You can't fact-check everything... ... there's just too much. It's the proverbial drinking from a fire hose... even if you're just a dog on the internet. ( Slightly obscure reference .) The past few…

This is the kind of image that accompanies stories on human-AI love. This robot looks horny, but it’s not. JERSEY CITY, OCTOBER 18, 2026.  Ayrin falls in love with Leo. He’s always there for…

Imagine you're building an eval to test whether an AI coding agent can deploy an application to your cloud platform. You want to verify that the generated application actually works, so you need to…

_Almost Timely News: 🗞️ What Are Open Weights Models? (2026-10-18) :: View in Browser_ The Big Plug(s) 🚗 Missed MAICON 2026? You can catch the replay of my session here! ✍️ Enroll in my new course,…

From Matheus Lima: Before you tell the agent to start building anything, you go there and read and understand it first. You don’t need a complete implementation plan, but you need a rough idea of…

Practical lessons on using AI in software engineering: beating procrastination, fighting code entropy, and why system design matters more than ever.

I finally read an article that I have in an open Tab since a couple of weeks. I procrastinated, because it was rather long (> 5500 words) and I didn’t just want to rush through it. The article…

Hi there, This post is an “I built a thing” post about krea2-prompting (opens in new tab), an agent skill for writing Krea 2 prompts I built recently. Basically, that’s it: try the…

Existing in 2026 means hearing all manner of AI-related discourse. I'm intensely bored of it. Wake up babe, FooAI dropped a new model; it's slightly better than the last model and marginally cheaper.…

Can artificial intelligence identify positive or negative sentiment in encyclopedic texts? Do different language models assess the same content in similar ways?These questions are explored in the…

Building lasteverLM in 78 hours with a head agent, a Linux VM and 20–24 hours of my own time — and what I want from fleet tooling next.

I think the title captures the idea fully but I'll dig into it because this ain't Twitter. You're working an idea, a piece of code, or a multibillion a year SaaS app which is probably…

This week the theme is scale: what breaks when agents write, merge and review code faster than any human can keep up. Five posts cover Anthropic’s new dynamic workflows for Claude Managed…

For the past few months, Silicon Valley has been pushing AI agents and agentic commerce on consumers. While there are open questions found with what problems are being addressed by the technology,…

A month ago, I received an email out of the blue from someone named Ioannis Tsiokos. He said that he had achieved a breakthrough in tiling theory: he had discovered a strongly aperiodic monotile in…

A playful thought experiment about simulated universes, time-skipped workflow tests, consciousness as a context window and the back door to God's admin panel.

ProtocolPolice I had the weirdest dream last night. I installed an app on my phone so I can look up the prices for couches. That makes sense, it's time to get a new couch in our living room. It…

Back in October 2023, I wrote about the multi-layered cake of GenAI development. My point then was simple: building a serious GenAI application involved much more than prompt engineering. Models,…

I have access to a Large Language Model. It can do things for me. Most anything I can imagine done with ones and zeroes. It used to be that I’d come up with an idea, and then had the puzzle part…

Laplace once said: An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed, if this intellect were also vast…

Après avoir passé l’habituel chatbot parfaitement inutile qui ne comprend pas une question élémentaire, le support bpost daigne me mettre en contact avec un humain. Mais est-ce vraiment un…

Sixteen months ago my CNN beat HastyBot 51.7% of the time. The new transformer value net, macondo-tf, wins 57.65% over 100,000 game pairs, level with a 2-ply Monte Carlo simulation while thinking…

Lab 1 looked at memorization in LLMs and at the several different things that can mean: extractability, membership inference, counterfactual memorization. This lab focuses on extractability, and uses…

I am excited to share that the complete manuscript of LLM Customization and Fine-Tuning: Adaptation, Distillation, and Alignment, written with Weehyong Tok , is now in Early Access from Manning…

Everyone is complaining about low-quality AI code and 'vibe coding'. But code quality didn't crash because of AI. It crashed because humans are rubber-stamping code they didn't review.

In this blog post, we will explore the economics of decision models like Jev, and discusses the formalization of decision hierarchies using Chow's reject rule.

Here’s Spencer Tweedy on the strange new layer of computing that is the large language model : In the earlier styles of computer instruction, the interpretation process could only translate your…

Agents sit between implementers and reviewers now, and code review has gotten lonelier because of it. Sharing agent sessions can bring the humans back.

Previous parts of Build a Basic AI Agent From Scratch: Basic Agent Tools Long Task Planning Human in the Loop & Security Security II Security III Memory You can find and clone this code in this…

This week, I learned: I tested a few tools that help agents work with mobile devices. On Android, apart from adb, mobile-mcp seems better than Android-MCP or mobilerun (which has a different purpose,…

We did the almost-annual unconference, this time at the Saalbau Bornheim. The attendance was expectedly low at 5 persons, one person joining us from Amsterdam and one person joining us from…

Eres un Knowledge Worker o trabajas en tecnología y tienes dudas sobre el futuro con todo esto de las LLM, agentes y demas?… aca dejo unos pensamientos que tengo desde hace rato, pero hoy como que…

In the previous post, I explained why I don’t want to use AI as a person who values understanding. In this post, I want to talk about why I don’t want to use AI as a person who values art. A few…

If you haven’t accepted that these things are smarter than you, even in your own domain of expertise, I’m sorry. And if you have, but still believe you should be making all the meaty calls in that…

I'd like you to believe I'm real person typing these words but can't think how to convince you. It's more likely this was bot-produced by some corporation or government with a profit motive. I aim to…

In which Ben asks the pertinent question:“If the Codex team has unlimited tokens on really powerful unreleased models, why is the product getting worse?” I think he underweights focus and…

Let's start with an analogy, because everybody loves that. You come to your kitchen and find a puddle of water forming around the sink. You call a plumber, takes a look, and gives you the verdict…

The first thing they teach you about fleet stewardship is that an agent never wants anything. It sounds like a trivial point until you are running forty of them.

The voice in your head is not yours. It’s an advertisement urging you to buy a drink from the nearby vending machine. By now, you’ve learned to ignore them. The messages are annoying, but it’s a…

This acceleration makes human attention the bottleneck for agent output. Even among non-technical employees, reading all the output of their Claude agents would take half their work week. In…

The “math vs. AI” wars are continuing to heat up. I predict things will become more intense in the coming months. OpenAI and Anthropic will likely release new proofs, possibly including…

Your coding agent just scored 100% on an eval, and every criterion passed. There's only one problem: the code it generated doesn't build. How is that possible? The eval checked 20 aspects of the…

I’ve said it before; I’m no programmer. One of the stalwarts for open-source has been Linus Torvalds, and he’s seen a lot of code over the decades. It’s just been…

Comparing J-space, emitted reasoning, and final responses in a small language model, using self-harm and suicide-related scenarios.

Disclaimer: These are my own thoughts and opinions. They don’t reflect those of my employer. I’m discussing my own personal feelings about AI. I think it’s fair to say everyone is having to grapple…

Every independent gaming festival and showcase in the year 2026 requires some kind of generative AI disclosure. That is great! Slop and vibe-coded games are a scourge, AI is a technofascist,…

Everything is AI now. If you read my blog tracking every game developers using AI (or just paid attention to the news), you’d know that AI already has its tendrils firmly wrapped around every aspect…

Everyone is riding the agent orchestration wave and reinventing bureaucracy. Instead of starting from scratch, I used Marine Corps staff doctrine.

An interview with Kazuto Yamaki of Sigma that appeared in LensVid raised an interesting point. I wonder how many people caught it. "Sigma doesn’t use AI to design lenses, for now. Mr. Yamaki said the…