CPUコアとGPUコアを“増量”したモデルを用意。　AMDは3月9日（米国太平洋夏時間）、組み込み機器向けAPU（GPU統合型CPU）「Ryzen AI Embedded P100プロセッサ」の追加ラインアップを発表した。

This podcast examines how HPE is meeting AI's soaring compute demands with CPU/GPU advances, hybrid cloud models, and Neo cloud services that enable faster upgrades, liquid cooling, and flexible, consumption-based infrastructure.

On this week's episode of The Circuit, Ben Bajarin and Jay Goldberg break down the surprising new partnership between NVIDIA and Intel, including a $5B investment and plans for co-developed CPU–GPU systems for data centers and PCs. They explore what this means for x86 in AI racks, AMD's competitive position, and why Intel's silence on Foundry raises big questions. The conversation also touches on Meta's new Ray-Ban AR glasses—demo fails and all—and Huawei's unexpected progress in advanced chipmaking. A packed episode full of strategy, irony, and hot takes on the future of semiconductors.

Big thanks to Brilliant for sponsoring this video. To try everything Brilliant has to offer for free for a full 30 days, visit https://brilliant.org/davidbombal or scan the QR code onscreen – You'll also get 20% off an annual premium subscription. Quantum computing isn't a far-off sci-fi threat. In a hybrid model (CPU + GPU + quantum), it could land within ~3 years—and that's enough to break today's asymmetric cryptography (RSA/ECC/PKI) used for TLS/HTTPS and VPN key exchange via Shor's algorithm. Result? Session keys exposed, mass decryption possible, and a risky “post-quantum gap” before new hardware/software are widely deployed. In this frank discussion, OTW joins David Bombal to explain: • Why asymmetric key exchange is the weak link for TLS/VPNs • How hybrid quantum + CUDA-Q accelerates timelines • Nation-states' “collect now, decrypt later” strategy • The painful window before post-quantum cryptography (PQC) rolls out • Risks of client-side scanning and on-device AI • NSA backdoors, PGP/Phil Zimmermann, and the San Bernardino iPhone case // Occupy The Web SOCIAL // X: / three_cube Website: https://hackers-arise.net/ // Occupy The Web Books // Linux Basics for Hackers 2nd Ed US: https://amzn.to/3TscpxY UK: https://amzn.to/45XaF7j Linux Basics for Hackers: US: https://amzn.to/3wqukgC UK: https://amzn.to/43PHFev Getting Started Becoming a Master Hacker US: https://amzn.to/4bmGqX2 UK: https://amzn.to/43JG2iA Network Basics for hackers: US: https://amzn.to/3yeYVyb UK: https://amzn.to/4aInbGK // OTW Discount // Use the code BOMBAL to get a 20% discount off anything from OTW's website: https://hackers-arise.net/ // Playlists REFERENCE // Linux Basics for Hackers: • Linux for Hackers Tutorial (And Free Courses) Mr Robot: • Hack like Mr Robot // WiFi, Bluetooth and ... Hackers Arise / Occupy the Web Hacks: • Hacking Tools (with demos) that you need t... // YouTube video REFERENCE // • USA stops UK creating Apple backdoor. Why ... • Will this Tiny Chip Change EVERYTHING in Q... // David's SOCIAL // Discord: / discord X: / davidbombal Instagram: / davidbombal LinkedIn: / davidbombal Facebook: / davidbombal.co TikTok: / davidbombal YouTube: / @davidbombal Spotify: open.spotify.com/show/3f6k6gE... SoundCloud: / davidbombal Apple Podcast: podcasts.apple.com/us/podcast... // MY STUFF // https://www.amazon.com/shop/davidbombal // SPONSORS // Interested in sponsoring my videos? Reach out to my team here: sponsors@davidbombal.com // MENU // 0:00 - Coming Up 01:00 - Intro 01:49 - OTW's Books 02:46 - Brilliant Advert 04:01 - Encryption and Apple Backdoors 05:12 - Nation-states Racing for Quantum Computing 08:35 - Jensen Huang's Hybrid Quantum Computer 10:57 - The Threat of National Agencies 12:10 - China Collecting Encrypted Data 13:56 - Post-Quantum Cryptography 17:58 - The NSA Collecting Data 19:21 - The NSA and Encryption 24:40 - Concerns Against Cryptography 28:09 - Client-Side Scanning 30:46 - Should you Run AI on your Devices? 34:56 - NVIDIA's Role on Cryptography 37:37 - Conclusion Please note that links listed may be affiliate links and provide me with a small percentage/kickback should you use them to purchase any of the items listed or recommended. Thank you for supporting me and this channel! Disclaimer: This video is for educational purposes only.

Press XJoin the Press X Discord: https://discord.gg/MAXtvmv2rw Topics:Doug Bowser explains why Switch 2 costs $150 more than Switch: 1) "a lot of improved features" with Joy-Con 2, LCD screen, and CPU/GPU processing power, and 2) the social features like GameChat - CNBC Video“the Super Mario Bros. 2 movie” will be released next March (movie hasn't been described as a “2” before)Bowser offers iffier quote about price to CBC that went viral https://www.cbc.ca/news/entertainment/nintendo-switch-2-price-tariffs-1.7503598Doug Bowser claims tariffs "weren't factored into the pricing" of Switch 2, but Nintendo is "actively assessing" the situation - Wired (non-paywall) / (gives lots of similar non-answers to NPR too)Doug Bowser explains the $80 price for Mario Kart World Washington Post (non-paywall)Switch 2 Joy-Con controller was designed "from the ground up" but doesn't use Hall sticks https://www.engadget.com/gaming/nintendo/nintendo-says-the-switch-2-joy-con-controllers-dont-have-hall-effect-thumbsticks-145541771.htmlNintendo Switch 2 Welcome Tour is $9.99; BotW & TotK Switch 2 upgrades are also $9.99 each https://www.ign.com/articles/our-big-nintendo-switch-2-interview-with-nintendo-of-americas-bill-trinenSwitch 2 carts still taste bad on purpose https://www.gamespot.com/articles/switch-2-carts-still-taste-bad-designed-purposefully-to-be-spat-out/1100-6530649/Switch 2 eShop will have no music either; Switch 2 has no achievementsBuilds for old or canceled Switch games have leaked, apparently including Retro Studios' canceled music RPG, Project Harmony https://nintendoeverything.com/cancelled-nintendo-switch-games-leaked-online-with-footage-of-scrapped-retro-studios-game-ridge-racer/Xbox Game Showcase set for June 8, followed immediately by The Outer Worlds 2 Direct https://news.xbox.com/en-us/2025/04/09/xbox-games-showcase-2025-outer-worlds-2-direct/ Questions from Discord: 

Подивитись відео на YouTube Patreon - https://www.patreon.com/androidstoryTelegram https://t.me/androidstory_chat00:00:00 - Враження від виставки CES. Порівняння поштових сервісів: Укрпошта та Нова Пошта, їхні переваги, недоліки, ринкові долі та державне значення00:27:12 - Історичний екскурс у розвиток поштових служб00:31:43 - Обговорення доставки дронами Amazon: технологія, обмеження та порівняння інженерних можливостей України та Заходу00:46:17 - Продовження теми відеокарт: порівняння продуктивності та енергоспоживання Nvidia 5090 та 4090, технологія MultiFrame Generation00:52:32 - Розмова про апскейлінг технології: DLSS від Nvidia, відкритий FSR від AMD та їх інтеграція в ігрові рушії01:00:12 - Інтегровані відеокарти AMD, спільна пам'ять CPU/GPU та їхнє значення для локальних нейромереж і прайвасі01:11:17 - Завершення: філософські роздуми про цінність "фейкових" фреймів та метрики оцінки продуктивності відеокарт

Lords: * Andi * Casey Topics: * Lifehacks as communion with the divine * I decided to fire my computer * Winston is starting to forget things Microtopics: * A Star Trek watchalong podcast that doesn't exist yet. * Positing that what you said is no longer an NDA violation by the time this episode comes out. * Plugging a fake game that you worked on. * Astrobot. * Horror movie clinky noises that you can't hear over the PS4 fan noises. * Caffeine-infused mints with Tux the Penguin branding on Think Geek dot com. * The pre-eminent source for Life Hacks. * Using a hotel shower cap to bake bread. * Anime girls that are happy to see you. * That one time Film Crit Hulk broke character. * The joy of moving efficiently through the world. * More efficient ways to set the microwave timer. * Hotel rooms that you can bake bread in. * Whether bread should contain hair. * Tricking yourself into not being bored while doing something you have to do. * Reading 50 life hacks and applying none of them because. * Viral Life Hack that's killed 33 people. * A life hack that already had a body count in the double digits before someone made a TikTok about it. * Getting really fed up with computers. * Cryptographic signing processes that you can't participate in. * The HDCP certification board taking steps to ensure nobody can take a screen shot of their Crunchy Roll anime. * The analog hole. * Open source web browsers that can't see DRM content. * Microsoft-authenticated Linux installations. * Designing a circuit that solves a math problem. * Stamping a circuit onto your circuit clay. * An independent circuit re-implementation of video game hardware. * Should you use FPGA to do a thing? * Ridiculous multi-level memory caching systems. * Bootstrapping an FPGA design tool that runs on an FPGA device. * Every single circuit doing something on every single cycle. * Voltages going high and/or low. * Making a bunch of CPUs and testing them afterwards to see how many GHz they have. * Why the PS3 Cell processor had 7 SPUs * The industrial uses of the Cell processor. * A GLSL compiler that outputs FPGA circuits. * Mr. MiSTer. * Open-hardware laptops. * Inventing an open-source GPU. * Multics or Minix. * Writing a Breakout clone in Rust targeting the weird CPU your friend just invented. * Making a terrible first effort that is the right kind of good enough. * A laptop that has a FPGA where the CPU/GPU usually goes. * 1970s-era TV games. * The Epoch Cassette Vision. * A game console with interchangeable cartridges where the CPU is on the cartridge. * The Glasgow Interface Explorer. * Describing your FPGA circuit in Python. * Manufacturing homebrew Cassette Vision Homebrew cartridges for the audience of zero Cassette Vision owners. * Making art just for you, in the most overly elaborate and overly complicated way possible. * The programmer equivalent of going to swim with the dolphins. * Diagonal pixels. * Childhood amnesia. * Remembering your memories. * Using 10% of your brain. (And also the other 90%.) * Knowing things about stuff. * When one brother dies, the other brother gets their memories. * Memories that are formed before vs. after you learn to talk. * Being persecuted for being friends with a girl. * Rules of heteronormativity being enforced by three year olds. * Getting off of Wordpress.

