Today in one line
Today: The Pentagon is planning for AI companies to train on classifie… · Mistral bets on ‘build-your-own AI’ as it takes on OpenAI, Anth… · AI: OpenAI ships GPT-5.4 mini and nano, faster and more capab….
Top 10 — News
- The Pentagon is planning for AI companies to train on classified data, defense official says The Pentagon is discussing plans to set up secure environments for generative AI companies to train military-specific versions of their mod…
- Mistral bets on ‘build-your-own AI’ as it takes on OpenAI, Anthropic in the enterprise Mistral Forge lets enterprises train custom AI models from scratch on their own data, challenging rivals that rely on fine-tuning and retri…
- Apple can delist apps “with or without cause,” judge says in loss for Musi app Judge tosses Musi case against Apple, sanctions lawyers for "mak[ing] up facts."
- I went to the Pentagon to watch Pete Hegseth scold war reporters It is day 13 of America's surprise war with Iran - by sheer coincidence, it's Friday the 13th - and I am delirious.
- Why Anthropic Thinks AI Should Have Its Own Computer — Felix Rieseberg of Claude Cowork & Claude Code Desktop Claude Cowork came out of an accident.
- Justice Department Says Anthropic Can’t Be Trusted With Warfighting Systems In response to Anthropic’s lawsuit, the government said it lawfully penalized the company for trying to limit how its Claude AI models coul…
- Microsoft restructures AI division to chase superintelligence after Nadella once called AI models a commodity Microsoft is restructuring its AI division to double down on its own AI models, all the way up to superintelligence.
- Get Shit Done: A meta-prompting, context engineering and spec-driven dev system Article URL: https://github.com/gsd-build/get-shit-done Comments URL: https://news.ycombinator.com/item?id=47417804 Points: 231 # Comments:…
- Launch an autonomous AI agent with sandboxed execution in 2 lines of code Article URL: https://amaiya.github.io/onprem/examples_agent.html Comments URL: https://news.ycombinator.com/item?id=47420493 Points: 17 # C…
- The Pleasures of Poor Product Design
Article URL: https://www.inconspicuous.info/p/the-pleasures-of-poor-product-design Comments URL: https://news.ycombinator.com/item?id=47420…
AI tools & model updates
- OpenAI ships GPT-5.4 mini and nano, faster and more capable but up to 4x pricier OpenAI has released two new compact models—GPT-5.4 mini and nano—built for coding assistants, subagents, and computer control.
- openai/openai-python: v2.29.0 Release update (see link for changelog).
- openai/openai-node: v6.32.0 Release update (see link for changelog).
- GPT-5.4 mini and GPT-5.4 nano, which can describe 76,000 photos for $52 OpenAI today: Introducing GPT‑5.4 mini and nano .
- langchain-ai/langchain: langchain-anthropic==1.4.0 Release update (see link for changelog).
- Inside NVIDIA Groq 3 LPX: The Low-Latency Inference Accelerator for the NVIDIA Vera Rubin Platform NVIDIA Groq 3 LPX is a new rack-scale inference accelerator for the NVIDIA Vera Rubin platform, designed for the low-latency and large-cont…
- Amazon SageMaker Unified Studio supports aggregated view of data lineage Amazon SageMaker Unified Studio now provides an aggregated view of data lineage, displaying all jobs contributing to your dataset.
- microsoft/onnxruntime: ONNX Runtime v1.24.4 Release update (see link for changelog).
- BerriAI/litellm: v1.82.0.dev.7 Release update (see link for changelog).
- ollama/ollama: v0.18.1
Release update (see link for changelog).
Full list — News (30)
The Pentagon is discussing plans to set up secure environments for generative AI companies to train military-specific versions of their mod…
Mistral Forge lets enterprises train custom AI models from scratch on their own data, challenging rivals that rely on fine-tuning and retri…
Thousands of people are trying Garry Tan's Claude Code setup, which was shared on GitHub.
After their dramatic falling-out, it doesn't seem as though Anthropic and the Pentagon are getting back together.
Judge tosses Musi case against Apple, sanctions lawyers for "mak[ing] up facts."
Iris-scan backed tokens could help stop agent swarms from overwhelming online systems.
Desert state becomes first to file criminal case against prediction platform.
Of the seven illnesses identified so far, four are in children age 3 or younger.
It's the latest salvo in an escalating battle between state regulators and an industry that claims it's not beholden to them.
10
Apple's first-ever "background security improvement" fixes a vulnerability in its Safari browser running its latest software.
Kagi's "Small Web" offers a handpicked collection of more than 30,000 non-commercial, human-authored websites, including personal blogs, we…
It is day 13 of America's surprise war with Iran - by sheer coincidence, it's Friday the 13th - and I am delirious.
