Weekly Digest
A weekly roll-up of the most important items across my daily digests.
Weekly roll-up (for bookmarking). Each week links to the daily issues and surfaces a few highlights.
2026-W26 (2026-06-22 → 2026-06-28)
Issues: 2026-06-22 · 2026-06-26
Highlights — News
- When the Trump administration cracks down on Anthropic, who benefits? (TechCrunch)
- Trump admin’s coal investments assist plants with repeated violations (Ars Technica)
- Bose thinks it can be a media company for some reason (The Verge)
- OpenAI tripled revenue to $5.7 billion in Q1 but burned through $3.7 billion to get there (The Decoder (AI))
- Samsung Electronics brings ChatGPT and Codex to employees (OpenAI Blog)
- The White House is asking OpenAI to slow roll the release of its new model over safety concerns (TechCrunch)
- YouTube updates Shorts to make it even more like TikTok (The Verge)
- Microsoft adds another year to Windows 10 extended update program (Ars Technica)
- Why Amazon Dropped Its OpenAI Movie, Data Center Workers Fight Back, and Meta Leaks Employee Data (WIRED - AI)
- What happened after 2k people tried to hack my AI assistant (Hacker News (front page))
Highlights — AI tools
- Comfy-Org/ComfyUI: comfyanonymous/ComfyUI: v0.25.1 (GitHub Releases - comfyanonymous/ComfyUI)
- Data2Story turns a CSV file into a verified interactive news article using seven AI agents (The Decoder (AI))
- Temporary Cloudflare Accounts for AI agents (Simon Willison (LLM/tools))
- BerriAI/litellm: v1.89.3 (GitHub Releases - BerriAI/litellm)
- langchain-ai/langchain: langchain==1.3.10 (GitHub Releases - langchain-ai/langchain)
- OpenAI GPT, OpenAI GPT OSS, and NVIDIA Nemotron models on Amazon Bedrock receive FedRAMP High and DoD IL-4/5 approval in AWS GovCloud (US) (AWS What's New (Recent Announcements))
- [AINews] OpenAI reports median internal Codex output tokens grew 56x in Research, 32x in Customer Support, 27x in Engineering, and 13x in Legal since November 2025. (Latent Space (LLM/tools))
- langchain-ai/langchain: langchain-fireworks==1.4.3 (GitHub Releases - langchain-ai/langchain)
- Streamlining Resource Binding with End-to-End Support for Vulkan Descriptor Heaps (NVIDIA Developer Blog)
2026-W25 (2026-06-15 → 2026-06-21)
Issues: 2026-06-15 · 2026-06-16 · 2026-06-17 · 2026-06-18 · 2026-06-21
Highlights — News
- Did a medieval flying monk spot Halley's comet, twice? It's complicated (Ars Technica)
- The AI layoff wave is becoming a powder keg (TechCrunch)
- Under-16 social media ban announced by UK government (The Verge)
- Introducing the OpenAI Partner Network (OpenAI Blog)
- Amazon and five other companies reportedly triggered the government crackdown on Anthropic's Fable model (The Decoder (AI))
- Chipmaker Nvidia seeks to raise over $25B in first bond deal since 2021 (Ars Technica)
- Why do South Koreans love AI so much? (MIT Technology Review)
- The US government’s Anthropic models ban was never about an AI jailbreak (TechCrunch)
- All the news about Anthropic’s new AI fight with the White House (The Verge)
- The US government may be asking Anthropic the impossible by demanding unhackable LLMs (The Decoder (AI))
Highlights — AI tools
- vllm-project/vllm: v0.23.0 (GitHub Releases - vllm-project/vllm)
- Google Cloud's Open Knowledge Format turns scattered docs into Markdown files for AI agents (The Decoder (AI))
- Amazon Bedrock AgentCore Memory now supports strictly consistent metadata for long-term memory (AWS What's New (Recent Announcements))
- Fine-Tuning Biological Foundation Models with LoRA Using NVIDIA BioNeMo Recipes (NVIDIA Developer Blog)