Les fabricants des puces qui composent vos processeurs et vos cartes graphiques cachent un lourd secret lors de la fabrication et de la mise en vente des cartes... Mais lequel est-ce ?Notes de l'épisode :L'architecture Von Neumann : https://code-garage.com/podcast/circuits/episode-8

12 月 3 日，英特尔的 CEO 帕特·基尔辛格（Pat Gelsinger）突然宣布辞去首席执行官职务。很多报道透露，基辛格其实是在董事会的压力下「被迫退休」。最近半年，关于英特尔这家芯片巨头的负面消息接连不断。从公司市值上看，过去五年的时间，英特尔的市值缩水了近一半。为了应对财务压力，英特尔近期也宣布了一项高达100亿美元的成本削减计划，以及将裁员约 1.5 万名员工的消息。这一系列的坏消息都预示着这个曾经的科技霸主正经历着其历史上最为艰难的时期。 我们今天的话题就是想一起来分析以下曾经辉煌的英特尔正在遭遇的困境，刚刚辞职的 CEO Pat Gelsinger 为挽回公司所做的一些努力，以及他离开后公司可能的未来。 本期人物 硅谷徐老师，AI 高管、连续创业者、斯坦福客座讲师，小红书和微信视频号：硅谷徐老师 ｜公众号：硅谷云｜ YouTube： Byte into Future Yaxian，「科技早知道」监制 Sam Li, 曾任职英特尔，现从事虚拟机和高性能计算架构的研发工作 Tom Qin, 连续创业者，曾任职小米，熟悉半导体产业及电子产品供应链体 主要话题 [03:22] Windows 操作系统与 x86 架构形成的 Wintel 联盟让英特尔几乎独占了个人计算机市场 [05:24] 英特尔卖掉 ARM 架构许可证，从此错失掉手机芯片市场 [13:14] 苹果曾希望与英特尔合作开发手机芯片，却遭英特尔拒绝 [15:51] 在成熟的 CPU 带来的市场成功面前，GPU 在英特尔的研发被边缘化 [19:23] 技术的护城河变成了创新的绊脚石，这样的例子在科技企业中屡见不鲜 [22:26] 失去手机芯片市场之后：英特尔的晶圆厂已无法再与台积电相抗衡 [26:46] 假如时光可以倒流，英特尔的 CEO 们还会做出同样错误的决定吗？ [32:58] CEO Pat Gelsinger 的传奇人生：18 岁加入英特尔，08 年负气出走，21 年重新回归 [34:43] Pat 上任后的晶圆厂复兴计划：一个正确但注定失败的选择 [38:57] 冒险精神：大公司最需要也最稀缺的企业文化 [42:10] 悲观的现实和乐观的未来：英特尔可能面临分拆，但硅谷精神将永续 幕后制作 监制：Yaxian 后期：Jack 运营：George 设计：饭团 商务合作 声动活泼商务合作咨询 (https://sourl.cn/6vdmQT) 加入声动活泼会员，支持我们创作好内容 2021 年我们发起了「声动胡同会员计划」，这是一个纯支持项目，支持我们不断制作可信赖的、扎实的商业信息。得益于这些支持，我们每档节目内容和形式不断突破，听友们越来越多、也常在苹果中国年度榜、CPA、喜马拉雅榜上有名。 为了感谢各位的支持，我们会在每周向付费会员们提供一封播客手记，手记中会记录我们在节目之外更多的思考和观察，你可以通过它和我们保持更深度的交流；付费会员还能免费收听所有我们旗下的付费内容，如「不止金钱」和「跳进兔子洞第三季」。新会员 365 元一年，相当于一天一块钱。欢迎点击此处成为好内容的支持者。 加入我们 加入我们：声动活泼正在寻找商业化合作经理、播客节目监制的全职伙伴，以及早咖啡内容实习生，详情点击招聘入口； 关于声动活泼 「用声音碰撞世界」，声动活泼致力于为人们提供源源不断的思考养料。 我们还有这些播客：声动早咖啡 (https://www.xiaoyuzhoufm.com/podcast/60de7c003dd577b40d5a40f3)、声东击西 (https://etw.fm/episodes)、吃喝玩乐了不起 (https://www.xiaoyuzhoufm.com/podcast/644b94c494d78eb3f7ae8640)、反潮流俱乐部 (https://www.xiaoyuzhoufm.com/podcast/5e284c37418a84a0462634a4)、泡腾 VC (https://www.xiaoyuzhoufm.com/podcast/5f445cdb9504bbdb77f092e9)、商业WHY酱 (https://www.xiaoyuzhoufm.com/podcast/61315abc73105e8f15080b8a)、跳进兔子洞 (https://therabbithole.fireside.fm/) 、不止金钱 (https://www.xiaoyuzhoufm.com/podcast/65a625966d045a7f5e0b5640) 欢迎在即刻 (https://okjk.co/Qd43ia)、微博等社交媒体上与我们互动，搜索 声动活泼 即可找到我们。 期待你给我们写邮件，邮箱地址是：ting@sheng.fm 声小音 https://files.fireside.fm/file/fireside-uploads/images/4/4931937e-0184-4c61-a658-6b03c254754d/gK0pledC.png 欢迎扫码添加声小音，在节目之外和我们保持联系。 Special Guests: Sam Li, Tom Qin, and 雅娴.

Surprise! There's an episode today! We are continuing our series on The Computer in this episode. We dive into CPUS and GPUS each of their individual parts, and the difference between them. Please DO Watch Part one of the video (S2 E14) for more context on what we are talking about! (Unless you have a general knowledge on RAM, SSD, HDD, and Display Technology) Here's the video I was talking about (CPU v/s GPU) https://www.youtube.com/watch?v=-P28LKWTzrI Sources: https://handwovenmagazine.com/green-dye-death/ https://www.techtarget.com/searchstorage/definition/cache-memory#:~:text=There%20are%20three%20general%20cache,often%20more%20capacious%20than%20L1. https://ms.codes/blogs/computer-hardware/how-many-alu-in-cpu#:~:text=In%20a%20single%20ALU%20architecture,all%20arithmetic%20and%20logical%20operations. https://www.redhat.com/en/blog/cpu-components-functionality https://aws.amazon.com/what-is/cpu/ https://edu.gcfglobal.org/en/computerbasics/inside-a-computer/1/ https://www.lenovo.com/us/en/glossary/arithmetic-logic-unit/ https://www.digitaltrends.com/computing/what-is-a-cpu/ https://www.youtube.com/watch?v=Axd50ew4pco

Episode 53: We've been testing games this week, including S.T.A.L.K.E.R. 2 and Flight Simulator 2024, so we discuss how these games run on PC. Loading issues, punishing CPU requirements, VRAM issues and more.CHAPTERS00:00 - Intro04:07 - Testing S.T.A.L.K.E.R. 214:51 - VRAM is An Issue Again34:18 - Floaty Controls and Frame Generation39:04 - Testing Flight Simulator 202446:51 - Updates From Our Boring LivesBluesky: https://bsky.app/profile/hardwareunboxed.bsky.socialSUBSCRIBE TO THE PODCASTAudio: https://shows.acast.com/the-hardware-unboxed-podcastVideo: https://www.youtube.com/channel/UCqT8Vb3jweH6_tj2SarErfwSUPPORT US DIRECTLYPatreon: https://www.patreon.com/hardwareunboxedFloatplane: https://www.floatplane.com/channel/HardwareUnboxedLINKSYouTube: https://www.youtube.com/@Hardwareunboxed/Twitter: https://twitter.com/HardwareUnboxed Hosted on Acast. See acast.com/privacy for more information.

Bernie Wu is VP of Business Development for MemVerge. He has 25+ years of experience as a senior executive for data center hardware and software infrastructure companies including companies such as Conner/Seagate, Cheyenne Software, Trend Micro, FalconStor, Levyx, and MetalSoft. Boosting LLM/RAG Workflows & Scheduling w/ Composable Memory and Checkpointing // MLOps Podcast #270 with Bernie Wu, VP Strategic Partnerships/Business Development of MemVerge. // Abstract Limited memory capacity hinders the performance and potential of research and production environments utilizing Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) techniques. This discussion explores how leveraging industry-standard CXL memory can be configured as a secondary, composable memory tier to alleviate this constraint. We will highlight some recent work we've done in integrating of this novel class of memory into LLM/RAG/vector database frameworks and workflows. Disaggregated shared memory is envisioned to offer high performance, low latency caches for model/pipeline checkpoints of LLM models, KV caches during distributed inferencing, LORA adaptors, and in-process data for heterogeneous CPU/GPU workflows. We expect to showcase these types of use cases in the coming months. // Bio Bernie is VP of Strategic Partnerships/Business Development for MemVerge. His focus has been building partnerships in the AI/ML, Kubernetes, and CXL memory ecosystems. He has 25+ years of experience as a senior executive for data center hardware and software infrastructure companies including companies such as Conner/Seagate, Cheyenne Software, Trend Micro, FalconStor, Levyx, and MetalSoft. He is also on the Board of Directors for Cirrus Data Solutions. Bernie has a BS/MS in Engineering from UC Berkeley and an MBA from UCLA. // MLOps Swag/Merch https://mlops-community.myshopify.com/ // Related Links Website: www.memverge.com Accelerating Data Retrieval in Retrieval Augmentation Generation (RAG) Pipelines using CXL: https://memverge.com/accelerating-data-retrieval-in-rag-pipelines-using-cxl/ --------------- ✌️Connect With Us ✌️ ------------- Join our slack community: https://go.mlops.community/slack Follow us on Twitter: @mlopscommunity Sign up for the next meetup: https://go.mlops.community/register Catch all episodes, blogs, newsletters, and more: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Bernie on LinkedIn: https://www.linkedin.com/in/berniewu/

Hello hello everyone, this is Alex, typing these words from beautiful Seattle (really, it only rained once while I was here!) where I'm attending Microsoft biggest developer conference BUILD. This week we saw OpenAI get in the news from multiple angles, none of them positive and Microsoft clapped back at Google from last week with tons of new AI product announcements (CoPilot vs Gemini) and a few new PCs with NPU (Neural Processing Chips) that run alongside CPU/GPU combo we're familiar with. Those NPUs allow for local AI to run on these devices, making them AI native devices! While I'm here I also had the pleasure to participate in the original AI tinkerers thanks to my friend Joe Heitzberg who operates and runs the aitinkerers.org (of which we are a local branch in Denver) and it was amazing to see tons of folks who listen to ThursdAI + read the newsletter and talk about Weave and evaluations with all of them! (Btw, one the left is Vik from Moondream, which we covered multiple times). I Ok let's get to the news: TL;DR of all topics covered: * Open Source LLMs * HuggingFace commits 10M in ZeroGPU (X)* Microsoft open sources Phi-3 mini, Phi-3 small (7B) Medium (14B) and vision models w/ 128K context (Blog, Demo)* Mistral 7B 0.3 - Base + Instruct (HF)* LMSys created a "hard prompts" category (X)* Cohere for AI releases Aya 23 - 3 models, 101 languages, (X)* Big CO LLMs + APIs* Microsoft Build recap - New AI native PCs, Recall functionality, Copilot everywhere * Will post a dedicated episode to this on Sunday* OpenAI pauses GPT-4o Sky voice because Scarlet Johansson complained* Microsoft AI PCs - Copilot+ PCs (Blog)* Anthropic - Scaling Monosemanticity paper - about mapping the features of an LLM (X, Paper)* Vision & Video* OpenBNB - MiniCPM-Llama3-V 2.5 (X, HuggingFace)* Voice & Audio* OpenAI pauses Sky voice due to ScarJo hiring legal counsel* Tools & Hardware* Humane is looking to sell (blog)Open Source LLMs Microsoft open sources Phi-3 mini, Phi-3 small (7B) Medium (14B) and vision models w/ 128K context (Blog, Demo)Just in time for Build, Microsoft has open sourced the rest of the Phi family of models, specifically the small (7B) and the Medium (14B) models on top of the mini one we just knew as Phi-3. All the models have a small context version (4K and 8K) and a large that goes up to 128K (tho they recommend using the small if you don't need that whole context) and all can run on device super quick. Those models have MIT license, so use them as you will, and are giving an incredible performance comparatively to their size on benchmarks. Phi-3 mini, received an interesting split in the vibes, it was really good for reasoning tasks, but not very creative in it's writing, so some folks dismissed it, but it's hard to dismiss these new releases, especially when the benchmarks are that great! LMsys just updated their arena to include a hard prompts category (X) which select for complex, specific and knowledge based prompts and scores the models on those. Phi-3 mini actually gets a big boost in ELO ranking when filtered on hard prompts and beats GPT-3.5

