pants2

I'd think an ethicist at Meta/OpenAI involves extremely challenging mental gymnastics

The frontier labs would rather charge a premium for the best capabilities. You joke but they will probably soon have a GPT-bio that they license to Pfizer and a GPT-quant that they license to Citadel, rather than keeping those valuable capabilities in the public models.

You're right. Wh/day is a terrible unit!

A quick search suggests one hamster can produce around 0.5 Wh/day, so 100 hamsters should easily power your phone laptop with a constant 50W!

There's a super obvious tell, 5.5 instant loves emojis.

Yeah, something very cool is XOR can only improve randomness as long as there's no correlation between the sources. So Linux RNG XOR'd with this comment is even better than it's source!

Crypto people make fun of me for making up passwords in my head, but my brain RNG has never been exploited, nor has anyone successfully Sherlock'd my password.

You must be fun at parties

There's a staggering amount of new, basic software which often goes unmaintained quickly because no discipline was needed to build it. But I don't think we're seeing, for example, Adobe releasing new creative suite tools faster than before. This isn't something I can quite quantify, though.

After all this time it's still a pain point just to pay them for a Gemini API key.

Blockchain protocols are also faced with the most active, persistent, and well-funded state sponsored attackers of any industry, so naturally formal verification should be of great interest.

I think Grok is replacing composer now yeah? Fast mode is similar speed to composer.

Right - one would assume that if they're on the verge of an ASI that will consume the entire economy, advertising is only a distraction.

Perhaps with strategic striping on the ball they could also get angular velocity, could come up with some very cool stats. Or if you're really having fun, microchip every ball with an accelerometer / gyro

At least you could display a message that says "this lane is reserved. Open lanes: 6, 7"

Forget Baba Is You, I want to see an LLM beat Elden Ring

Near the end of 2025, Google was on top - best LLM (Gemini 3 Pro), best video model (Veo 3.1), and best image model (Nano Banana). They seemed unstoppable.

And then, they stopped. Why? Are they having trouble with TPU?

When even Meta can come out with a better LLM than Google, something must be terribly wrong.

The Artifical Analysis has a link on their homepage but it 404's :/

https://artificialanalysis.ai/models/inkling

I typically disable autocorrect on Apple products because of this, cautiously optimistic about their improved speech models, but definitely worried that it's going to 'correct' technical jargon to more common words.

Kinda funny that their "cost-vs-performance" chart looks the same as the one for Composer 2.5[1], except that it includes Composer 2.5 at a completely different spot.

What are the chances that CursorBench ranks Cursor's model highest, and Cognition's bench ranks Cognition's model highest? Both are to be RL'd from Kimi as a base model, BTW.

I'd posit that it's not deliberate deception, but for both companies their training data and benchmarks come from the same dataset (Devin/Cursor interaction logs) so they naturally overfit.

1. https://cursor.com/blog/composer-2-5

If you're developing on top of LLM APIs directly, this is definitely not true. There are differences in how context caching works, in what's available through native harnesses, the types of tools you're fine-tuned on (GPT uses apply_patch while Claude uses edit, with different formats), the API surface (Agents SDK, Responses API, Managed Agents), cost structures, and best-practice guidance all around.

Not to mention the meta of account limits, billing, ZDR contracts, etc.

Only because one is a short answer and the other is not

When you're a student in a competitive program at a top university, graded on a curve, and you know your fellow classmates are cheating with AI, you have little choice but to do the same. Especially when jobs for new grads are harder to come by and there's more pressure to also go above and beyond with internships and side projects during your time in school. There's no way to compete without cheating.

Yep, I knew this was coming and did the same. Can't believe my MBP is now an appreciating asset!

This is insanely cool. My wallet is ready.

I built something similar at my org. Users simply connect the agent via OAuth and that inherits all of their permissions, so it acts as them.

What's cooler is then it can view/add/remove people from channels, so it can conduct access reviews -- overall I consider it a security improvement.

That has been going on for a while AFAIK

Is this new news?

> This policy was published on June 8, 2026 with an effective date of July 8, 2026

I recently set up an agent with a Base wallet and made a few x402 API calls. Pretty convenient!

Yeah, if they can fit an 8B model that's really good at improving the output by thinking, running at 16K tok/s on Taalas would be mind-blowing.