- "They screwed us": Personality clashes sent Anthropic's models offline (Simon Willison (LLM/tools))
- openai/openai-python: v2.42.0 (GitHub Releases - openai/openai-python)
- openai/openai-node: v6.43.0 (GitHub Releases - openai/openai-node)
- Comfy-Org/ComfyUI: comfyanonymous/ComfyUI: v0.25.0 (GitHub Releases - comfyanonymous/ComfyUI)
- Building AI Agents for AR Glasses and XR Devices with NVIDIA XR AI (NVIDIA Developer Blog)
- Amazon Bedrock Guardrails announces a new API targeting agentic AI workflows (AWS What's New (Recent Announcements))
2026-W24 (2026-06-08 → 2026-06-14)
Issues: 2026-06-08
Highlights — News
- Hacked, leaked, and held for ransom: the worst breaches of 2026 so far (TechCrunch)
- RIP Anthony Head: Our 10 favorite moments of Buffy's Giles (Ars Technica)
- Anthropic poaches OpenAI's second-ever chip engineer as both companies race toward IPOs (The Decoder (AI))
- NASA will wear high-tech Prada long johns to the Moon (The Verge)
- OpenAI and Anthropic May Be Rivals, but Investors Aren’t Picking Sides (WIRED - AI)
Highlights — AI tools
- OpenAI says "chat is dead" and plans to rebuild ChatGPT as a full-blown agent app (The Decoder (AI))
- datasette-agent-edit 0.1a0 (Simon Willison (LLM/tools))
- BerriAI/litellm: v1.88.0 (GitHub Releases - BerriAI/litellm)
- Amazon Bedrock AgentCore Runtime introduces interactive shells for terminal access into agent sessions (AWS What's New (Recent Announcements))
- ollama/ollama: v0.30.6 (GitHub Releases - ollama/ollama)
2026-W23 (2026-06-01 → 2026-06-07)
Issues: 2026-06-01
Highlights — News
- Meta is reportedly developing an AI pendant (TechCrunch)
- How to watch Nvidia’s Computex keynote (The Verge)
- On its 40th anniversary, we reassess 1986's SpaceCamp (Ars Technica)
- How Turkey Hacked the Hair Transplant Industry (WIRED - AI)
- Anthropic bans AI tools during job interviews to see how candidates actually think (The Decoder (AI))
Highlights — AI tools
- Microsoft and Nvidia reportedly team up on AI PCs that run actual agents instead of Copilot (The Decoder (AI))
- BerriAI/litellm: v1.84.4 (GitHub Releases - BerriAI/litellm)
- How we contain Claude across products (Simon Willison (LLM/tools))
2026-W22 (2026-05-25 → 2026-05-31)
Issues: 2026-05-30 · 2026-05-31
Highlights — News
- After Nvidia’s $20B not-acqui-hire, AI chip startup Groq reportedly raising $650M (TechCrunch)
- Botnet of more than 17 million devices dismantled (Ars Technica)
- This is the James Bond game we’ve been waiting for (The Verge)
- OpenAI's Codex can now operate your Windows PC autonomously, hunting bugs and testing apps on its own (The Decoder (AI))
- Do You Actually Need to Pay for Transcription Software? (WIRED - AI)
- Meta is reportedly developing an AI pendant (TechCrunch)
- Grifters, cynics, and true believers: The family tree of vaccine opponents (Ars Technica)
- Backrooms is a certified blockbuster with a $38 million opening day (The Verge)
- Poetry for Engineers: Cyborg Laboratory (IEEE Spectrum)
Highlights — AI tools
- OpenAI is giving away its life sciences AI model to help governments prepare for the next pandemic (The Decoder (AI))
- [AINews] Founders and Forward Deployed Engineers (Latent Space (LLM/tools))
- DynoSim: Simulating the Pareto Frontier (NVIDIA Developer Blog)
- Hands-On With Gemini Spark: I Gave It Access to My Life and It Friend-Zoned My Boyfriend (WIRED - AI)
- Microsoft and Nvidia reportedly team up on AI PCs that run actual agents instead of Copilot (The Decoder (AI))
- BerriAI/litellm: v1.84.4 (GitHub Releases - BerriAI/litellm)