Mojo is the latest language from the creator of Swift and LLVM. It's an attempt to take some of the best techniques from CPU/GPU-level programming and package them up in a Python-compatible syntax.In this episode we explore why Mojo was created, and what it offers to Python programmers and non-Python programmers alike. How is it built for performance, and which performance features matter? What's its take on functional programming and type systems? And can it marry the high-level programming of Python with the low-level programming of LLVM/MLIR?If you're a Python programmer who needs better performance, a C programmer who expects more from a ‘scripting language', or just someone who'd be happier if Python had a first-class type system, Mojo might well be for you…–Mojo: https://www.modular.com/max/mojoMojo's Roadmap: https://docs.modular.com/mojo/roadmap.htmlThe Mojo Discord: https://discord.com/invite/modularMLIR: https://mlir.llvm.org/Chris's Talks: https://nondot.org/sabre/Resume.html#talksChris on Twitter: https://twitter.com/clattner_llvmKris on Mastodon: http://mastodon.social/@krisajenkinsKris on LinkedIn: https://www.linkedin.com/in/krisjenkins/Kris on Twitter: https://twitter.com/krisajenkins–#software #podcast #mojolang #ml #pythonml

Episode 27: Two discussion topics in this podcast episode, firstly on CPU/GPU scaling and when it makes sense to upgrade a CPU vs a GPU, and then we re-rank all the current GPUs from worst to best. Lots of interesting chat in this one!CHAPTERS00:00 - Intro00:25 - CPU and GPU Scaling Discussion28:25 - Re-Ranking All Current GPUs1:05:10 - Updates From Our Boring LivesSUBSCRIBE TO THE PODCASTAudio: https://shows.acast.com/the-hardware-unboxed-podcastVideo: https://www.youtube.com/channel/UCqT8Vb3jweH6_tj2SarErfwSUPPORT US DIRECTLYPatreon: https://www.patreon.com/hardwareunboxedFloatplane: https://www.floatplane.com/channel/HardwareUnboxedLINKSYouTube: https://www.youtube.com/@Hardwareunboxed/Twitter: https://twitter.com/HardwareUnboxed Hosted on Acast. See acast.com/privacy for more information.

1:15 短線股價跌深，為財報跳漲貢獻最大利多(先前集數分析過) 3:00 整理過去12次+本次財報公布反映，3日前報酬確實負相關 7:30 Double beat正常不過，且台積電財報也可看出 ----- NVIDIA 法說的摘要 8:30 AI引爆各區、各行業需求，公司產品需求遠大於供給 9:07 伺服器CPU轉至GPU趨勢未改變，看至2025有好表現 10:10 上一季現金回購23億元季減年增(合計財報年共92億美元) 12:12 NVIDIA與法人都只聚焦Data Center(伺服器晶片)，營運成長單一且強大!? 13:56 對中限制之晶片許可尚未取得，逐步有少量對中出貨(替代的閹割版AI晶片) ----- 17:00 毛利率最亮眼，也要注意毛利率經過一年可能觸頂成長趨緩 20:10 AI伺服器供應鏈，改變趨勢是買家(CSP廠商) CSP三個影響關鍵: 25:00 商業化變現華爾街開始定價，AI支出與回報將開始受到檢視 28:00 股價的高低，也會影響公司被迫砍支出(想想2年前資料中心的故事) 30:33 CSP廠商吃人夠夠，著作、隱私權限制歐盟可能出手? 33:00 自由現金流成長趨緩，帶動股價財報公布後趨平，資金轉換至… 相關圖表資料與完整文章: https://www.big-econ.com/index.php?sec=article&ID=3729

「我們幾乎是用盡全力，以過去自家累積的優勢基礎和架構，一路摸索、建立起SOC測試設備和開發能力!」愛德萬測試副總經理蘇勇鴻與副總經理陳建州，對當年如何大幅跨越記憶體測試領域到SOC測試設備的歷程，仍是記憶猶新。本集節目就邀請兩人連袂分享，愛德萬如何努力突破，以單晶片測試優等生角色，參與建立STC/OpenStar開放架構平台, 獲致在SoC-T2000領域中，並於CPU、GPU(play station)嶄露頭角的市場目標，歡迎收聽。全球自動化測試與量測設備領導製造商­-愛德萬測試技術開發執全球牛耳，擁有70年多元發展技術優勢和價值的愛德萬測試，又如何在此關鍵時刻承先啟後，聚焦全球自動化測試設備，見證半導體世界的風起雲湧?為此，由出身工研院的資深產業主持人謝美芳，從2024年1/2起到4/16，於每週二上午8:15播出的節目，製播專題，帶領聽眾一窺全球半導體產業最大自動化測試設備的發展史，如何在早期透過半導體製程倚靠的精確測試系統，逐漸在其後陸續締造出包括5G通訊、高效能運算、人工智慧，以及自駕車等領域的豐碩成果，透過半導體測試設備繼續精準地領航科技大未來。 

This paper introduces PowerInfer, a high-speed Large Language Model (LLM) inference engine on a personal computer (PC) equipped with a single consumer-grade GPU. The key underlying the design of PowerInfer is exploiting the high locality inherent in LLM inference, characterized by a power-law distribution in neuron activation. This distribution indicates that a small subset of neurons, termed hot neurons, are consistently activated across inputs, while the majority, cold neurons, vary based on specific inputs. PowerInfer exploits such an insight to design a GPU-CPU hybrid inference engine: hot-activated neurons are preloaded onto the GPU for fast access, while cold-activated neurons are computed on the CPU, thus significantly reducing GPU memory demands and CPU-GPU data transfers. PowerInfer further integrates adaptive predictors and neuron-aware sparse operators, optimizing the efficiency of neuron activation and computational sparsity. Evaluation shows that PowerInfer attains an average token generation rate of 13.20 tokens/s, with a peak of 29.08 tokens/s, across various LLMs (including OPT-175B) on a single NVIDIA RTX 4090 GPU, only 18% lower than that achieved by a top-tier server-grade A100 GPU. This significantly outperforms llama.cpp by up to 11.69x while retaining model accuracy. 2023: Yixin Song, Zeyu Mi, Haotong Xie, Haibo Chen https://arxiv.org/pdf/2312.12456v1.pdf

