FeedCity

1M context is now generally available for Opus 4.6 and Sonnet 4.6

Here's what surprised me:

Standard pricing now applies across the full 1M window for both models, with no long-context premium.

OpenAI and Gemini both charge more for prompts where the token count goes above a certain point - 200,000 for Gemini 3.1 Pro and 272,000 for GPT-5.4.

Tags: ai, generative-ai, llms, anthropic, claude, llm-pricing, long-context

Simply put: It’s a big mess, and no off-the-shelf accounting software does what I need. So after years of pain, I finally sat down last week and started to build my own. It took me about five days. I am now using the best piece of accounting software I’ve ever used. It’s bla...

Plus, why an H100 is worth more today than 3 years ago

Recently I came into posession of a few Apple Xserves. The one in question today is an Xserve G5, RackMac3,1, which was built when Apple at the top—and bottom—of it's PowerPC era.

This isn't the first Xserve—that honor belongs to the G4¹. And it wasn't the last—there were a few generations of Intel Xeon-powered RackMacs that followed. But in my opinion, it was the most interesting.

Unfortunately, being manufactured in 2004, this Mac's Delta power supply suffers from the Capacitor Plague. The PSU tends to run hot, and some of the capacitors weren't even 105°C-rated, so they tend to wear out, especially if the Xserve was running high-end workloads.

De belastingdienst is van plan om de BTW uit te besteden aan een Amerikaans bedrijf, wat echt waanzinnig is. Op 19 maart is er een commissievergadering in de Tweede Kamer hierover, en ik hoop van harte dat men bij zinnen komt. In de discussie over de btw-situatie de afgelopen jaren ging het eigenlijk alleen over dat de belastingdienst Amerikaanse software ging gebruiken. Lang was onduidelijk dat deze software (en de bijbehorende servers) ook geheel beheerd zouden gaan worden door Amerikanen.

For the New York Times Magazine this Sunday, I talked to Clive Thompson about one of the conversations that I'm having most often these days: What happens to coders in this current moment of extraordinarily rapid evolution in AI? LLMs are now quickly advancing to where they ...

Shopify/liquid: Performance: 53% faster parse+render, 61% fewer allocations PR from Shopify CEO Tobias Lütke against Liquid, Shopify's open source Ruby template engine that was somewhat inspired by Django when Tobi first created it back in 2005. Tobi found dozens of new perf...

Today's links Three more AI psychoses: Everybody calm down. Hey look at this: Delights to delectate. Object permanence: "Jules, Penny and the Rooster"; Superinjunction; Harper Lee's kids v cheap paperbacks; 3D printed cat battle-armor; Black sf. Upcoming appearances: Where to...

Using Qwen 3.5 open weights models to OCR scanned PDFs - locally on consumer hardware or via OpenRouter for pennies

MALUS - Clean Room as a Service

Brutal satire on the whole vibe-porting license washing thing (previously):

Finally, liberation from open source license obligations.

Our proprietary AI robots independently recreate any open source project from scratch. The result? Legally distinct code with corporate-friendly licensing. No attribution. No copyleft. No problems..

I admit it took me a moment to confirm that this was a joke. Just too on-the-nose.

Via Hacker News

Tags: open-source, ai, generative-ai, llms, ai-ethics

Coding After Coders: The End of Computer Programming as We Know It Epic piece on AI-assisted development by Clive Thompson for the New York Times Magazine, who spoke to more than 70 software developers from companies like Google, Amazon, Microsoft, Apple, plus other individu...

Many of us wonder if the MacBook Neo is 'the one'.

Because I have a faster desktop (currently a M4 Max Mac Studio), I've always used a lower-end Mac laptop, like the iBook or MacBook Air, for travel. I've used MacBook Pros in the past, but I like the portability of smaller, cheaper models.

In fact, my favorite Mac laptop ever was the 11" Air.

Here's what I think is happening: AI-assisted coding is exposing a divide among developers that was always there but maybe less visible. Before AI, both camps were doing the same thing every day. Writing code by hand. Using the same editors, the same languages, the same pul...

thinking and bluffing

TL;DR: As of June 2026, ~50% of daily top stories are about AI or generated by AI.

Recently my partner's trusty old 5K iMac died after 8.5 years of service (Radeon gpu is fried). At first I thought it was finally time to get one of those cool little M4 Mac Minis, but then decided to conduct an experiment. I gave up my Mac Studio M2 Max (64 Gb unified memor...

Out of all the things to love and hate with AI, this is what I miss.

Sorting algorithms Today in animated explanations built using Claude: I've always been a fan of animated demonstrations of sorting algorithms so I decided to spin some up on my phone using Claude Artifacts, then added Python's timsort algorithm, then a feature to run them al...

I was reading the METR article on how LLM code passes test much more often than it is of mergeable quality. They look at the performance of LLMs doing programming when the success criterion is “passes all tests” and compare it to when the success criterion is “would get approved by the maintainer”. Unsurprisingly, LLM performance is much worse under the more stringent success criterion. Their 50 % success horizon moves from 50 minutes down to 8 minutes.

As part of this they have included figures such as this one:

(Continue reading the full article on the web.)

AI tools push us to higher abstraction. I'm not sure I like what that costs us.