Microsoft, Google, Red Hat, Grafana ve LlamaIndex public GitHub repolarına canlı verified credential pushladı. Pipeline'ı ben yazdım: Gemini 2.5 ile etiketle, regex'e distil et, TruffleHog'u active…

Grimmett and Li solved a small statistical-mechanics model — the polygon model, a way of weighting the closed loops you can draw on a lattice — on the honeycomb, and then named the next step they…

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…

Many developers are watching AI agents generate bad code, introduce bugs, and erode their skills. I'm having a different experience. The post AI took my job as a coder. Then it made me a better…

Yesterday 28 June 2026, I had the honor of speaking at the YMCA National Planning Forum in Alwaye, Kerala. We tackled a big topic: how Artificial Intelligence fits into a historic service…

Scientists have a name for what happens when artificial intelligence runs out of human data to learn from. They call it model collapse. It does not happen all at once. It happens in stages, and the…

Generative AI has made plausible ideas nearly free, and a fashionable argument says this changes everything. It doesn't. Ideas were never scarce—judgement was, and remains so. On Grub Street at…

Tech bros still dominate the discussions about so-called “AI” with false claims. Even most “AI”-critical researchers spend much of their time meticulously debunking (always…

I've avoided going deep with LLM usage, including for agentic coding. This is has been for a number of reasons. I don't want to be sending data to US tech platforms is a big one. I don't want to…

There are two ways to use a coding agent: sit in its terminal, or take its face off and drive the engine from your own code. If you're building anything around agents, the second mode is the whole…

AI-generated poetry is preferable to human-written poetry? Oh, we are actually done for. The authors describe the reason in the results: The AI poems are more accessible. My house is full of cowards…

Josh chats with Allan Friedman about all things Bill of Materials. Allan did a ton of work to help turn SBOM into what it is today. He has many thoughts and ideas around the new types of BOMs, a…

The moment you start tracing an AI app, three words turn up everywhere: span, trace, session. People sling them around as if they’re obviously different, and if you’re new to this you nod…

Is Agent Code Less Maintainable? What CodeThread Reveals About Codex CLI Output and How to Defend Against Drift

Over-Privileged Tool Selection: Why Your Coding Agent Reaches for the Admin Key — and How Codex CLI’s Hook Pipeline Stops It

The model is the easy part — you can rent it via an API. Building a product around it is software development. Want to sell AI? You need developers.

Human Agent in the loop I dislike the phrase “human in the loop” because it cedes authority to the machines. Let’s flip the narrative. It’s our loop, we work the same way we always have, now we…

Earlier this week I read Kyle Chayka’s latest for The New Yorker , called “ The A.I.-Design Aesthetic That’s Taking Over the Internet .” It’s about Anthropic’s Claude Design and some of the visual…

I think I'm probably gonna go to the grave bitter than tech bros and the billionaire class are starting to make me hate robots. Robots (in fiction) have been my beloved little guys, my conduit, my…

At what stage of AI model evolution are we at?

I recently added Ornith 1.0 to the Will It Mythos benchmark. Ornith is a Qwen or Gemma 4 post-train (the one I tested is based on Qwen 3.5 35B MoE, I believe) that teaches the model to build its own…

I’ve been thinking a lot about code generation with AI agents lately, and wanted to start writing a bit about how I’m approaching the pro

I quite enjoyed this talk from Bryan Cantrill where he discusses the difficult engineering problems they overcame while working on their company Oxide. Some of the problems they ran into were bugs.…

A new Berlin Story report, about drones attacking Ukraine, discusses the NVidia AI hardware used by Russia. Inside the Russian Zala drone, we found the NVidia TX2-A (Jetson Tegra X2) AI chip with 8GB…

A 2026 field guide to ChatGPT, Claude, Gemini, Grok, Meta AI, Mistral, Cohere, Qwen, Perplexity, companions, and agentic AI.

I recently spent a few hours viewing a huge trove of videos on the YouTube channel Fantasy Universe. These are all generated by some unidentified person or group using artificial intelligence.…

Mythos gets reinstated, but we may be seeing the end of widely shared leading AI models.

Everyone talks about what AI can do. Nobody talks about what it costs you. A practical look at context switching, the vampiric effect, and how I keep agents on my terms.

OpenAI flinched at its own IPO, Micron printed software margins on RAM, and the robots learned to fear a safety review. The AI bill has arrived.

Lior Div runs an AI cybersecurity company. He wrote something this week that the AI industry needs to read twice. AI doesn’t scale by removing people. It scales by moving them closer. That is…

undefined

Via: Telegraph: Substantial debt-fuelled spending on AI is driving up the risk of a global financial crisis, central bankers have warned. The Bank for International Settlements (BIS) said on Sunday…

A human future, without horror.

Some time ago I vented about A.I. in a post named Assistive Incompetence . All my posts are written from a humanist and disinterested perspective. I have no control whatsoever about what people do…

Anthropic ran a randomized trial in early 2026 where engineers learned a new Python library, half with AI assistance and half without. Both groups finished the tasks at the same speed. But the AI…

Data scientists and clinicians provide their perspectives on how humans and AI can safely work together in intensive care units.

Starting is free now, so I start everything and finish nothing. On the compulsion loop of agentic coding — and why stopping and resuming are _the_ skills to play with now.

Before I go further, let me explain how this usually goes without AI. I leave the hotel, I walk around, I look for a place with lots of local people in it and if I don't find one soon enough, I pick…

A bronze pin leaves the earth, and the future discovers it has been shouting over older machines.

In five to ten years, frontier-grade AI runs on home hardware and the state loses central control of inference. So it does what it did with drugs: criminalize the capability and hunt it — a smarter…

Three ways to do it, and the plumbing that's reshaping all of them.

In this post I like to describe the current dilemma of using JSON Schema for structured AI outputs and how TypeSchema can be used as a clean, alternative data-modeling layer to generate both native…

Recently I’ve been experimenting with local LLMs for coding. Up until a couple of months ago I was a happy GitHub Copilot subscriber. I’d been quite satisfied to let other companies take on the cost…

blockquote.large { font-size: 1.2em; padding-top: 1em; } blockquote.large span { font-style: italic; } figcaption { font-size: 0.8em; text-align: center; } .image-columns { display: flex;…

_Almost Timely News: 🗞️ How To Improve Advertising with AI (2026-06-28) :: View in Browser_ The Big Plug 👉 My new course, GEO 201 on competitive GEO measurement, is now for sale. Content Authenticity…

This article discusses the reality of AI hallucinations in chatbots, why they occur, and how to mitigate their impact. Understanding AI's limitations is crucial.

Your weekly digest of Cloud-Native AI, Docker Sandboxes, and Model Context Protocol innovations.

Here’s a question I can’t stop chewing on: why is some software so damn good, and most of it so… not? Not buggy. Not slow. Just bad. Bad in a way you feel in your gut the moment…