Dans cet épisode, Katia, Arnaud et Emmanuel discutent les nouvelles de cette fin 2023. Le gatherer dans les stream Java, les exceptions, JavaScript dans la JVM, recherche vectorielle, coût du cloud, Gemini, Llama et autres animaux fantastiques et pleins d'outils sympathiques pour fêter la fin de l'année. Enregistré le 15 décembre 2023 Téléchargement de l'épisode LesCastCodeurs-Episode-304.mp3 News Aide Les Cast Codeurs et remplis un petit formulaire pour nous guider l'année prochaine https://lescastcodeurs.com/sondage Langages Avec JEP 461, arrivée dans en preview dans Java 22 de la notion de “gatherer” pour les streams https://groovy.apache.org/blog/groovy-gatherers dans cet article de Paul King, de l'équipe Groovy, il montre et contraste ce que l'on pouvait faire en Groovy depuis des années, comme des sliding windows, par exemple explique l'approche des gatherers avec ses opérations intermédiaires gatherer sont des operations intermediaires custom qui prennent un etat et le prochain element pour decided quoi faire, et meme changer le stream d'elements suivants (en publier) (via la fonction integrate certains peuvent permettre de combiner les resultats intermediaires (pour paralleliser) Examples : fenetres de taille fixe, fenettres glissantes Joe Duffy, qui est CEO de Pulumi, mais qui avait travaillé chez Microsoft sur le project Midori (un futur OS repensé) parle du design des exceptions, des erreurs, des codes de retour https://joeduffyblog.com/2016/02/07/the-error-model/ Il compare les codes d'erreurs, les exceptions, checked et non-checked il separe les bugs des erreurs attendues (bugs doivent arreter le process) il raconte l'histoire des unchecked exception et leurs problemes et des checked exceptopns et poourquoi les developeurs java les detestent (selon lui) long article maisn interessant dans ses retours mais lon je ne suis pas allé au bout :smile: Après la disparition de Nashorn dans le JDK, on peut se tourner vers le projet Javet https://www.caoccao.com/Javet/index.html Javet permet d'intégrer JavaScript avec le moteur V8 Mais aussi carrément Node.js c'est super comme capacité car on a les deux mielleurs moteurs, par contre le support hors x86 est plus limité (genre arm sous windows c'est non) Librairies Une partie de l'équipe Spring se fait lourder après le rachat effectif de Broadcom https://x.com/odrotbohm/status/1729231722498425092?s=20 peu d'info en vrai à part ce tweet mais l'acquisition Broadcome n'a pas l'air de se faire dans le monde des bisounours Marc Wrobel annonce la sortie de JBanking 4.2.0 https://www.marcwrobel.fr/sortie-de-jbanking-4-2-0 support de Java 21 possibilité de générer aléatoirement des BIC amélioration de la génération d'IBAN jbanking est une bibliotheque pour manipuler des structures typiques des banques comme les IBAN les BIC, les monnaies, les SEPA etc. Hibernate Search 7 est sorti https://in.relation.to/2023/12/05/hibernate-search-7-0-0-Final/ Support ElasticSearch 8.10-11 et openSearch 2.10-11 Rebasé sur Lucerne 9.8 support sur Amazon OpenSearch Serverless (experimental) attention sous ensemble de fonctionnalités sur Serverless, c'est un API first search cluster vendu a la lambda En lien aussi sur la version 7.1 alpha1 Hibernate ORM 6.4 est sorti https://in.relation.to/2023/11/23/orm-640-final/ support pour SoftDelete (colonne marquant la suppression) support pour les operations vectorielles (support postgreSQL initialement) les fonctions vectorielles sont particulièrement utilisées par l'IA/ML événement spécifiques JFR Intégration de citrus et Quarkus pour les tests d'intégrations de pleins de protocoles et formats de message https://quarkus.io/blog/testing-quarkus-with-citrus/ permet de tester les entrees / sorties attendues de systèmes de messages (HTTP, Kafka, serveur mail etc) top pour tester les application Event Driven pas de rapport mais Quarkus 3.7 ciblera Java 17 (~8% des gens utilisaient Java 11 dans les builds qui ont activé les notifications) Hibernate Search 7.1 (dev 7.1.0.Alpha1) avec dernière version de Lucene (9.8), Infinispan rajoute le support pour la recherche vectorielle. https://hibernate.org/search/releases/7.1/ https://infinispan.org/blog/2023/12/13/infinispan-vector-search Hibernate Search permet maintenant la recherche vectorielle La dernière version est intégrée en Infinispan 15 (dev) qui sortira La recherche vectoriolle et stockage de vecteurs, permettent convertir Infinispan en Embedding Store (langchain) Cloud Comment choisir sa region cloud https://blog.scottlogic.com/2023/11/23/conscientious-cloud-pick-your-cloud-region-deliberately.html pas si simple le coût la securité légale de vos données la consommation carbone de la région choisie (la France est top, la Pologne moins) la latence vs où sont vos clients les services supportés Web Vers une standardisation des Webhooks ? https://www.standardwebhooks.com/ Des gens de Zapier, Twilio, Ngrok, Kong, Supabase et autres, se rejoignent pour essayer de standardiser l'approche des Webhooks La spec est open source (Apache) sur Github https://github.com/standard-webhooks/standard-webhooks/blob/main/spec/standard-webhooks.md Les objectifs sont la sécurité, la reliabilité, l'interopérabilité, la simplicité et la compatibilité (ascendante / descendante) sans la spec, chaque webhook est different dans son comportement et donc les clients doivent s'adapter dans la sematique et les erreurs etc la (meta-) structure de la payload, la taille, la securisation via signature (e.g. hmac), les erreurs (via erreurs HTTP), etc Data et Intelligence Artificielle Google annonce Gemini, son nouveau Large Language Model https://blog.google/technology/ai/google-gemini-ai/#sundar-note modèle multimodal qui peut prendre du texte, en entrée, mais aussi des images, du son, des vidéos d'après les benchmarks, il est largement aussi bon que GPT4 plusieurs tailles de modèles disponible : Nano pour être intégré aux mobiles, Pro qui va être utilisé dans la majeure partie des cas, et Ultra pour les besoins de réflexion les plus avancés Android va rajouter aussi des librairies AICore pour utiliser Gemini Nano dans les téléphones Pixel https://android-developers.googleblog.com/2023/12/a-new-foundation-for-ai-on-android.html Gemini Pro va être disponible dans Bard (en anglais et dans 170 pays, mais l'Europe va devoir attendre un petit peu pour que ce soit dispo) Gemini Ultra devrait aussi rejoindre Bard, dans une version étendue https://blog.google/products/bard/google-bard-try-gemini-ai/ Gemini va être intégré progressivement dans plein de produits Google DeepMind parlant de Gemini https://deepmind.google/technologies/gemini/#introduction Un rapport de 60 pages sur Gemini https://storage.googleapis.com/deepmind-media/gemini/gemini_1_report.pdf Gemini a permis aussi de pouvoir développer une nouvelle version du modèle AlphaCode qui excelle dans les compétitions de coding https://storage.googleapis.com/deepmind-media/AlphaCode2/AlphaCode2_Tech_Report.pdf Liste de petites vidéos sur YouTube avec des interviews et démonstrations des capacités de Gemini https://www.youtube.com/playlist?list=PL590L5WQmH8cSyqzo1PwQVUrZYgLcGZcG malheureusement certaines des annonces sont un peu fausse ce qui a amené un discrédit (non du) sur Gemini par exemple la video “aspirationelle” était vendue comme du réel mais ce n'est pas le cas. et ultra n'est pas disponible encore ausso la comparaison de ChatGPT sur la page (initialement au moins) comparait des choux et des carottes, meme si le papier de recherche était correct Avec la sortie de Gemini, Guillaume a écrit sur comment appeler Gemini en Java https://glaforge.dev/posts/2023/12/13/get-started-with-gemini-in-java/ Gemini est multimodèle, donc on peut passer aussi bien du texte que des images, ou même de la vidéo Il y a un SDK en Java pour interagir avec l'API de Gemini Facebook, Purple Llama https://ai.meta.com/blog/purple-llama-open-trust-safety-generative-ai/ Opensource https://ai.meta.com/llama/ dans l'optique des modeles GenAI ouverts, Facebook fournit des outils pour faire des IA responsables (mais pas coupables :wink: ) notament des benchmarks pour evaluler la sureté et un classifier de sureté, par exemple pour ne pas generer du code malicieux (ou le rendre plus dur) llama purple sera un projet parapluie D'ailleurs Meta IBM, Red Hat et pleins d'autres ont annoncé l'AI Alliance pour une AI ouverte et collaborative entre académique et industriels. Sont notammenrt absent Google, OpenAI (pas ouvert) et Microsoft Juste une annouce pour l'instant mais on va voir ce que ces acteurs de l'AI Alliance feront de concret il y a aussi un guide d'utilisateur l'usage IA responsable (pas lu) Apple aussi se met aux librairies de Machine Learning https://ml-explore.github.io/mlx/build/html/index.html MLX est une librairie Python qui s'inspire fortement de NumPy, PyTorch, Jax et ArrayFire Surtout, c'est développé spécifiquement pour les Macs, pour tirer au maximum parti des processeurs Apple Silicon Dans un des repos Github, on trouve également des exemples qui font tourner nativement sur macOS les modèles de Llama, de Mistral et d'auters https://github.com/ml-explore/mlx-examples non seulement les Apple Silicon amis aussi la memoire unifiee CPU/GPU qui est une des raisons clés de la rapidité des macs Faire tourner Java dans un notebook Jupyter https://www.javaadvent.com/2023/12/jupyter-notebooks-and-java.html Max Andersen explore l'utilisation de Java dans les notebooks Jupyter, au lieu du classique Python il y a des kernels java selon vos besoins mais il faut les installer dans la distro jupyter qu'on utilise et c'est la que jbang installable via pip vient a la rescousse il installe automatiquement ces kernels en quelques lignes Outillage Sfeir liste des jeux orientés développeurs https://www.sfeir.dev/tendances/notre-selection-de-jeux-de-programmation/ parfait pour Noël mais c'est pour ceux qui veulent continuer a challenger leur cerveau après le boulot jeu de logique, jeu de puzzle avec le code comme forme, jeu autour du machine learning, jeu de programmation assembleur Les calendriers de l'Avent sont populaires pour les développeurs ! En particulier avec Advent of Code https://adventofcode.com/ Mais il y a aussi l'Advent of Java https://www.javaadvent.com/ Ou un calendrier pour apprendre les bases de SVG https://svg-tutorial.com/ Le calendrier HTML “hell” https://www.htmhell.dev/adventcalendar/ qui parle d'accessibilité, de web components, de balises meta, de toutes les choses qu'on peut très bien faire en HTML/CSS sans avoir besoin de JavaScript Pour les développeurs TypeScript, il y a aussi un calendrier de l'Avent pour vous ! https://typehero.dev/aot-2023 Un super thread de Clara Dealberto sur le thème de la “dataviz” (data visualization) https://twitter.com/claradealberto/status/1729447130228457514 Beaucoup d'outil librement accessibles sont mentionnés pour faire toutes sortes de visualisations (ex. treemap, dendros, sankey…) mais aussi pour la cartographie Quelques ressources de site qui conseillent sur l'utilisation du bon type de visualisation en fonction du problème et des données que l'on a notemment celui du financial time qui tiens dans une page de PDF Bref c'est cool mais c'est long a lire Une petite liste d'outils sympas - jc pour convertir la sortie de commandes unix en JSON https://github.com/kellyjonbrazil/jc - AltTab pour macOS pour avoir le même comportement de basculement de fenêtre que sous Windows https://alt-tab-macos.netlify.app/ - gron pour rendre le JSON grep-able, en transformant chaque valeur en ligne ressemblant à du JSONPath https://github.com/tomnomnom/gron - Marker, en Python, pour transformer des PDF en beau Markdown https://github.com/VikParuchuri/marker - n8n un outil de workflow open source https://n8n.io/ gron en fait montre des lignes avec des assignments genre jsonpath = value et tu peux ungroner apres pour revenir a du json Marker utilise du machine learning mais il halklucine moins que nougat (nous voilà rassuré) Docker acquiert Testcontainers https://techcrunch.com/2023/12/11/docker-acquires-atomicjar-a-testing-startup-that-raised-25m-in-january/ Annonce par AtomicJar https://www.atomicjar.com/2023/12/atomicjar-is-now-part-of-docker/ Annonce par Docker https://www.docker.com/blog/docker-whale-comes-atomicjar-maker-of-testcontainers/ Architecture Comment implémenter la reconnaissance de chanson, comme Shazam https://www.cameronmacleod.com/blog/how-does-shazam-work il faut d'abord passer en mode fréquence avec des transformées de Fourrier pour obtenir des spectrogrammes puis créer une sorte d'empreinte qui rassemble des pics de fréquences notables à divers endroits de la chanson d'associer ces pics pour retrouver un enchainement de tels pics de fréquence dans le temps l'auteur a partagé son implémentation sur Github https://github.com/notexactlyawe/abracadabra/blob/e0eb59a944d7c9999ff8a4bc53f5cfdeb07b39aa/abracadabra/recognise.py#L80 Il y avait également une très bonne présentation sur ce thème par Moustapha Agack à DevFest Toulouse https://www.youtube.com/watch?v=2i4nstFJRXU les pics associés sont des hash qui peut etre comparés et le plus de hash veut dire que les chansons sont plus similaires Méthodologies Un mémo de chez ThoughtWorks à propos du coding assisté par IA https://martinfowler.com/articles/exploring-gen-ai.html#memo-08 Avec toute une liste de questions à se poser dans l'utilisation d'un outil tel que Copilot Il faut bien réaliser que malheureusement, une IA n'a pas raison à 100% dans ses réponses, et même plutôt que la moitié du temps, donc il faut bien mettre à jour ses attentes par rapport à cela, car ce n'est pas magique La conclusion est intéressante aussi, en suggérant que grosso modo dans 40 à 60% des situations, tu peux arriver à 40 à 80% de la solution. Est-ce que c'est à partir de ce niveau là qu'on peut vraiment gagner du temps et faire confiance à l'IA ? Ne perdez pas trop de temps non plus à essayer de convaincre l'IA de faire ce que vous voulez qu'elle fasse. Si vous n'y arrivez pas, c'est sans doute parce que l'IA n'y arrivera même pas elle même ! Donc au-delà de 10 minutes, allez lire la doc, chercher sur Google, etc. notamment, faire genrer les tests par l'IA dans al foulée augmente les risques surtout si on n'est pas capable de bien relire le code si on introduit un choix de pattern genre flexbox en CSS, si c'est sur une question de sécuriter, vérifier (ceinture et bretelle) est-ce le framework de la semaine dernière? L'info ne sera pas dans le LLM (sans RAG) Quelles capacités sont nécessaires pour déployer un projet AI/ML https://blog.scottlogic.com/2023/11/22/capabilities-to-deploy-ai-in-your-organisation.html C'est le MLOps et il y a quelques modèles end to end Google, IBM mais vu la diversité des organisations, c'est difficile a embrasser ces versions completes ML Ops est une métier, data science est un metier, donc intégrer ces competences sachez gérer votre catalogue de données Construire un process pour tester vos modèles et continuellement La notion de culture de la recherche et sa gestion (comme un portefeuille financier, accepter d'arrêter des experience etc) la culture de la recherche est peu présente en engineering qui est de construire des choses qui foncitonnent c'est un monde pre LLM Vous connaissez les 10 dark patterns de l'UX ? Pour vous inciter à cliquer ici ou là, pour vous faire rester sur le site, et plus encore https://dodonut.com/blog/10-dark-patterns-in-ux-design/ Parmi les dark patterns couverts Confirmshaming Fake Urgency and the Fear of Missing Out Nagging Sneaking Disguised Ads Intentional Misdirection The Roach Motel Pattern Preselection Friend Spam Negative Option Billing or Forced Continuity L'article conclut avec quelques pistes sur comment éviter ces dark patterns en regardant les bons patterns de la concurrence, en testant les interactions UX, et en applicant beaucoup de bon sens ! les dark patterns ne sont pas des accidents, ils s'appuient sur la psychologie et sont mis en place specifiquement Comment choisir de belles couleurs pour la visualisation de données ? https://blog.datawrapper.de/beautifulcolors/ Plutôt que de penser en RGB, il vaut mieux se positionner dans le mode Hue Saturation Brightness Plein d'exemples montrant comment améliorer certains choix de couleurs Mieux vaut éviter des couleurs trop pures ou des couleurs trop brillantes et saturées Avoir un bon contraste Penser aussi aux daltoniens ! j'ai personnellement eu toujours du mal avec saturationm vs brightness faire que les cloueirs en noir et blanc soient separees evant de le remettre (en changeant la brightness de chaque couleur) ca aide les daltoniens eviter les couleurs aux 4 coins amis plutot des couleurs complementaires (proches) rouge orange et jaune (non saturé) et variations de bleu sont pas mal les couleurs saturées sont aggressives et stressent les gens Pourquoi vous devriez devenir Engineering Manager? https://charity.wtf/2023/12/15/why-should-you-or-anyone-become-an-engineering-manager/ L'article parle de l'évolution de la perception de l'engineering management qui n'est plus désormais le choix de carrière par défaut pour les ingénieurs ambitieux. Il met en évidence les défis auxquels les engineering managers sont confrontés, y compris les attentes croissantes en matière d'empathie, de soutien et de compétences techniques, ainsi que l'impact de la pandémie de COVID-19 sur l'attrait des postes de management. L'importance des bons engineering mnanagers est soulignée, car ils sont considérés comme des multiplicateurs de force pour les équipes, contribuant de manière significative à la productivité, à la qualité et au succès global dans les environnements organisationnels complexes. L'article fournit des raisons pour lesquelles quelqu'un pourrait envisager de devenir Engineering Manager, y compris acquérir une meilleure compréhension de la façon dont les entreprises fonctionnent, contribuer au mentorat et influencer les changements positifs dans la dynamique des équipes et les pratiques de l'industrie. Une perspective est présentée, suggérant que devenir Engineering manager peut conduire à la croissance personnelle et à l'amélioration des compétences de vie, telles que l'autorégulation, la conscience de soi, la compréhension des autres, l'établissement de limites, la sensibilité à la dynamique du pouvoir et la maîtrise des conversations difficiles. L'article encourage à considérer la gestion comme une occasion de développer et de porter ces compétences pour la vie. Sécurité LogoFAIL une faille du bootloader de beaucoup de machines https://arstechnica.com/security/2023/12/just-about-every-windows-and-linux-device-vulnerable-to-new-logofail-firmware-attack/ en gros en changeant les eimages qu'on voit au boot permet d'executer du code arbitraire au tout debuit de la securisation du UEFI (le boot le plus utilisé) donc c'est game over parce que ca demarre avant l'OS c'est pas une exploitation a distance, il faut etre sur la machine avec des droits assez elevés deja mais ca peut etre la fin de la chaine d'attaque et comme d'hab un interpreteur d'image est la cause de ces vulnerabilités Conférences L'IA au secours de conférences tech: rajoute des profile tech femme comme speaker au programme pour passer le test diversité online via des profiles fake. https://twitter.com/GergelyOrosz/status/1728177708608450705 https://www.theregister.com/2023/11/28/devternity_conference_fake_speakers/ https://www.developpez.com/actu/351260/La-conference-DevTernity-sur-la-technologie-s-e[…]s-avoir-cree-de-fausses-oratrices-generees-automatiquement/ j'avais lu le tweet du createur de cette conf qui expliquait que c'etait des comptes de tests et que pris dans le rush ils avaient oublié de les enlever mais en fait les comptes de tests ont des profils “Actifs” sur le reseaux sociaux apparemment donc c'était savamment orchestré Au final beaucoup de speakers et des sponsors se desengagent La liste des conférences provenant de Developers Conferences Agenda/List par Aurélie Vache et contributeurs : 31 janvier 2024-3 février 2024 : SnowCamp - Grenoble (France) 1 février 2024 : AgiLeMans - Le Mans (France) 6 février 2024 : DevFest Paris - Paris (France) 8-9 février 2024 : Touraine Tech - Tours (France) 15-16 février 2024 : Scala.IO - Nantes (France) 6-7 mars 2024 : FlowCon 2024 - Paris (France) 14-15 mars 2024 : pgDayParis - Paris (France) 19 mars 2024 : AppDeveloperCon - Paris (France) 19 mars 2024 : ArgoCon - Paris (France) 19 mars 2024 : BackstageCon - Paris (France) 19 mars 2024 : Cilium + eBPF Day - Paris (France) 19 mars 2024 : Cloud Native AI Day Europe - Paris (France) 19 mars 2024 : Cloud Native Wasm Day Europe - Paris (France) 19 mars 2024 : Data on Kubernetes Day - Paris (France) 19 mars 2024 : Istio Day Europe - Paris (France) 19 mars 2024 : Kubeflow Summit Europe - Paris (France) 19 mars 2024 : Kubernetes on Edge Day Europe - Paris (France) 19 mars 2024 : Multi-Tenancy Con - Paris (France) 19 mars 2024 : Observabiity Day Europe - Paris (France) 19 mars 2024 : OpenTofu Day Europe - Paris (France) 19 mars 2024 : Platform Engineering Day - Paris (France) 19 mars 2024 : ThanosCon Europe - Paris (France) 19-21 mars 2024 : IT & Cybersecurity Meetings - Paris (France) 19-22 mars 2024 : KubeCon + CloudNativeCon Europe 2024 - Paris (France) 26-28 mars 2024 : Forum INCYBER Europe - Lille (France) 28-29 mars 2024 : SymfonyLive Paris 2024 - Paris (France) 4-6 avril 2024 : Toulouse Hacking Convention - Toulouse (France) 17-19 avril 2024 : Devoxx France - Paris (France) 18-20 avril 2024 : Devoxx Greece - Athens (Greece) 25-26 avril 2024 : MiXiT - Lyon (France) 25-26 avril 2024 : Android Makers - Paris (France) 8-10 mai 2024 : Devoxx UK - London (UK) 16-17 mai 2024 : Newcrafts Paris - Paris (France) 24 mai 2024 : AFUP Day Nancy - Nancy (France) 24 mai 2024 : AFUP Day Poitiers - Poitiers (France) 24 mai 2024 : AFUP Day Lille - Lille (France) 24 mai 2024 : AFUP Day Lyon - Lyon (France) 2 juin 2024 : PolyCloud - Montpellier (France) 6-7 juin 2024 : DevFest Lille - Lille (France) 6-7 juin 2024 : Alpes Craft - Grenoble (France) 27-28 juin 2024 : Agi Lille - Lille (France) 4-5 juillet 2024 : Sunny Tech - Montpellier (France) 19-20 septembre 2024 : API Platform Conference - Lille (France) & Online 7-11 octobre 2024 : Devoxx Belgium - Antwerp (Belgium) 10-11 octobre 2024 : Volcamp - Clermont-Ferrand (France) 10-11 octobre 2024 : Forum PHP - Marne-la-Vallée (France) 17-18 octobre 2024 : DevFest Nantes - Nantes (France) Nous contacter Pour réagir à cet épisode, venez discuter sur le groupe Google https://groups.google.com/group/lescastcodeurs Contactez-nous via twitter https://twitter.com/lescastcodeurs 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/

