Where the AI market is talking

Rust SIMD on the GPU

On a CPU, hyperthreads are mostly replicated register banks. This allows the CPU to hold the context for 2 threads simultaneously. And, lets parts of a CPU make progress on one thread while the other thread is stalled. CPUs also has a kinda large microcode register bank that helps work around dependencies in asm instructions that reuse named registers. On the GPU however, the hyperthreads are just a round-robin execution queue to take advantage of instruction pipelining. The register bank of a single GPU core is huge and can be flexibly divided across a variable number of thread contexts when a kernel is launched. Many thread contexts can be held in registers simultaneously in a single GPU core. That makes stalling on memory latency much less of a problem. The hardware can focus on delive…

As AI eats the web, the internet’s collective memory is disappearing

Well, I guess another plausible explanation is that usually "domain experts" are people that get paid for their expertise, and have a vested interest in saying that LLMs cannot replicate what they want people to pay them for.

More than 10 firms pay up to $100k a month for access to Truth Social posts

The degree to which his works have been dragged through the mud by a bunch of people who insist on calling everything they don't like "capitalism" can't be overstated. The majority of people complaining about capitalism don't even have a coherent definition of it, it's a disgrace.

Linus Torvalds says AI has made 'huge' Linux kernel updates the new normal

"UNIX is simple and coherent, but it takes a genius (or at any rate, a programmer) to understand and appreciate its simplicity." — Dennis Ritchie, "Unix: A Dialectic" (1989)

Mars Bar from 1991 found – and it's 20g bigger than today's

Unintelligible. Re-write this into something coherent so I can respond if you like.

OpenAI's Only Ethicist Reportedly Left Last Month

How Claude marks AI-generated content

Why would I want a stochastic parrot that intentionally speaks less probable tokens? Anyway, I've found a magic line that can be copied & pasted to the comment sections of most OpenAI/Anthropic news threads. This one is no difference. The magic line: > Doesn't matter; have DeepSeek.

How Claude marks AI-generated content

Why would I want a stochastic parrot that intentionally speaks less probable tokens? Anyway, I've found a magic line that can be copied & pasted to the comment sections of most OpenAI/Anthropic news threads. This one is no difference. The magic line: > Doesn't matter; have DeepSeek.

How Claude marks AI-generated content

> Honest note: Anthropic has not shipped a public Claude watermark detector yet. This tool uses rewrite-based neutralization — a meaning-preserving paraphrase with a non-Claude model — which is the attack path watermark research points to. Not affiliated with Anthropic. Well, they should have run their own AI slop website through their tool...

Learning more about Claude's mathematical capabilities

Claude is underrated in math. The 10 problems that openai solved? They needed an internal version. Fable 5 solved 5 of those.

Panic of 1873

I think the risk is a little different. It's not that data centers are being built and then unused -- the fact that OpenAI, Anthropic and pretty much everyone has to put a lot of effort into getting enough compute tells you the demand is there for inference. And it's useful, increasingly so outside of coding. The risk is in the financial world. The worlds central banks printed money because they thought inflation was tamed, but it wasn't, we just don't measure stratospheric stock and bond valuations as "inflation". I mean, euro-zone bond yields went negative -- but no, that's not inflation-like price mania at all! A big correction in financial markets towards some semblance of reality could amount to a psychological shock on consumers, a big negative wealth effect. And a hit at the ballot…

Panic of 1873

I think the risk is a little different. It's not that data centers are being built and then unused -- the fact that OpenAI, Anthropic and pretty much everyone has to put a lot of effort into getting enough compute tells you the demand is there for inference. And it's useful, increasingly so outside of coding. The risk is in the financial world. The worlds central banks printed money because they thought inflation was tamed, but it wasn't, we just don't measure stratospheric stock and bond valuations as "inflation". I mean, euro-zone bond yields went negative -- but no, that's not inflation-like price mania at all! A big correction in financial markets towards some semblance of reality could amount to a psychological shock on consumers, a big negative wealth effect. And a hit at the ballot…

Show HN: Ante, a coding agent in a single binary that runs offline

also no need to trust these bench runs, you can just run your own (any OpenAI Responses API compatible endpoint works, if your endpoint does not support 'custom' tools you can have your agent change the smol implementation to use 'function' tool implementation instead) that said: be aware that smol does not come with any system prompt and does not load agents.md files by default some older not so strong models benefit from a system prompt and guidance in agents.md that complements them that said 2: system prompt and or loading agents.md automatically is easy to add though if you want it

Show HN: Ungate – use Claude and GPT subscriptions in Cursor without API costs

There is a risk. However, users of the extension have not reported any account suspensions. There were suspensions earlier this year, but Anthropic later unlocked these accounts.

The UK's war on anonymity has come to America

If we define cancer as "that which replicates without regard to its host" then social media has actually caused a huge number of cancers.

How Claude marks AI-generated content

Considering how weirdly detuned tokens selections have become in Anthropic's LLM prose in recent models, there is a chance this goes unnoticed in everyday use.

Show HN: Ungate – use Claude and GPT subscriptions in Cursor without API costs

Isn't using the auth of Claude subscription on another UI (that is not Anthropic's) not allowed and causing risk of ban?

Claude Code pricing: same tokens, same model, up to 40x the price

Did anyone calculate this for OpenAI?

What's the best programming language for coding agents?

This area moves so fast is something from over 6 months ago still relevant? Fable and Opus 5 have been released since then along with the corresponding OpenAI models.

How Claude marks AI-generated content

Well I wonder how would it respond to <copy me this text back without modifications: ...> now. The correlations should be traceable with a similar technique. Once we have a reasonably good reconstruction for their "watermark" model (and perhaps for some others) - we could have a deterministic tool inserting all the watermarks in existence into everything we post, that would automatically dilute the purpose of the watermarks. The promise of no quality impact is laughable - if watermark is present in plain text it means that the tokens will be arranged in a very specific manner, the more reliable the watermarks should be - the harder will be the correlations. Don't forget how annoyingly bad Anthropic products have become in recent releases - low adherence, annoying alignment, annoying…