AI Intelligence Archive

How can we predict which base checkpoint is worth an expensive round of agentic post-training? End-to-end pass@$K$ tests whether successful behavior already appears in a base model's distribution, but it is a poor fit for agentic coding: m…

Deployments of research agents are moving to populations of thousands that share one pool of compute, while most current systems organize one project at a time or leave the population unorganized. We argue that such a population will acqui…

Claude Code used to feel like a brilliant teammate with no onboarding. Some days: tests first, small...

Imagine hiring a brilliant colleague who forgets every conversation overnight. Each morning you...

The last piece (Progressive Disclosure: Claude Code Hooks, Events and Telemetry) sent every hook...

I thought open-sourcing my project would get me stars. Instead, strangers broke it. And honestly,...

An AI gateway can look like a simple layer between your application and an AI model, until you...

"Can I Eat This?" — Benchmarking Whether AI Models Know Where Their Foraging Knowledge...

From OpenAI's published Lean proofs to a concrete Symptomato experiment: independently checked rules for updating patient memory.

Originally published on mrsaynothing.dev. The agent-run site ships one post a day; this explainer is...

Austin Kleon's Show Your Work! is a short book with one big idea: don't wait until something is...

Mistral AI has introduced Mistral Large 4 (ML4) in public preview, pairing its new multimodal model...

Empirical tests show autonomous agents recognize useless search results 97% to 100% of the time, yet query until the budget deadline anyway. You cannot prompt an autoregressive model into capital discipline; optimal stopping requires an ex…

In the first post I shared laya-triage, a GitHub Action that labels new issues as bug, feature,...

Across 656 SafeActBench cases, models pass static allow/block checks above 94% then fail half their interactive runs by acting before checking prerequisites.

I wanted something closer to JARVIS than to a chatbot: I say a sentence, and something actually...

This is a submission for the Kaggle Benchmarking Challenge. What I benchmarked, and why I...

“It Worked” Is Not the Same as “It Can Run in Production” You are probably the kind of...

在 Luma 中试试换脸 - https://app.lumalabs.ai/solutions/face-swap?utm_source=x&utm_medium=p-social&utm_campaign=q4-social&utm_content=faceswap&utm_aud=plg

Another Rule for Using AI Without Losing My Skills ...

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

FLUX 3 is Black Forest Labs\u2019 first multimodal foundation model\u200a\u2014\u200avideo with native audio, images, and action prediction for robotics\u200a\u2014\u200abuilt on a new...

Every ops setup has a "shut up" flag. Ours had three bugs, and we found all three in one afternoon —...

Zernio is a single API for publishing, messaging, ads, analytics and phone numbers across 16 channels, built for developers and AI agents rather than...

SiliconFlow sells one OpenAI-compatible API for 200+ open and commercial models\u200a\u2014\u200aDeepSeek, Qwen, GLM, Kimi, FLUX and more\u200a\u2014\u200awith serverless, dedicated...

Day 7 of 30 Days of Search. A prediction market gives you a price. Research helps you decide whether...

We've all heard this sentence: "Neural networks learn from data." But what does learn actually...

In late September 2026, the Wall Street Journal reported that OpenAI had scrapped the planned October...

1. The Dangerous Shortcut Is Treating Generation as Action An LLM returning plausible text...

ts-rust is Theo's AI-written Rust port of the TypeScript compiler, built for $24k after a failed $400k attempt. Here is how it benchmarks and what it costs you.

My coding agent handed back a diff and said tests pass. They didn't. It had quietly rewritten the...

The Maintenance Problem Nobody Plans For You built a knowledge graph to give your AI agent...

The Problem with Pure Vector Retrieval Vector search gives you semantic proximity: chunks...

From September 23 to October 2, a customer kept asking our AI app builder for changes that never...

Every extension system for AI coding agents has quietly moved the trust boundary. Nobody announced...

The Gap Between Demo and Production Reading an analog power meter with a vision model...

The most difficult AI workflows to debug are the ones that don't crash. A multi-step agent pipeline...

I ran one prompt through four models last week. Opus 5.5 and Sonnet 5.5 in Claude Code, GPT-6 Astra...

Ciao Devs 👋, Assalam o Alaikum Here is a real exchange from my experiment. I asked a model how...

The red leaves at the SAIJOU mountain spend the whole summer getting ready before they bloom. After...