PROMO Instant Gaming: https://www.instant-gaming.com/?igr=kristalcrossgaming CAPITOLI: 00:00 Benvenuti su Kristal News! 00:24 The Last of Us Multiplayer, QUINDI è ANCORA "VIVO"? 06:15 Detective Cross 08:02 BLOODBORNE, QUALCOSA si MUOVE? 13:07 PS5 PRO, ALTRI LEAK CPU/GPU 17:37 Ho bisogno del vostro aiuto. 22:28 Detective Cross 23:14 NEWS A RUOTA 32:20 Offerta del GIORNO 33:21 Grazie della visione! Offerta del giorno: For The King 2 ▶ -30% (17,49€) https://www.instant-gaming.com/it/13618-comprare-steam-for-the-king-ii-pc-gioco-steam-europe/?igr=kristalcrossgaming Le #KristalNews sono anche in PODCAST! https://www.spreaker.com/show/kristalnews-il-podcast Disponibile anche su SPOTIFY, APPLE PODCAST, ecc ecc PLAYLIST Kristal News https://www.youtube.com/playlist?list=PLiAZAlZgYAdpnFFrYRcfnqVI6WCld_E5g SECONDO CANALE: https://www.youtube.com/channel/UCO13aJlGKgOVYq7-yMe_lPA SEGUIMI IN LIVE SU TWITCH: https://www.twitch.tv/kristalcrossgaming SUPPORTAMI SU PATREON! https://www.patreon.com/kristalcrossgaming SOCIAL & COMMUNITY: FACEBOOK: https://www.facebook.com/kristalcrossgaming/ INSTAGRAM: https://www.instagram.com/kristalcrossgaming/ GRUPPO TELEGRAM: https://t.me/joinchat/APAtYQ5DPmVA3AHTl8mBcA VUOI GIOCHI SUPERSCONTATI? CLICCA QUI! https://www.instant-gaming.com/igr/kristalcrossgaming/ https://www.cdkeys.com/?mw_aref=kristalcrossgaming https://www.epicgames.com/store/it/ (TAG: KRISTAL-CROSS) LINK AFFILIATO AMAZON: https://amzn.to/3tXIq3d SPONSOR - ABBONAMENTO AL CANALE: https://www.youtube.com/channel/UC-isEtNtS2_30c5ANuEEyKQ/join DONAZIONI: https://paypal.me/kristalcross VUOI SPEDIRMI QUALCOSA? MANDALO A QUESTO INDIRIZZO! Flow Rock Live SRLS (per Kristal) Via Fioravanti, 3 San Benedetto del Tronto, AP 63074 "LA CRICCA DEL SOTTOSUOLO": Link Utili! DISCORD: https://discord.gg/BD7vqk TELEGRAM: https://t.me/LaCriccadelSottosuolo FACEBOOK: https://www.facebook.com/LaCriccaDelSottosuolo/ ---------------------------------------- VUOI SENTIRE I MIEI LAVORI? IL SITO DEL MIO STUDIO http://www.kristalcrossmusicproduction.com/page0/page0.html LE MIE BAND: THE LOTUS: https://www.youtube.com/user/TheLotusChannel LOST RESONANCE FOUND: https://www.youtube.com/user/lostresonancefound Attenzione, quasi tutte le musiche che sentite nei video (quando presenti) sono o fatte da me, o dalle band di cui faccio parte, o delle band registrate e/o mixate da me nel mio studio. Uso Final Cut Pro X Per editare i video e Logic Pro X per l'editing audio. Grazie per la visita da Kristal Cross!