Earlier this week, Apple unveiled the AirPods Max 2.
Remedy is winding down its team shooter FBC: Firebreak with a big update that launches today.
Starting Tuesday, Spotify users will have access to "Exclusive Mode," a new feature in the streaming service's Windows app that can help ma…
Ikea's new Matter-over-Thread products were supposed to prove that the smart home could be cheap, accessible, and reliable.
Claude Cowork came out of an accident.
In response to Anthropic’s lawsuit, the government said it lawfully penalized the company for trying to limit how its Claude AI models coul…
Microsoft is restructuring its AI division to double down on its own AI models, all the way up to superintelligence.
Article URL: https://github.com/gsd-build/get-shit-done Comments URL: https://news.ycombinator.com/item?id=47417804 Points: 231 # Comments:…
Article URL: https://amaiya.github.io/onprem/examples_agent.html Comments URL: https://news.ycombinator.com/item?id=47420493 Points: 17 # C…
Article URL: https://www.inconspicuous.info/p/the-pleasures-of-poor-product-design Comments URL: https://news.ycombinator.com/item?id=47420…
Article URL: https://rovarma.com/articles/a-tale-about-fixing-ebpf-spinlock-issues-in-the-linux-kernel/ Comments URL: https://news.ycombina…
Article URL: https://arxiv.org/abs/2603.15381 Comments URL: https://news.ycombinator.com/item?id=47418722 Points: 41 # Comments: 15
Article URL: https://mistral.ai/news/forge Comments URL: https://news.ycombinator.com/item?id=47418295 Points: 181 # Comments: 27
https://web.archive.org/web/20260317185928/https://terathon....
Article URL: https://hanno.codes/2026/03/17/java-26-is-here/ Comments URL: https://news.ycombinator.com/item?id=47416548 Points: 197 # Comm…
Article URL: https://fidget-spinner.github.io/posts/jit-on-track.html Comments URL: https://news.ycombinator.com/item?id=47416486 Points: 3…
Article URL: https://wasmer.io/posts/edgejs-safe-nodejs-using-wasm-sandbox Comments URL: https://news.ycombinator.com/item?id=47416081 Poin…
Full list — AI tools & model updates (30)
OpenAI has released two new compact models—GPT-5.4 mini and nano—built for coding assistants, subagents, and computer control.
At GTC 2026, Nvidia is massively expanding its platform for physical AI.
Release update (see link for changelog).
Release update (see link for changelog).
5
OpenAI now believes its strategy of shipping as many products as possible at once left the company exposed.
Release update (see link for changelog).
Release update (see link for changelog).
OpenAI today: Introducing GPT‑5.4 mini and nano .
Release update (see link for changelog).
AI-native services are exposing a new bottleneck in AI infrastructure: As millions of users, agents, and devices demand access to intellige…
At GTC 2026, Nvidia expanded the Vera Rubin platform it introduced at CES with custom CPU racks, dedicated inference chips, a new storage a…
The point of the blackmail exercise was to have something to describe to policymakers—results that are visceral enough to land with people,…
13
NVIDIA Groq 3 LPX is a new rack-scale inference accelerator for the NVIDIA Vera Rubin platform, designed for the low-latency and large-cont…
Reasoning models are growing rapidly in size and are increasingly being integrated into agentic AI workflows that interact with other model…
Amazon SageMaker Unified Studio now provides an aggregated view of data lineage, displaying all jobs contributing to your dataset.
Release update (see link for changelog).
Release update (see link for changelog).
Release update (see link for changelog).
SageMaker Training Plans allows you to reserve GPU capacity within specified time frames in cluster sizes of up to 64 instances.
AWS Blu Insights capabilities are now available as part of AWS Transform, enabling customers to launch mainframe refactoring projects from…
Amazon Bedrock AgentCore Runtime now supports InvokeAgentRuntimeCommand, a new API that lets you execute shell commands directly inside a r…
22
Agentic Engineering Patterns > LLMs are restricted by their context limit - how many tokens they can fit in their working memory at any giv…
Mistral AI has released Mistral Small 4, combining fast text responses, logical reasoning, and image processing in one model.
Starting today, customers can use Amazon Bedrock in the Asia Pacific (New Zealand) Region to easily build and scale generative AI applicati…
Release update (see link for changelog).
Release update (see link for changelog).
Use subagents and custom agents in Codex Subagents were announced in general availability today for OpenAI Codex, after several weeks of pr…
Release update (see link for changelog).
Release update (see link for changelog).
Healthcare faces a structural demand–capacity crisis: a projected global shortfall of ~10 million clinicians by 2030, billions of diagnosti…