- How we contain Claude across products (Simon Willison (LLM/tools))
2026-W21 (2026-05-18 → 2026-05-24)
Issues: 2026-05-24
Highlights — News
- Four Russian satellites are now within striking distance of an ICEYE radarsat (Ars Technica)
- Hackers are learning to exploit chatbot ‘personalities’ (The Verge)
- I tried Amazon’s Bee wearable and am both intrigued and slightly creeped out (TechCrunch)
- Anthropic may keep supplying Claude to the NSA despite being flagged as a supply chain risk by the Pentagon (The Decoder (AI))
- These Robots Are Making Meals for a Nonprofit in San Francisco’s Tenderloin (WIRED - AI)
Highlights — AI tools
- Alibaba's latest AI model ran autonomously for 35 hours to optimize code for its own custom chip (The Decoder (AI))
- BerriAI/litellm: v1.86.0 (GitHub Releases - BerriAI/litellm)
- [AINews] All Model Labs are now Agent Labs (Latent Space (LLM/tools))
2026-W20 (2026-05-11 → 2026-05-17)
Issues: 2026-05-12 · 2026-05-13
Highlights — News
- Instructure strikes deal with hackers who breached it twice (TechCrunch)
- Linux bitten by second severe vulnerability in as many weeks (Ars Technica)
- Canvas owner reaches ‘agreement’ with hackers to secure stolen data (The Verge)
- The Download: a Nobel winner on AI, and the case for fixing everything (MIT Technology Review)
- Sam Altman's personal investments face political scrutiny ahead of OpenAI's planned IPO (The Decoder (AI))
- “Will I be OK?” Teen died after ChatGPT pushed deadly mix of drugs, lawsuit says (Ars Technica)
- Musk mulled handing OpenAI to his children, Altman testifies (TechCrunch)
- Sam Altman says Elon Musk’s mind games were damaging OpenAI (The Verge)
- Elon Musk Had ‘Hair-Raising’ Idea of Passing OpenAI On to His Kids, Sam Altman Says (WIRED - AI)
- Google says it stopped a mass cyberattack after AI was used to discover a zero-day exploit (The Decoder (AI))
Highlights — AI tools
- Thinking Machines Lab ships its first model and argues interactivity is what OpenAI gets wrong about voice (The Decoder (AI))
- Introducing NVIDIA Fleet Intelligence for Real-Time GPU Fleet Visibility and Optimization (NVIDIA Developer Blog)
- langchain-ai/langchain: langchain==1.3.0 (GitHub Releases - langchain-ai/langchain)
- [AINews] Thinking Machines' Native Interaction Models - TML-Interaction-Small 276B-A12B - advances SOTA Realtime Voice and kills standard VAD (Latent Space (LLM/tools))
- Anthropic expands legal AI offerings with new Claude Cowork plugins (The Decoder (AI))
- How to Eliminate Pipeline Friction in AI Model Serving (NVIDIA Developer Blog)
- Amazon SageMaker Feature Store now supports SageMaker Python SDK V3 (AWS What's New (Recent Announcements))
2026-W19 (2026-05-04 → 2026-05-10)
Issues: 2026-05-07 · 2026-05-08 · 2026-05-09
Highlights — News
- DOJ says ransomware gang tapped into Russian government databases (TechCrunch)
- TSMC taps wind power as AI chip demand soars, Taiwan feels energy crunch (Ars Technica)
- Musk’s biggest loyalist became his biggest liability (The Verge)
- OpenAI built a networking protocol with AMD, Broadcom, Intel, Microsoft, and NVIDIA to fix AI supercomputer bottlenecks (The Decoder (AI))
- Elon Musk’s Last-Ditch Effort to Control OpenAI: Recruit Sam Altman to Tesla (WIRED - AI)
- The Download: AI malaise and babymaking tech (MIT Technology Review)
- Everyone’s a loser in Strait of Hormuz game that simulates global crisis (Ars Technica)
- Last 24 hours to get 50% off a second pass to TechCrunch Disrupt 2026 (TechCrunch)
- Boox’s new page-turning e-reader remote is a tiny two-button keyboard (The Verge)