QualcommがPC向けハイエンドSoC「Snapdragon X Elite」を発表　CPUもGPUもNPUも高速なのに省電力。　Qualcommは10月24日（ハワイ時間）、PC向けの新型SoC（System-on-a-Chip）「Snapdragon X Elite」を発表した。搭載製品は2024年中盤から順次登場する見通しだ。

Benjamin and Chance discuss the iOS 17 Accessibility features Apple previewed this week. Apple also unveiled some new live concert features for Apple Music. Rumors give more color on the iPhone 16 spec situation, and anticipation for M3 Apple Silicon Macs grow. Plus, Chance goes hands-on with the newly-announced Beats Studio Buds+. Sponsored by Zocdoc: Go to Zocdoc.com/happyhour and download the Zocdoc app to sign-up for free and book a top-rated doctor. Many are available as soon as today. Sponsored by Fast Growing Trees: Join 1.5 million happy customers and save 15% off your perfect plants, shrubs, and trees! Follow Twitter: @ChanceHMiller Mastodon: @chancehmiller@mastodon.social Benjamin Mayo @bzamayo or @bzamayo@mastodon.social Read More Apple previews iOS 17 accessibility features: Assistive Access, Personal Voice and Live Speech, more Review: The new transparent Beats Studio Buds Plus First M3 Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Subscribe or Follow Apple Podcasts Overcast Spotify

Benjamin and Chance discuss the iOS 17 Accessibility features Apple previewed this week. Apple also unveiled some new live concert features for Apple Music. Rumors give more color on the iPhone 16 spec situation, and anticipation for M3 Apple Silicon Macs grow. Plus, Chance goes hands-on with the newly-announced Beats Studio Buds+. Sponsored by Zocdoc: Go to Zocdoc.com/happyhour and download the Zocdoc app to sign-up for free and book a top-rated doctor. Many are available as soon as today. Sponsored by Fast Growing Trees: Join 1.5 million happy customers and save 15% off your perfect plants, shrubs, and trees! Follow Twitter: @ChanceHMiller Mastodon: @chancehmiller@mastodon.social Benjamin Mayo @bzamayo or @bzamayo@mastodon.social Read More Apple previews iOS 17 accessibility features: Assistive Access, Personal Voice and Live Speech, more Review: The new transparent Beats Studio Buds Plus First M3 Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Subscribe or Follow Apple Podcasts Overcast Spotify

App Store stopped more than $2 billion in fraudulent transactions in 2022. Apple launches new concert discovery features on Apple Maps and Apple Music. Apple introduces new features for cognitive accessibility, along with Live Speech, Personal Voice, and Point and Speak in Magnifier. Apple devices to occupy 90% of TSMC 3nm capacity in 2023 First M3 Pro Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Pro. Apple executive in charge of TV+ and sports businesses to depart. Apple service outages are at unacceptable level given the importance of the ecosystem. Leaker claims Apple used 'multi-step sting' operation to identify and fire their source ahead of WWDC. Most states halt use of Google and Apple's COVID-19 exposure notification system.  Picks of the Week Alex Pick: Ainope USBC 3.2 Gen 2 Andy's Pick: Whats.New Jason's Pick: Laya's Horizon Hosts: Leo Laporte, Alex Lindsay, Andy Ihnatko, and Jason Snell Download or subscribe to this show at https://twit.tv/shows/macbreak-weekly. Get episodes ad-free with Club TWiT at https://twit.tv/clubtwit Sponsors: hellofresh.com/macbreak16 and use code macbreak16 Miro.com/podcast

App Store stopped more than $2 billion in fraudulent transactions in 2022. Apple launches new concert discovery features on Apple Maps and Apple Music. Apple introduces new features for cognitive accessibility, along with Live Speech, Personal Voice, and Point and Speak in Magnifier. Apple devices to occupy 90% of TSMC 3nm capacity in 2023 First M3 Pro Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Pro. Apple executive in charge of TV+ and sports businesses to depart. Apple service outages are at unacceptable level given the importance of the ecosystem. Leaker claims Apple used 'multi-step sting' operation to identify and fire their source ahead of WWDC. Most states halt use of Google and Apple's COVID-19 exposure notification system.  Picks of the Week Alex Pick: Ainope USBC 3.2 Gen 2 Andy's Pick: Whats.New Jason's Pick: Laya's Horizon Hosts: Leo Laporte, Alex Lindsay, Andy Ihnatko, and Jason Snell Download or subscribe to this show at https://twit.tv/shows/macbreak-weekly. Get episodes ad-free with Club TWiT at https://twit.tv/clubtwit Sponsors: hellofresh.com/macbreak16 and use code macbreak16 Miro.com/podcast

App Store stopped more than $2 billion in fraudulent transactions in 2022. Apple launches new concert discovery features on Apple Maps and Apple Music. Apple introduces new features for cognitive accessibility, along with Live Speech, Personal Voice, and Point and Speak in Magnifier. Apple devices to occupy 90% of TSMC 3nm capacity in 2023 First M3 Pro Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Pro. Apple executive in charge of TV+ and sports businesses to depart. Apple service outages are at unacceptable level given the importance of the ecosystem. Leaker claims Apple used 'multi-step sting' operation to identify and fire their source ahead of WWDC. Most states halt use of Google and Apple's COVID-19 exposure notification system.  Picks of the Week Alex Pick: Ainope USBC 3.2 Gen 2 Andy's Pick: Whats.New Jason's Pick: Laya's Horizon Hosts: Leo Laporte, Alex Lindsay, Andy Ihnatko, and Jason Snell Download or subscribe to this show at https://twit.tv/shows/macbreak-weekly. Get episodes ad-free with Club TWiT at https://twit.tv/clubtwit Sponsors: hellofresh.com/macbreak16 and use code macbreak16 Miro.com/podcast

App Store stopped more than $2 billion in fraudulent transactions in 2022. Apple launches new concert discovery features on Apple Maps and Apple Music. Apple introduces new features for cognitive accessibility, along with Live Speech, Personal Voice, and Point and Speak in Magnifier. Apple devices to occupy 90% of TSMC 3nm capacity in 2023 First M3 Pro Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Pro. Apple executive in charge of TV+ and sports businesses to depart. Apple service outages are at unacceptable level given the importance of the ecosystem. Leaker claims Apple used 'multi-step sting' operation to identify and fire their source ahead of WWDC. Most states halt use of Google and Apple's COVID-19 exposure notification system.  Picks of the Week Alex Pick: Ainope USBC 3.2 Gen 2 Andy's Pick: Whats.New Jason's Pick: Laya's Horizon Hosts: Leo Laporte, Alex Lindsay, Andy Ihnatko, and Jason Snell Download or subscribe to this show at https://twit.tv/shows/macbreak-weekly. Get episodes ad-free with Club TWiT at https://twit.tv/clubtwit Sponsors: hellofresh.com/macbreak16 and use code macbreak16 Miro.com/podcast

App Store stopped more than $2 billion in fraudulent transactions in 2022. Apple launches new concert discovery features on Apple Maps and Apple Music. Apple introduces new features for cognitive accessibility, along with Live Speech, Personal Voice, and Point and Speak in Magnifier. Apple devices to occupy 90% of TSMC 3nm capacity in 2023 First M3 Pro Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Pro. Apple executive in charge of TV+ and sports businesses to depart. Apple service outages are at unacceptable level given the importance of the ecosystem. Leaker claims Apple used 'multi-step sting' operation to identify and fire their source ahead of WWDC. Most states halt use of Google and Apple's COVID-19 exposure notification system.  Picks of the Week Alex Pick: Ainope USBC 3.2 Gen 2 Andy's Pick: Whats.New Jason's Pick: Laya's Horizon Hosts: Leo Laporte, Alex Lindsay, Andy Ihnatko, and Jason Snell Download or subscribe to this show at https://twit.tv/shows/macbreak-weekly. Get episodes ad-free with Club TWiT at https://twit.tv/clubtwit Sponsors: hellofresh.com/macbreak16 and use code macbreak16 Miro.com/podcast

Listen to a recap of the top stories of the day from 9to5Mac. 9to5Mac Daily is available on iTunes and Apple's Podcasts app, Stitcher, TuneIn, Google Play, or through our dedicated RSS feed for Overcast and other podcast players. Sponsored by Kolide: Kolide's method means fewer support tickets, less frustration, and most importantly: 100% fleet compliance. Visit  kolide.com/9to5macdaily  to learn more or book a demo. New episodes of 9to5Mac Daily are recorded every weekday. Subscribe to our podcast in Apple Podcast or your favorite podcast player to guarantee new episodes are delivered as soon as they're available. Stories discussed in this episode: First M3 Macs may launch by end of the year, chip has more CPU/GPU cores compared to M2 Apple already anticipating more Reality Pro headset delays iPhone 14 satellite features now available in Australia and New Zealand Follow Chance: Twitter: @ChanceHMiller Mastodon: @chancehmiller@mastodon.social Listen & Subscribe: Apple Podcasts Overcast RSS Spotify TuneIn Google Podcasts Catch up on 9to5Mac Daily episodes! Enjoy the podcast? Shop Apple at Amazon Don't miss out on our other daily podcasts: Quick Charge 9to5Toys Daily The Buzz Share your thoughts! Drop us a line at happyhour@9to5mac.com. You can also rate us in Apple Podcasts or recommend us in Overcast to help more people discover the show.

In this episode of the AI Today podcast hosts Kathleen Walch and Ron Schmelzer define the terms CPU, GPU, TPU, and Federated Learning, explain how these terms relate to AI and why it's important to know about them. Show Notes: FREE Intro to CPMAI mini course CPMAI Training and Certification AI Glossary Glossary Series: Artificial Intelligence AI Glossary Series – Machine Learning, Algorithm, Model Glossary Series: (Artificial) Neural Networks, Node (Neuron), Layer Glossary Series: Natural Language Processing (NLP), NLU, NLG, Speech-to-Text, TTS, Speech Recognition Continue reading AI Today Podcast: AI Glossary Series – CPU, GPU, TPU, and Federated Learning at AI & Data Today.

ElixirConf US 2022 just finished! We cover the big announcements, talk highlights, and other relevant tech news. We discuss what some of these big announcements and projects represent and what they might mean for the Elixir community going forward. We talk about the Elixir 1.14 release, Livebook advances, Phoenix 1.7, machine learning progress, and the surprise announcement of Phoenix LiveView Native! Show Notes online - http://podcast.thinkingelixir.com/115 (http://podcast.thinkingelixir.com/115) Elixir Community News - https://elixir-lang.org/blog/2022/09/01/elixir-v1-14-0-released/ (https://elixir-lang.org/blog/2022/09/01/elixir-v1-14-0-released/) – Elixir v1.14 officially released - https://github.com/elixir-lang/elixir/blob/v1.14.0/CHANGELOG.md#changelog-for-elixir-v114 (https://github.com/elixir-lang/elixir/blob/v1.14.0/CHANGELOG.md#changelog-for-elixir-v114) – Elixir 1.14 changelog - https://github.com/elixir-lang/elixir/blob/v1.14.0/CHANGELOG.md#changelog-for-elixir-v114 (https://github.com/elixir-lang/elixir/blob/v1.14.0/CHANGELOG.md#changelog-for-elixir-v114) – Nerves v1.9.0 fixed Elixir 1.14 warnings - Phoenix 1.7 upcoming release discussed - Phoenix 1.7 generators will use Tailwind CSS - New phx.gen.auth --live option - https://github.com/liveviewnative/liveview-client-swiftui (https://github.com/liveviewnative/liveview-client-swiftui) – Phoenix LiveView Native was announced - https://github.com/liveviewnative/elixirconf_chat (https://github.com/liveviewnative/elixirconf_chat) – ElixirConf Chat project created using Phoenix LiveView Native - https://getfirefly.org (https://getfirefly.org) – Lumen was renamed to Firefly - https://twitter.com/HoldenOullette/status/1565486046237921280 (https://twitter.com/HoldenOullette/status/1565486046237921280) – Podium released an OWASP security training LiveBook for Elixir developers. - https://github.com/podium/elixir-secure-coding (https://github.com/podium/elixir-secure-coding) – Elixir Secure Coding Training (ESCT) - https://www.ectoinproduction.com (https://www.ectoinproduction.com) – Ecto In Production future home - https://github.com/liveshowy/webauthnlivecomponent (https://github.com/liveshowy/webauthn_live_component) – SmartLogic released a LiveComponent to support WebAuthn authentication for your LiveView app - https://github.com/liveshowy/webauthnlivecomponent_demo (https://github.com/liveshowy/webauthn_live_component_demo) – WebAuthn authentication demo page - https://github.com/kipcole9/tempo (https://github.com/kipcole9/tempo) – Kip Cole released a new kind of DateTime library called Tempo - https://kipcole9.github.io/tempo/2021-01-04-its-about-time/ (https://kipcole9.github.io/tempo/2021-01-04-its-about-time/) – Temp blog post explains more about it. - https://twitter.com/steveschoger/status/1562117153591107586 (https://twitter.com/steveschoger/status/1562117153591107586) – Heroicons v2.0 released. Used in TailwindUI templates. - https://twitter.com/louispilfold/status/1564247740879609860 (https://twitter.com/louispilfold/status/1564247740879609860) – Louie Pilford showed a screenshot of Gleam compiling Elixir's Plug - https://blog.heroku.com/next-chapter (https://blog.heroku.com/next-chapter) – Heroku, a popular PaaS made significant policy changes. Ending free tier and more. - https://spectrum.ieee.org/top-programming-languages-2022 (https://spectrum.ieee.org/top-programming-languages-2022) – IEEE Top Programming Languages 2022 - https://twitter.com/josevalim/status/1565408635961884673 (https://twitter.com/josevalim/status/1565408635961884673) – José Valim shared they are porting non-neural algorithms to Elixir/Nx which runs on both CPU/GPU. Shared impressive performance comparisons. - Chris Grainger gave a keynote about how Elixir is ready for real, production machine learning work. - https://www.lambdadays.org/lambdadays2022 (https://www.lambdadays.org/lambdadays2022) – Lambda Days conference. 5-6 June 2023 in Krakow, Poland Do you have some Elixir news to share? Tell us at @ThinkingElixir (https://twitter.com/ThinkingElixir) or email at show@thinkingelixir.com (mailto:show@thinkingelixir.com) Find us online - Message the show - @ThinkingElixir (https://twitter.com/ThinkingElixir) - Email the show - show@thinkingelixir.com (mailto:show@thinkingelixir.com) - Mark Ericksen - @brainlid (https://twitter.com/brainlid) - David Bernheisel - @bernheisel (https://twitter.com/bernheisel) - Cade Ward - @cadebward (https://twitter.com/cadebward)

