Simon Willison explored running Python ASGI applications directly in the browser using Pyodide and a service worker. This research aims to overcome limitations of previous in-browser Python implementations, specifically enabling JavaScript execution within generated HTML. The project, Datasette Lite, leverages WebAssembly and a service worker to intercept requests and run the Python app, improving compatibility with Datasette plugins.
→ This is a high-signal development for on-device AI, pushing local Python execution further into the browser, which could enable more sophisticated client-side AI tools and interactive data applications without server-side dependencies.
NVIDIA has released Cosmos 3, an open omni-model designed for physical AI reasoning and action. This model integrates various modalities to understand and interact with the physical world, aiming to bridge the gap between AI and robotics. It represents a significant step towards more capable and autonomous physical AI systems.
→ Cosmos 3 is a high-signal release for physical AI, potentially impacting on-device robotics and offering new avenues for local model integration beyond just text.
Hugging Face published the first part of a beginner's guide to `torch.profiler` for PyTorch users. This guide introduces the tool's basic functionalities, including how to set it up, collect traces, and interpret the output for performance analysis. It aims to help developers identify bottlenecks and optimize their PyTorch models more effectively.
→ Essential for anyone optimizing local LLMs like Gemma or Llama, as `torch.profiler` offers deep insights into performance bottlenecks, crucial for on-device efficiency and contest submissions.
Simon Willison explored running Python ASGI applications directly in the browser using Pyodide and a service worker. This research aims to overcome limitations of previous in-browser Python implementations, specifically enabling JavaScript execution within generated HTML. The project, Datasette Lite, leverages WebAssembly and a service worker to intercept requests and run the Python app, improving compatibility with Datasette plugins.
→ This is a high-signal development for on-device AI, pushing local Python execution further into the browser, which could enable more sophisticated client-side AI tools and interactive data applications without server-side dependencies.
Hugging Face published the first part of a beginner's guide to `torch.profiler` for PyTorch users. This guide introduces the tool's basic functionalities, including how to set it up, collect traces, and interpret the output for performance analysis. It aims to help developers identify bottlenecks and optimize their PyTorch models more effectively.
→ Essential for anyone optimizing local LLMs like Gemma or Llama, as `torch.profiler` offers deep insights into performance bottlenecks, crucial for on-device efficiency and contest submissions.
NVIDIA has released Cosmos 3, an open omni-model designed for physical AI reasoning and action. This model integrates various modalities to understand and interact with the physical world, aiming to bridge the gap between AI and robotics. It represents a significant step towards more capable and autonomous physical AI systems.
→ Cosmos 3 is a high-signal release for physical AI, potentially impacting on-device robotics and offering new avenues for local model integration beyond just text.
Mistral AI has released Mistral Medium 3.5, their latest flagship model, which demonstrates improved performance across various benchmarks. This new model is positioned as a powerful option for complex reasoning tasks and multilingual applications. It is now available through their API.
→ Another strong contender in the frontier model space, but the real interest for us is how this might influence future local LLM developments and open-source releases.
Anthropic announced a Series H funding round of $65 billion at a $900 billion pre-money valuation, with $15 billion from hyperscalers. The company also reported a $47 billion revenue run-rate, significantly up from $9 billion in December. Additionally, Anthropic released Opus 4.8 and introduced Dynamic Workflows/ultracode.
→ While the funding is massive, the release of Opus 4.8 and Dynamic Workflows/ultracode could signal new tooling and model capabilities relevant to the Claude Code ecosystem.
Simon Willison explored running Python ASGI applications directly in the browser using Pyodide and a service worker. This research aims to overcome limitations of previous in-browser Python implementations, specifically enabling JavaScript execution within generated HTML. The project, Datasette Lite, leverages WebAssembly and a service worker to intercept requests and run the Python app, improving compatibility with Datasette plugins.
→ This is a high-signal development for on-device AI, pushing local Python execution further into the browser, which could enable more sophisticated client-side AI tools and interactive data applications without server-side dependencies.
NVIDIA has released Cosmos 3, an open omni-model designed for physical AI reasoning and action. This model integrates various modalities to understand and interact with the physical world, aiming to bridge the gap between AI and robotics. It represents a significant step towards more capable and autonomous physical AI systems.
→ Cosmos 3 is a high-signal release for physical AI, potentially impacting on-device robotics and offering new avenues for local model integration beyond just text.
Hugging Face published the first part of a beginner's guide to `torch.profiler` for PyTorch users. This guide introduces the tool's basic functionalities, including how to set it up, collect traces, and interpret the output for performance analysis. It aims to help developers identify bottlenecks and optimize their PyTorch models more effectively.
→ Essential for anyone optimizing local LLMs like Gemma or Llama, as `torch.profiler` offers deep insights into performance bottlenecks, crucial for on-device efficiency and contest submissions.
Mistral AI has released Mistral Medium 3.5, their latest flagship model, which demonstrates improved performance across various benchmarks. This new model is positioned as a powerful option for complex reasoning tasks and multilingual applications. It is now available through their API.
→ Another strong contender in the frontier model space, but the real interest for us is how this might influence future local LLM developments and open-source releases.