- Canvas Breach Disrupts Schools & Colleges Nationwide (Krebs on Security)
Highlights — AI tools
- Amazon Bedrock AgentCore Memory announces metadata for long-term memory (AWS What's New (Recent Announcements))
- openai/openai-python: v2.35.1 (GitHub Releases - openai/openai-python)
- langchain-ai/langchain: langchain==1.3.0a2 (GitHub Releases - langchain-ai/langchain)
- langchain-ai/langchain: langchain==1.2.18 (GitHub Releases - langchain-ai/langchain)
- AI safety tests have a new problem: Models are now faking their own reasoning traces (The Decoder (AI))
- [AINews] GPT-Realtime-2, -Translate, and -Whisper: new SOTA realtime voice APIs (Latent Space (LLM/tools))
- BerriAI/litellm: v1.83.14-stable.patch.3 (GitHub Releases - BerriAI/litellm)
- [AINews] Anthropic growing 10x/year while everyone else is laying off >10% of their workforce (Latent Space (LLM/tools))
- Streaming Tokens and Tools: Multi-Turn Agentic Harness Support in NVIDIA Dynamo (NVIDIA Developer Blog)
- Using Claude Code: The Unreasonable Effectiveness of HTML (Simon Willison (LLM/tools))
2026-W18 (2026-04-27 → 2026-05-03)
Issues: 2026-04-27 · 2026-04-28 · 2026-04-29
Highlights — News
- To buy this Bay Area home, you’ll need Anthropic equity (TechCrunch)
- Strange New Worlds S4 teaser strikes a more serious tone (Ars Technica)
- Google’s new gradient icon design is coming to more apps (The Verge)
- OpenAI kills its dedicated coding model Codex again, folding it into GPT-5.5 (The Decoder (AI))
- Discord Sleuths Gained Unauthorized Access to Anthropic’s Mythos (WIRED - AI)
- Elon Musk and Sam Altman are going to court over OpenAI’s future (MIT Technology Review)
- Musk and Altman face off in trial that will determine OpenAI's future (Ars Technica)
- OpenAI ends Microsoft legal peril over its $50B Amazon deal (TechCrunch)
- Google is testing AI chatbot search for YouTube (The Verge)
- Some Musk v. Altman Jurors Don't Like Elon Musk (WIRED - AI)
Highlights — AI tools
- langchain-ai/langchain: langchain-openai==1.2.1 (GitHub Releases - langchain-ai/langchain)
- Amazon Bedrock AgentCore Gateway and Identity support VPC egress (AWS What's New (Recent Announcements))
- open-webui/open-webui: v0.9.2 (GitHub Releases - open-webui/open-webui)
- Comfy-Org/ComfyUI: comfyanonymous/ComfyUI: v0.20.1 (GitHub Releases - comfyanonymous/ComfyUI)
- Introducing talkie: a 13B vintage language model from 1930 (Simon Willison (LLM/tools))
- vllm-project/vllm: v0.20.0 (GitHub Releases - vllm-project/vllm)
- BerriAI/litellm: v1.83.10-stable (GitHub Releases - BerriAI/litellm)
- Amazon SageMaker HyperPod now supports G7e and r5d.16xlarge instances (AWS What's New (Recent Announcements))
- langchain-ai/langchain: langchain-anthropic==1.4.2 (GitHub Releases - langchain-ai/langchain)
- Scaling Biomolecular Modeling Using Context Parallelism in NVIDIA BioNeMo (NVIDIA Developer Blog)
2026-W17 (2026-04-20 → 2026-04-26)
Issues: 2026-04-24 · 2026-04-25 · 2026-04-26
Highlights — News
- In a first, a ransomware family is confirmed to be quantum-safe (Ars Technica)
- Vercel says some of its customers’ data was stolen prior to its recent hack (TechCrunch)
- Meta is laying off 10 percent of its staff (The Verge)
- French government agency confirms breach as hacker offers to sell data (Hacker News (front page))
- What Anthropic’s Mythos Means for the Future of Cybersecurity (IEEE Spectrum)
- In another wild turn for AI chips, Meta signs deal for millions of Amazon AI CPUs (TechCrunch)
- Google will invest as much as $40 billion in Anthropic (Ars Technica)