The team catches up with Bill Paxson and Rob Futrick to learn about Azure HPC and the unique way in which the Microsoft Support team is supports HPC customers     Media file: https://azpodcast.blob.core.windows.net/episodes/Episode435.mp3 YouTube: https://youtu.be/1y1iDUvx75A Resources: High-performance computing (HPC) on Azure - Azure Architecture Center | Microsoft Docs High Performance Computing – HPC | Microsoft Azure   Free MS Learn modules covering many HPC AI topics, from intro to HPC to advanced optimization: Run high-performance computing (HPC) applications on Azure - Learn | Microsoft Docs Overview from HPC AI Day conference earlier this year: (101) Azure HPC+AI Day: HPC + AI software - YouTube Great overview of the entire HPC platform, including CPU + GPU hardware, HPC Storage considerations + options, etc.  https://www.youtube.com/watch?v=U54kF0EZ9DM Azure CycleCloud specifically:  https://www.youtube.com/watch?v=SyrsXUwm0vA. It has a live demo 19 minutes in.   Other updates: Monday, August 15, 2022 9:00 PM Use Container Storage Interface (CSI) driver for Azure Blob storage on Azure Kubernetes Service (AKS) - Azure Kubernetes Service | Microsoft Docs     Azure Data Explorer: Log and telemetry analytics benchmark https://azure.microsoft.com/en-us/blog/azure-data-explorer-log-and-telemetry-analytics-benchmark/   Announcing Microsoft Dev Box Preview   https://azure.microsoft.com/en-us/blog/announcing-microsoft-dev-box-preview/

Data are not all created equal. There are structured (machine-generated) and unstructured (human-generated) data, and each requires a different approach. As a result of the increased use of digital applications and services, unstructured information is growing rapidly. It is estimated that 80-90% of company data is unstructured, and the amount of unstructured data continues to grow at an alarming rate per year. In this episode of The Analytic Mind Podcast, Sam McKay speaks with Kirk Marple, the founder, and CEO of Unstruk Data. Hear about the hidden value in unstructured data and how it impacts the future of data and analytics. Kirk Marple is a Customer-focused Technology leader with extensive expertise with cloud-based microservices, scalable multimedia data ingestion, knowledge graphs (entity extraction/enrichment), machine learning and computer vision integration, and CPU/GPU-based file and data processing workflows. He is highly skilled at architecting and managing teams developing APIs and scalable services, leveraging on-premise and cloud infrastructure, generating significant cost savings, and supporting revenue growth for startups through Fortune 25 companies. Unstruk Data, is a company that is building the industry's leading Unstructured Data Warehouse for automating data preparation via metadata enrichment, integrated compute, and graph-based search. Unstructured data can include anything from audio to visual data, and Kirk explains that analytics can be done on geospatial and temporal unstructured data. Then by using machine learning and artificial intelligence, it is possible to generate knowledge graphs based on this data.

Kirk Marple is the CEO and Founder of Unstruk Data, a new company that is Building the industry's leading Unstructured Data Warehouse for automating data preparation via metadata enrichment, integrated compute, and graph-based search. As a customer-focused technology leader, Kirk has extensive expertise in cloud-based microservices, scalable multimedia data ingestion, knowledge graphs (entity extraction/enrichment), machine learning and computer vision integration, and CPU/GPU-based file and data processing workflows. Kirk has over 25 years of experience developing media management pipelines, leading DevOps at venture backed companies and structuring successful exits. He holds multiple patents and industry awards, and has truly established himself as an industry leader.

By Michael G. Hull Modern technology has evolved at an even faster pace in the last decade generating cutting-edge innovations and breakthroughs – from artificial intelligence, augmented and virtual reality, decentralized finance, to blockchain and crypto technologies. Today, the world is highly connected more than ever. With the emergence of the next-generation cellular network, 5G is the latest iteration of the global wireless technology standard. Numerous countries are securing the availability of 5G coverage in their areas while there are places with limited availability. In 2021, South Korea, Saudi Arabia, Kuwait, Hong Kong, and the USA lead the countries with 5G network availability and user connection. Addressing Speed and Latency with Edge Computing and 5G Connection Along with the advancement of mobile network technology, the 5G network offers faster data speeds, bigger capacity, reduced latency, and improved mobile network services than its 4G predecessor. The 5G network enables significantly improved features of the previous networks to deliver a better user experience and connectivity. In comparing 4G and 5G speeds, the average 4G speed rate can reach from 10-50 Mbps (megabits per second) with a maximum speed ranging from 150 Mbps-1Gbps (gigabits per second) while the average 5G speed rate starts from 50 Mbps with a max speed rate of 1-10 Gbps. Speed rates may also depend on factors like wireless network connection, geographical location, and device unit. The latest 5G network has another major improvement: low latency. Latency is the measured gap or delay between the data transmission and response. Under the 5G network, reduced latency means faster response to transmitted data in real time, improving the overall mobile network user experience, and opening up opportunities across various industries. Nope — this is not an April fool's joke ? $SNE is coming to #LATOKEN this Monday, April 4th. Stay tuned for the official details ? StrongNode.io (@StrongNodeEdge) April 1, 2022 StrongNode powers mobile edge computing The mobile network industry is fast becoming decentralized with the rise of the 5G network. Shifting to a distributed and decentralized model places the transmission and processing of data towards the edge or closer to the data source with the help of edge computing technology. 5G and edge computing goes hand-in-hand as 5G is necessary to improve mobile broadband network performance and minimize latency for faster processing of high-volume data and increased, even instant, response time. 5G mobile edge computing can provide improved services and performances in gaming on 5G-enabled devices, telecommunications, IoT sensors from manufacturing plants, food delivery, video streaming, media, and entertainment among others. This is where StrongNode.io, an Infrastructure-as-a-Service tech company comes in. Leveraging edge computing technology, StrongNode aims to address high latency with the optimization of devices such as laptops, mobile phones, tablets, and eventually smart TV. The StrongNode Edge technology will harness the underutilized computing resources. CPU/GPU cycles, network bandwidth, and data storage – of digital devices commonly found in households, schools, and enterprises to build an edge network that can power the computing needs of other individuals, institutions, and organizations. With the proliferation of 5G-enabled devices, there will be an increase in the efficiency of use in computing resources leaving a higher percentage that can be shared with other users in the network. StrongNode CEO and Co-founder Daniel Saito said that with the rollout of 5G/6G technology, StrongNode could develop products and services that will help democratize computing for the masses regardless of the digital divide between different societies. With 5G/6G standards designed to have 10x or more faster network speed than the current 4G network, then 5G/6G can potentially enhance not only mobile broadband services ...

Follow-up: iPad Air does have stereo speakers Mac Studio HDMI is still 2.0 HDMI 2.1 27” Studio Display Pro coming soon? “Five-head” Origins The Flop House #94: Conan the Barbarian @ 29:46 Victor Borge’s Inflationary Language 20 Fenchurch Street The A7III does have animal eye detection (via Alec Pulianas) Base Mac Pro gets a spec bump MaskerAid Updates Photo library access Method Swizzling ARM Mac Pro worst-case scenarios Max Tech: Apple M2 Duo? Patent: Systems and Methods for Interconnecting Dies Patent: Fully Interconnected Heterogeneous Multi-layer Reconstructed Silicon Device Majin Bu’s tweet Universal Control Impressions Synergy #askatp: Did the DTK set reasonable expectations for M1 devices? (via Ryan Petriello) Does switching to higher-resolution displays put a strain on the CPU/GPU? (via Peter Wagoner) What are the three ugliest Macs of all time? (via Wade) John Molar Mac (Power Macintosh G3 All-In-One) LC 580 Casey Polybook Flower Power All 13 Colors of iMac 2004 17” iMac G5 Mac Pro Marco 2006 17” MacBook Pro John’s PowerBook G4 Review 11” MacBook Air 2004 17” iMac G5 2019 Mac Pro Rack-mount Version Post-show: A preview of Casey’s new setup (really) with a CalDigit TS4 Sponsored by: Squarespace: Make your next move. Use code atp for 10% off your first order. Linode: Instantly deploy and manage an SSD server in the Linode Cloud. New accounts get a $100 credit. Kolide: Meet compliance objectives in a remote-first world without resorting to rigid device management. Try Kolide for free! Become a member for ad-free episodes and our early-release, unedited “bootleg” feed! Become a member!

