CEOs Who Think AI Replaces Their Employees Are Just Bad CEOs
In the last three months I’ve had people forward me four separate examples of a CEO losing his or her mind over AI. What’s been striking to me is the similarity in each case: It would b…
Claude Fable 5 and Claude Mythos 5
Today we’re launching Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
GitHub - luke8086/gentleos32: Hobby operating system for vintage 32-bit PCs
Hobby operating system for vintage 32-bit PCs. Contribute to luke8086/gentleos32 development by creating an account on GitHub.
Microsoft's open source tools were hacked to steal passwords of AI developers | TechCrunch
Microsoft shut down dozens of GitHub code repositories for Azure and AI coding tools after a reported hack.
Catlantean 3D - Making Graphics Like It's 1993
Personal website and blog.
Run AI models in your app on Apple silicon.
Week 5 on Nick Launches: founders on launch timing and why you'll never feel ready, a featured builder, a reciprocal shoutout, the top launches, the numbers, and what's next.
OpenAI confirms a confidential S-1 submission to the SEC and has not yet determined timing for further action.
Popular initiative: ‘No to a Switzerland with 10 million! (Sustainability Initiative)’
On 14 June 2026, Swiss citizens will vote on the popular initiative: ‘No to a Switzerland with 10 million! (Sustainability Initiative)’.
The Future Is Undecidable. Duty Is Not.
A follow-up to the India 2047 song essay, moving from Godel, Turing, and the halting problem toward anxiety, controllability, the Gita, Stoicism, and the old problem of duty.
Joshua Hedtke
Joshua Hedtke
I used the leaked Claude Code to make a coding agent which is self improving, more efficient, and a better, self-improving coder.
SQL is the language for databases. A SQL program is also called a query. Each query takes one or more tables as input, and produces one table as output. A table is a collection of rows, where each row has the same number of columns. Each column has a name and …
When I first saw the title of OpenAI’s blog, I wasn’t too excited. Here it is: Building self-improving tax agents with Codex. Tax agents don’t sound exciting. Everything is self-improving. I got it. But this short post is a masterclass in practical, hands-on l…
If you self-host LLMs, at some point, you’ll likely experience a stuck GPU consuming significant electricity for long periods of time. This is my summary of discovering the problem & then implementing automation that corrects the issue. I want automation to …
I'm building a parallel internet, and it's called The Thinnernet
Since 2020, I've taken on a Steve Jobs alter-ego (on a very, very part time basis, and often for humor purposes ). The firs...
The Data Center Bankshot
The
A colleague of mine was recently discussing a connectivity monitoring system he is working on with me. It’s nothing fancy, just sending ICMP Echo Requests to a couple of different servers, and monitoring latency and dropped packet averages over 1-minute, 5-min…
2026 Apr 28 See all posts
I have been feeling that I am not enjoying the process lately, I am not having fun while I am working, previously I used to think about the problems even on weekends while out with friends, because they were my problems and I used to own them end to end, solvi…
Most advice on performance will only help you to get better incrementally. Here are some mantras I think are undervalued for true breakthroughs 🧘.
The economics of speculative decoding
Two underexplored axes: what MoE routing does to the decode roofline, and how compressed attention takes away the slack that used to make speculated tokens free.
Making numpy-ts as fast as native — nico.codes
How numpy-ts went from 15x slower than NumPy to performance parity
Cannibalism
For a long time the tech industry revelled in the distruption of what it saw as old, legacy industries. But now, as AI takes over tech, we’re now starting to eat our own, and it’s dark and ironic.