- Three reasons why DeepSeek’s new model matters (MIT Technology Review)
- Microsoft will let you pause Windows Updates indefinitely, 35 days at a time (The Verge)
- OpenAI's chief scientist says AI progress has been "surprisingly slow" and promises big leaps ahead (The Decoder (AI))
Highlights — AI tools
- OpenAI unveils GPT-5.5, claims a "new class of intelligence" at double the API price (The Decoder (AI))
- Amazon SageMaker supports notebooks and data agent for IdC domains (AWS What's New (Recent Announcements))
- An update on recent Claude Code quality reports (Simon Willison (LLM/tools))
- BerriAI/litellm: v1.83.7-stable.patch.1 (GitHub Releases - BerriAI/litellm)
- langchain-ai/langchain: langchain-fireworks==1.2.0 (GitHub Releases - langchain-ai/langchain)
- langchain-ai/langchain: langchain-openai==1.2.1 (GitHub Releases - langchain-ai/langchain)
- Amazon Bedrock AgentCore Gateway and Identity support VPC egress (AWS What's New (Recent Announcements))
- open-webui/open-webui: v0.9.2 (GitHub Releases - open-webui/open-webui)
2026-W05 (2026-01-26 → 2026-02-01)
Issues: 2026-01-26 · 2026-01-27
Highlights — News
- Inside OpenAI’s big play for science (MIT Technology Review)
- Data center power outage took out TikTok first weekend under US ownership (Ars Technica)
- TikTok attributes recent glitches to a power outage at a US data center (TechCrunch)
- MCP unites Claude chat with apps like Slack, Figma, and Canva (The Verge)
- MCP Apps, the Model Context Protocol's first official extension, turns AI responses into interactive interfaces (The Decoder (AI))
- Measles is surging in the US. Wastewater tracking could help. (MIT Technology Review)
- Poland's energy grid was targeted by never-before-seen wiper malware (Ars Technica)
- ChatGPT is pulling answers from Elon Musk’s Grokipedia (TechCrunch)
- Microsoft handed the government encryption keys for customer data (The Verge)
- OpenAI's GPT-5.2 Pro solves math problems that stumped every AI model before it (The Decoder (AI))
Highlights — AI tools
- ChatGPT fails to spot 92% of fake videos made by OpenAI's own Sora tool (The Decoder (AI))
- Scaling NVFP4 Inference for FLUX.2 on NVIDIA Blackwell Data Center GPUs (NVIDIA Developer Blog)
- modelcontextprotocol/servers: Release 2026.1.14 (GitHub Releases - modelcontextprotocol/servers)
- Comfy-Org/ComfyUI: comfyanonymous/ComfyUI: v0.10.0 (GitHub Releases - comfyanonymous/ComfyUI)
- Amazon Quick Suite launches expanded size, faster ingestion, and richer data type support for SPICE datasets (AWS What's New (Recent Announcements))
- First impressions of Claude Cowork, Anthropic's general agent (Simon Willison (LLM/tools))
2026-W04 (2026-01-19 → 2026-01-25)
Issues: 2026-01-20 · 2026-01-21 · 2026-01-22 · 2026-01-23 · 2026-01-24 · 2026-01-25
Highlights — News
- OpenAI partners with Cerebras (OpenAI Blog)
- Differential Transformer V2 (Hugging Face Blog)
- The Conversational Exam: A Scalable Assessment Design for the AI Era (arXiv)
- Open Responses: What you need to know (Hugging Face Blog)
- Evaluating 21st-Century Competencies in Postsecondary Curricula with Large Language Models: Performance Benchmarking and Reasoning-Based Prompting Strategies (arXiv)
- US regulator appeals Meta’s antitrust win (Financial Times - Technology)
- Global markets on alert as Europe to suspend approval of US trade deal (BBC News - Business)
- Anthropic’s CEO stuns Davos with Nvidia criticism (TechCrunch)
- Netflix earned $1.5 billion from ads in 2025 (The Verge)
- Dollar and US stocks fall as Trump says ‘no going back’ on Greenland bid (Financial Times - World)