Announcement, slides At The Economist's Impact 2nd AI Compute Summit, I talked about building smarter AI systems with smaller models. I explored in depth why AI spending keeps growing even though…

Weekly roundup of changes to AI developer tools: pricing shifts, new releases, feature updates, and user reaction. Week of May 12 – 18, 2026.

Last year I wrote a small WASM runtime in Go. How many security vulnerabilities do you think AI agents found in it?

Silicon Valley executives used to complain about “the Great Resignation” to justify replacing people with machines. Now, they have suddenly changed their tune to sound like generous…

Over the weekend the Cartoonists’ Club of Great Britain posted their position on their member’s use of generative AI. The policy was originally printed in their July 2025 edition of The…

What is the structural cost of the IndiaAI Mission? MeitY's ₹4,563 crore allocation creates a ₹25,437 crore deficit for 100,000 GPUs. Analysis reveals how data egress fees legally lock MSMEs into…

Plus, the Standard gets tech help from journalism’s newest whiz kid

AI Luddites • An Elon/OpenAI Verdict • Big Tech's Emissary Movement • Apple's 'Dinged Can' Strategy • Siri Heads Back to "Beta" • The Great SF AI Divide

I saw something in the Hermes Agent Discord that stopped me mid-scroll. Not a feature announcement. Not a breaking change notice. Not a philosophical debate about agent architectures. Just someone —…

Here's Zac Hill's excellent new substack, Manual Transmission , and here's Nate Meyvis' Bear blog . These are two of my favorite writers on the internet right now. Highly recommended. 1 Both of them…

You finish the day with thirty things still on your list. You ask your AI assistant to prepare a summary of the week’s meetings to send to the team. You give it five minutes, trust the output,…

AI can do this, AI can do that. Automate this, automate that. You don't know. You come from another time, where most sales were done face to face. Based on trust. But this new technology is here. The…

John Voorhees: Late yesterday, Anthropic announced messaging support for Claude Code, allowing users to connect to a Claude Code session running on a Mac from a mobile device using Telegram and…

Broken clay tablets have carried the written memory of the ancient Near East for thousands of years. Now, a new AI-powered tool is helping scholars compare the smallest details of Hittite cuneiform…

A few years ago, I wrote a post describing how I asked ChatGPT to solve a couple of elementary beam bending problems and how its answers were persistently wrong, even after I told it the mistakes it…

Speaking of the major LLM players, xAI is the only one that hasn't been monetizing developers and programmers until now. It seems they are starting to change that. Cursor and…

On June 30th, 2025, I started working at Stainless as a part-time contractor. A couple months later, on August 21st, I joined the team full-time. Today, on May 18th, Stainless announced they've been…

PolyAI opened its enterprise voice-agent platform to every builder, free for two months. It is practical, overconfident, and much sharper than the average AI call-center sermon.

A follow-up point on my “ AI Is Technology, Not a Product ” column over the weekend. Here’s a repeat of Steven Levy’s argument that John Ternus must direct Apple towards building “a killer AI…

Via: Boston Dynamics:

AI productivity gains are real. Whether a firm captures them is decided by the operating model around the tool. A note on what comes before procurement.

The frontier labs are building deployment companies for enterprise AI. OpenAI should steal the better Apple playbook: make AI physical, approachable, and easy to try.

The Anatomy of an AI-Native Org argued that the middle of the org chart collapses because translation work is the work that gets eaten. This is the role-by-role walkthrough for anyone who asked the…

Agents need job descriptions. If you’re building with true agentic frameworks – OpenClaw, Claude Managed Agents, Hermes Agent, etc. – you’ve likely noticed that getting an…

Build production-grade stateful AI agents in 30 minutes: master graph-based workflows, tool nodes, human-in-the-loop checkpoints, and persistent memory with local Ollama. No cloud APIs required.

I work on AI token streaming at Ably. Mental model Most people’s mental model of AI applications is stuck in November 2022. It’s a mental model of a chat window, a back and forth…

Around July last year I decided I was going to go all in on technical AI safety research.

While everyone is expecting Gemini AI 3.5 to be released tomorrow, May 19, at Google I/O 2026, some changes took place in advance. Starting on May 17, 2026, there are […]

Abstract/TLDR: LLMs are notoriously formulaic at writing, overusing certain tokens or phrases. I show that models trained with SFT fail to match the distribution of the training data by using Maximum…

Today, we're talking prediction markets, higher education, and data center job creation. The post The data center buildout won’t save employment in the AI era appeared first on Cautious…

Parallels with the Age of AI...

AI can be used to model the spread of Alzheimer’s disease through the brain and has now provided researchers with a more biologically grounded way to predict cognitive decline. Details are reported…

At Devoxx France 2026, I had the pleasure of presenting a session on Agentic Design Patterns. In this talk, I explore how to move beyond basic LLM wrappers to build reliable, scalable, and…

Next Wednesday I am giving an in-depth workshop on how to build and operate your own Second Brain for your AI agent.

| LessWrong post | I couldn’t find a recent write-up from a MATS alum about what attending MATS was like, so this is the thing that I wish I had.

Most AI second brain setups fail because the vault isn't built for it. This covers the structure, linking scheme, and agent protocol that actually make it work.

Today, AI tools are mostly solo sports. Developers write more code. Designers create more images. PMs crank out more docs. That's cool... but it's more cool to work together. So how does that work…

AI is being sold as a uniformly transformative force. The reality is more nuanced. The most impressive gains come from one domain, driven by years of focused engineering. Here's what that means for…

A rant about a rant you'll never read

A synopsis from NotebookLM …’almost a love story’ Mon 18th May 2026 The sources suggest that this is indeed an "almost love story," but one that shifts its focus from human-AI romance to a more…

https://miroslawstaron.github.io/hallucinations.html#/5 The shift from manual coding to AI-augmented orchestration is no longer a future – it is a reality. Software engineers adopt AI…

From Theater of War: A.I. Is About to Solve Loneliness. That’s a Problem , an adaptation of an essay by Paul Bloom , performed by Oscar Isaac, Bill Murray, and Lois Smith. On stage at the Brooklyn…

There is something quietly political about choosing a typeface. It says: I thought about how this would arrive. Gotham Variable arrived this month, and almost nobody noticed. Perhaps that is the most…

Debates over disruptive tools follow a pattern. The vibe coding fight is at a recognisable stage of it. Notes from someone who passed through an earlier version of the same argument.

I can only imagine what it's like learning the skill of programming in this day and age. What does an average college class look like? What is the CS professor teaching? And students, how do you…

"Tokenmaxxing" is, roughly, attempting to use as many tokens as possible. Much of the conversation about it focuses on questions like these: Should we really be doing everything with AI? How much are…

The graduates were not wrong to boo. Eric Schmidt stood at a podium at the University of Arizona on Sunday and did what every technology executive does when they face a room full of people who are…

So, não se tem uma tradução, mas pego emprestado a tradução mesmo para discutir aqui porque no próprio recurso se usa also, besides harness, o conceito de saddle. E se usa uma analogia com um…

Dear American researchers claiming “states and powerful institutions have increased strategic incentives to leverage media control”, please start with America. Here’s a quick recap…