初の公開収録回です。「本は頭に今日のタスクがないうちに読め」「堀元のでまかせ力は〇〇に出続けたから」「CPUは高橋名人、GPUは赤ちゃん20億人」など参加者からの質問に直接答えつつ進めていきます。 【目次】 00:00 オープニング。初の公開収録がスタート！ 03:13 何でコミュニティをYouTubeのメンバーシップでやらないの？ 11:19 情報整理について語ったらメタモンが暴走した 16:55 モヤリが人生を切り開く(？) 19:28 堀元、水野はいつ本を読んでいる？ 24:32 リクエスト「消えた言語について話して」ウ～ン1年後ね。 27:57 コミュニティ内の自由研究ラジオが楽しい 30:10 どこまで台本？どこからアドリブ？ 34:58 このラジオ年配の人にも好評だってよ 39:25 誕生の前に命名するのは〇〇のため 46:55 エンディング。コミュニティメンバー募集中！ 【サポーターコミュニティへの加入はこちらから！】 https://yurugengo.com/support 【姉妹チャンネル：ゆるコンピュータ科学ラジオ】 https://www.youtube.com/channel/UCpLu0KjNy616-E95gPx7LZg 【Twitterあるよ！】 ゆる言語学ラジオのTwitterアカウントがあるので、是非フォローしてください！面白語源ネタなどが流れてきてあなたの知識欲が満たされます。 →https://twitter.com/yuru_gengo 【おたよりフォーム】 https://forms.gle/mTGM7A9QNqgjZMgN7 ※皆様からの楽しいおたよりをお待ちしています！ 【お仕事依頼はこちら！】 yurugengo@gmail.com 【堀元見プロフィール】 慶應義塾大学理工学部卒。専門は情報工学。WEBにコンテンツを作り散らかすことで生計を立てている。現在の主な収入源は「アカデミックに人の悪口を書くnote有料マガジン」。 Twitter→https://twitter.com/kenhori2 noteマガジン→https://note.com/kenhori2/m/m125fc4524aca 個人YouTube→https://www.youtube.com/channel/UCYJ335HO_qLZDr7TywpI0Gg 【水野太貴プロフィール】 名古屋大学文学部卒。専門は言語学。 某大手出版社で編集者として勤務。言語学の知識が本業に活きてるかと思いきや、そうでもない。 【BGM提供】 ・フリーBGM・音楽素材MusMus様 https://musmus.main.jp ・OtoLogic様 https://otologic.jp/ #公開収録 #ゆる言語学ラジオ_公開収録

On this episode of Talkin' Tech, Lee Harris talks about a new ASUS Rog Flow, a hybrid laptop/Tablet with all the CPU/GPU power of a gaming PC.

最近一段时间有一个与科技相关的重大热点话题，就是中国的中兴通讯公司被美国禁售的事件，好多听众都来问我对这事怎么看？我国为什么都已经厉害了，但是芯片却还是罩门呢？我想，既然是做节目，还是要简单介绍一下中兴通讯被美国禁售是怎么回事。2018 年 4 月 16 日，美国的商务部下达了对中兴长达七年的禁售令，这个禁售包含两部分意思。一是禁止美国公司向中兴公司出售任何有技术含量的软硬件，准确地说这个出售是广义的，免费赠送也算是出售，用商务部的原话就是不得订购、买卖、使用、送交、存储、交易、转让、运输，包括开源的安卓系统中兴也不能用，因为安卓是谷歌的，说的通俗点就是只要是美国货，中兴公司连碰都不能碰。另一层含义是中兴公司的产品不能在美国销售。可能大家没想到，2017 年的最新数据显示，中兴的手机是墙内开花墙外香，中兴手机在美国的市场份额排名第四，占 11.6%，排在苹果、三星、LG 之后，排名第一的苹果大约是中兴的三倍，中兴手机在美国的出货量大约是 460 万台。禁令是即时生效的，所以也就意味着，从禁令下达的当天开始，中兴原计划销往美国的手机瞬间就变成几乎不值一文的库存了。这个打击对中兴来说是休克性的，中兴的股票也马上就停牌了。在中兴工作的据说八万名员工瞬间不知道该干什么了，当然这是比较夸张的说法，中兴的产品线也并不是完全依赖美国货。我相信在这种公司的危难时刻，一定还是会有大量的中兴员工同心协力、共渡难关的。那这次美国人为什么下手这么狠呢，原因到底是什么？根据美国商务部的文件，是这么说的，中兴公司违反了 2016 年与美国商务部的一份调解协议，当时中兴公司因为违反了美国商务部关于美国产品的出口限制条例，违规将含有美国技术的产品出售给伊朗等国家，所以美国商务部对中兴实施了 8 亿多美元的处罚，其中有 3 亿美元缓期执行，但前提是中兴公司要承诺做到很多事情，如果不能做到，那就会遭到长达 7 年的禁售令。结果中兴公司实际上没有按照协议规定开除 35 名高管，被美国抓住了证据，于是美国商务部就按照当初的协议下达了禁售令。大致来说就是这样一个过程。我相信详细的具体过程肯定是非常复杂的，中兴公司也一定会有自己的理由。那这中间的是非功过我没有资格去评说，也不是我这个节目的宗旨，我只是把事件的背景给听众们交代一下。这件事情，我想，会给很多并不理性的爱国者上了一课，那些整天用一种很极端的方式号召抵制美货的人在我看来就并不是十分理性的，这次事件，就告诉他们，有些产品，我们买不到才是更大的悲剧。我们总不能一边高喊抵制美货，一边又骂美国人不卖东西给我们吧，这种情况是很矛盾的。同时，这件事情也让很多普通中国人吃了一惊，原来在市场经济中，也并不都是顾客就是上帝啊。可能有很多人都觉得现在这个时代甲方都是腰杆子最硬的，谁有钱谁有底气，只有卖货的求着买货的，哪有买家求着卖家的道理呢。在我们作为个人的日常消费中，确实是这样的。但是在世界贸易的大格局中，尤其是涉及到高科技产品的贸易中，其实并不都是这样。大家可能不知道，很多中国公司，包括中兴、华为这样的大公司，拿着钱去买美国、德国、日本等国家的高技术含量的产品时，那都是要低三下四，有时候还要通过一些有活动能力的中间人才能买到。即便买到了，也必须签署同意非常多的附加条件。但也有很多情况下，我国的公司费了很大的周折，填写了大量调查表格，受尽了屈辱后，人家回一封邮件，写道：很抱歉，我们不能将该产品卖给贵公司。有时候，中国的公司花很高的价钱去收购国外一家负资产的公司，可能就是为了这家公司拥有的某一台先进的设备或者某个有研发能力的团队。我们经常说，钱买不来知识、人品、气质等等非物质层面的东西。实际上，有时候，连物质层面的科技产品也是没法用钱买来的。其中，最最重要的高科技产品就是芯片——一块小小的集成电路。它的原料就是一把沙子，提纯成单晶硅后，在上面刻上成千上亿个半导体单元，就是我国现在面临的一道高高的技术门槛，如果我们不能自己迈过去，就有可能遭到别国一剑封喉式的打击。芯片到底是怎么做出来的呢？一块芯片的诞生，要经历设计、制造、封装、测试四个阶段，缺一不可，越是前面的步骤技术含量就越高，越往后就相对容易一些。中国和美国目前最大的差距就在芯片的设计上，这种差距是全方位的差距，不但有技术本身，还有人家先发的专利保护优势。芯片设计本身又要分成很多个层次，最基础的层次就是指令集架构，相当于建楼房的基本框架结构，这种框架结构目前百分之一百掌握在美国人的手上。在框架结构上继续细化的设计中国和美国、日本、德国的差距也很大。如果从销售额上来说，中国的紫光公司是中国最大的芯片设计企业，可是它目前也仅仅具备设计内存芯片的能力。从先进性上来说，华为的海思麒麟芯片是中国自主研发的最高成就了，不过，它依然是建立在美国人搞出来的 ARM 框架下，当然用别人的框架还有很多其他复杂的因素，并不仅仅是技术因素，但技术因素也是最重要的因素之一。还有，麒麟芯片中非常至关重要的数字信号处理器 DSP 买的是美国德州仪器的技术，华为还研发不出替代品。所以，如果美国对华为也禁售，那么理论上，华为的麒麟芯片也无法生产了。除了手机芯片我们勉强还能算有自主研发能力外，其他广泛使用的民用领域芯片，例如 CPU、GPU等等，我们几乎都还是空白。芯片的设计图出来了，就要进入到制造环节，如果制造技术跟不上，设计的再好也不行。衡量制造水平的高低有一个最重要的参数，就是芯片半导体单元的密度，这个一般用多少纳米来衡量。这个领域最强的是美国、中国台湾和韩国。中国大陆目前具备的最先进的生产线是 28nm 的，但是国际最先进的已经小于 10nm 了，我们还与国际先进水平有至少两代的差距，我这还是用的中国的最好的水平跟国际上的平均水平比，如果跟他们最先进的去比，那差距就更大。总之，我们在制造领域的差距，业内人士都知道，也是非常巨大的。如果台湾能回归祖国的话，这一块的差距倒是可以瞬间抹平，只是台湾问题就更复杂了。相比设计和制造，我们在封装和测试这两个环节上的差距倒是并不大，甚至在工厂规模上还远超美国，但工厂规模还是不能代表技术，用到的设备和技术依然还严重依赖进口。总之，如果大家看到有些乐观主义的文章说，设计制造芯片也没什么了不起的，技术没那么神秘，只要愿意投入研发资金，搞出来并不难，这么说是不客观的，如果真这么容易，中国早就搞出来了。用大炼钢铁的精神去搞芯片也就是历史重演罢了。有些人在谈论中国的成就和世界的地位时，往往眼里只有 GDP，尤其是在谈到日本时，总是拿 GDP 来嘲笑他们发展停滞了，可是，如果从科技的发展来看，日本不但没有停滞，还把中国甩在了后面。我觉得，中国人在对待发展这个问题上，我们也需要一次认知升级，我们不能再只盯着 GDP 来与世界各国比较了，我们还要跟他们比全民的科学素养、科技企业的软实力、科研体制的活力和全民对科学活动的重视程度。在这次中兴事件后，我也看了大量的评论文章。有两种倾向我都是反对的，一个是速胜论，以为我们只要拿出铁人的牺牲精神，不怕苦不怕累，咬紧牙关就能在三五年内打破国外的技术垄断。还有一个是亡国论，从一个芯片技术的落后就全面否定我国这三十年来在科技发展上取得的所有重大成果。我国的军事技术、航空航天技术、巨型计算机、量子通信、超级工程建设技术等等，那也都是实打实的取得了巨大的进步，该有的自信我们绝不能丢掉，要相信中华民族的伟大复兴迟早会到来。我做了那么多期节目，今天是第一次在节目中喊了一次口号，那是看了很多亡国论的文章实在忍不住了。但是，我其实心理很清楚，喊口号其实是最没有实际作用的做法。芯片很难做，无论是设计、制造还是封装，都有非常高的技术含量，也没办法靠几个归国的高级人才就能解决。它是一个国家综合科技实力的体现，是一个国家在厚厚的土壤上才能成长出来的花朵，要想造出好芯片，要从娃娃抓起。全民科学素养就是这层厚厚的土壤。我们与美国的真正差距就是这层土壤的差距。在面对中兴禁售事件中，有些人高喊着口号，有些人痛骂着国民劣根性。也有一些人，默默地咬紧牙关，用自己力所能及的点滴工作为我国的科技发展做一些贡献。而我，通过中兴事件，越发感到自己正在进行的这项科普事业的重大意义，我虽然不能直接为中国今天的科研起到促进作用，但是我在做的是默默地耕耘土壤的工作，或许，明天中国取得的科技成就会有我的点滴贡献，我今天就像是一个农民，为了这片土壤的肥沃，辛勤地翻土、浇水、施肥，这些工作说实话并没有太多的技术含量，只是一些体力活，但这种工作也必须要有人做，一片茂盛庄稼的诞生，每一个环节都需要做好。而你作为我的听众，或许正在竖立世界观的过程中，或许正在选择未来攻读的方向，也或许是年轻的父母，我想，我们都少喊一点口号，少骂几声娘，咬紧牙关，志存高远，或许我们今天的努力，可以让中华民族的伟大复兴早一秒钟到来。