Anthropic announced a Series H funding round of $65 billion at a $900 billion pre-money valuation, with $15 billion from hyperscalers. The company also reported a $47 billion revenue run-rate, significantly up from $9 billion in December. Additionally, Anthropic released Opus 4.8 and introduced Dynamic Workflows/ultracode.
→ While the funding is massive, the release of Opus 4.8 and Dynamic Workflows/ultracode could signal new tooling and model capabilities relevant to the Claude Code ecosystem.
Research: pydantic-monty investigation It's been a few months since I last poked at Monty , the sandboxed subset of Python implemented in Rust. I had Claude Code look at the most recent release. Importantly the max_duration_secs , max_memory , max_allocations , and max_recursion_
Boston Children’s Hospital uses OpenAI technology to improve patient care, reduce operational burden, and help diagnose more than 40 rare disease cases.
Most people are still digesting the massive Anthropic news from yesterday. We’re taking the opportunity to solicit the leading AI FDE’s in the world for AIE’s new Forward Deployed Engineer track, mirroring similar pushes from both OpenAI DeployCo and Anthropic DeployCo : as well
The new AIEWF website is live! CFPs close in 2 days and we will run our first New Engineer Orientation this weekend, get your tickets booked ASAP as they -will- sell out. Take the AI Engineering Survey and get >$2k in credits and free AIE WF tickets ! One of the central tensions
Anthropic releases Claude Opus 4.8 with improved honesty, stronger self‑verification, and multi‑agent dynamic workflows for large code tasks. Benchmark scores narrow versus OpenAI's GPT‑5.5 while debate grows over harness quality and real‑world tradeoffs. Headlines also cover Kir
How we contain Claude across products A complaint I often have about sandboxing products is that they are rarely thoroughly documented , and in the absence of detailed documentation it's hard to know how much I can trust them. Anthropic just published a fantastic overview of how
Release: datasette 1.0a31 Another significant alpha release, with two new headline features. Datasette now offers users with the necessary permissions the ability to both execute write queries against their database and to save stored queries (renamed from "canned queries") both
Anthropic shipped Claude Opus 4.8 today. My favourite thing about it is this note in the release announcement: Users will find Opus 4.8 to be a modest but tangible improvement on its predecessor. There’s still more to be done: we’re working on developing and releasing models that
Release: llm-anthropic 0.25.1 New model: Claude Opus 4.8 ( claude-opus-4.8 ). New -o fast 1 option for fast mode , for organizations with that feature enabled on their account. Default max_tokens for each model now defaults to that model's maximum output rather than 8,192. #72 Se
Watch 9 videos showing the capabilities of Gemini Omni and Gemini 3.5, announced at Google I/O 2026.
University of Waterloo students develop AI prototypes like sign language tutors to reshape the future of education and work.
OpenAI shares guidance on third-party AI evaluations, covering how to assess model capabilities, safeguards, and validity for frontier systems.
MUFG uses ChatGPT Enterprise to build an AI-native organization, improve workflows, and deliver new AI-powered financial services at scale.
Here's the AI News from this past week. Launch your own AI agents with Hermes at http://hostinger.com/mattwolfehermes and get 10% off with code MATTWOLFE Discover More: 🛠️ Explore AI Tools & News: https://futuretools.io/ 📰 Weekly Newsletter: https://futuretools.io/newsletter Soci
DataCurve's DeepSWE benchmark exposes large performance gaps on realistic, long-horizon coding tasks. Summer AI slowdown panic returns alongside renewed debate over job displacement and deployment frictions. Token shortages and massive inference funding for Base10 and OpenRouter
I just sent out the May edition of my sponsors-only monthly newsletter . If you are a sponsor (or if you start a sponsorship now) you can access it here . This month: Al got expensive, and Anthropic had a really good month The model releases were a little disappointing Conference
The solution might be cancelling my AI subscription I find this post by David Wilson very relatable. David lists 16+ projects he's spun up with AI tooling, and concludes: I didn't mean to build most of these things. Usually the Claude session started with something like " write a
I Am Retiring from Tech to Live Offline I've seen a lot of posts on forums from people threatening to quit their careers over AI. This is not one of those: Chad Whitacre is taking concrete steps, starting with this typewritten, scanned letter I'm retiring from tech. Well, "retiri
The most interesting thing about Anthropic's $65B Series H announcement is this line (emphasis mine): Since our Series G in February, adoption has continued to grow across global enterprise customers, and our run-rate revenue crossed $47 billion earlier this month. Anthropic have
Tool: markdown-svg-renderer A slightly customized Markdown rendering tool with special treatment for fenced code SVG blocks - it both renders the image and provides a tab for switching to the code view. You can paste in Markdown or give it a URL to a CORS-enabled Markdown file or
How Braintrust engineers use Codex with GPT-5.5 to run experiments and code faster.
OpenAI launches Rosalind Biodefense, expanding trusted access to GPT-Rosalind for vetted developers and U.S. government partners advancing biodefense, public health, and pandemic preparedness through frontier AI.
Explore OpenAI’s Frontier Governance Framework and how our AI safety, security, and risk practices align with emerging EU and California regulations.
If you want to prevent AI from taking your job in the future, these are my top 3 pieces of advice: 1. Fall in love with learning. It’s one of the most common trait amongst successful people. 2. Build stuff. Companies don't just want book smarts anymore, they want people who can b
tp3_memories_localgemma3:4b)