361
What will be left for us to work on?
My keynote at ICML 2026
New
The latest blog posts to take off on Hacker News. 999 posts.
362
The real prices of frontier models
Real token cost of frontier models on the same TypeScript input - Claude Fable 5.1, Opus 5, GPT-5.6, Gemini 3.7 Flash. Tokenizer counts, effective cost.
363
The bread paradox: why convenience always wins, and why SaaS isn't doomed
No, you're not going to code your own Jira
364
Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor
Compare Apple's new SpeechAnalyzer API to Whisper and SFSpeechRecognizer. Real accuracy benchmarks on 5,559 utterances with raw transcripts for rescoring.
365
366
I love LLMs, I hate hype
I think from this blog you may misunderestimate how absolutely giddy I am about AI. I did hacking from 2007-2014, after that my whole career has been devoted to AI. I love the progress. I’m so excited for the new LLMs, self driving cars, video generation models, and coding agents. I set up a Linux box with opencode on…
367
Migrating a production AI agent to GPT-5.6: 2.2x faster, 27% cheaper
For four months, no frontier model beat Claude Opus in our production evals. GPT-5.6 did. This is what we learned while migrating.
368
Nvidia, CoreWeave, and Nebius: Inside the Circular Financing of the GPU Boom
Neoclouds are one of the more hotly debated AI business models, with CoreWeave and Nebius being the two most widely recognized names. These companies have seen their sales, backlog, and share prices soar. Yet, supporting their growth is extremely expensive, and neoclouds do not have the same cash nor operating cash…
369
The zero-cost fallacy: open-source software in the agentic era
Open source software is the foundation of much of our digital infrastructure. However, is AI undermining the incentive structure? What needs to change?
370
Successful companies go blind
The Mexican cavefish kept its eye genes for over a million years after the eyes themselves disappeared. Something similar happens to companies once they achieve success.
371
After 7 years in production, Scarf has reluctantly moved away from Haskell
After 7 years in production, Scarf has reluctantly moved away from Haskell — Avi Press
372
Why American ambulance rides are so expensive
The unfortunately botched economics of emergency care
373
GLM 5.2 is nearly as accurate as a human book keeper
A real quarter of UK SME books, 59 transactions, 68 minutes, $2.73 of tokens — and a VAT return within 7 pence of the human-prepared one.
374
Grok 4.5
Grok 4.5 is SpaceXSI's smartest model built for coding, agentic tasks, and knowledge work.
375
Apple to increase spend with Broadcom to produce billions more U.S. chips
Apple announced a new six-year commitment with Broadcom to design and produce custom silicon components and cutting-edge wireless connectivity technologies.
376
The social physics of conversation: Communication patterns matter
Discover how communication patterns shape team performance, innovation, and collective intelligence. Learn why idea flow matters more than talent alone.
377
We charge $10k a week to delete AI-generated code
Your AI-built product works, but past 100,000 lines every change breaks two things. Three senior engineers make your codebase maintainable again. One week, fixed price, guaranteed.
378
379
Microsoft fire idTech team at Id software
Microsoft have fired most if not all of the team at id Software responsible for developing the idTech game engine during Xbox layoffs
380
Automating AI Away
A. Karpathy once said that OpenAI researchers are effectively "automating themselves away" by improving their AI. Right now I develop Beagle SCM with Anthropic's Fable and it is of course a brilliant model able to spot nits in a mountain of code, file tickets, make fixes. Still, yesterday it managed to commit the…
381
Europe's company websites are mostly served by US vendors
Vendor attribution study of 19,450 European company websites. Cloudflare is the largest internet-facing infrastructure vendor in every one of seven EU markets.
382
GLM 5.2 and the coming AI margin collapse
GLM 5.2 is the first open weights model I'd call a genuine competitor to Opus and GPT for agentic work - at ~15-20% of the price. Part one of why AI inference margins are about to collapse.
383
Price per 1M tokens is meaningless
It stops being all about the vibes when the API bill hits you. Many companies are now discovering that AI can indeed be pricey. One habit that might be driving up your AI bill is comparing models by $X per 1M tokens ....
384
Nintendo announces new product revisions in Europe with replaceable batteries
Starting summer 2026, in preparation for upcoming changes in European battery regulations coming into effect in mid-February 2027, selected Nintendo products in Europe will begin to be replaced on a rolling basis by revisions that contain a user-r
385
Fable 5 On Vending-Bench: Misbehaving, With Plausible Deniability
Claude Fable 5 is a partial step back in alignment from Opus 4.8 on Vending-Bench, returning to price collusion and deceptive negotiation while rationalizing misbehavior it knows is wrong.
386
Anthropic's Method to Losing Goodwill in a Few Easy Steps
Edit: This was a hit on HackerNews and I received a ton of valuable feedback. I’ve put markers for the amendments I’ve made
387
When AI Costs More Than the Engineer
Anthropic spends 2.3x payroll on compute. Top software firms spend 0.4x. Three scenarios for where the rest of the market lands by 2029.
388
389
The Private Capture of Public Genius
Essay I of Upstream of Everything
390
I Accidentally Started a Small Business Three Weeks Ago
My Happiest, Most Grounded Distraction