Viewing profile — stingraycharles

Ok so this is a bit of a side note, but when reading this, did anyone else have the feeling that, for all their messaging around “we are so afraid that our models will be used for …

It is a systemic issue, though, when the system enables these certain actors to remain in position and cause the community to break apart like it’s doing.

“Our experience is that the Steering Committee as an institution lacks a native instinct for the delegation envisioned by the constitution, while also not being sufficiently engage…

That’s very much up for debate.

It’s temporary revenue, it’s not as if those businesses are suddenly fundamentally better. To me it seems like they’re just riding the AI hype out, expecting a crash after, and not…

What’s wrong with a signature?

You’re right, but it seems like the punishments are not really about that, but just a general “protect the children” and age verification system. > Instead, the court ordered Meta …

I think Apple much rather tells users that this is exactly what they warned them about when they were forced to support USB-C and will advise users to only buy official cables from…

This really needs to have a lot more evidence that it’s because GitHub’s code is being written by AI vs them being bombarded by activity from all kinds of AI agents all over the wo…

Didn’t Anthropic acquire Cerebras? Seems like a move into the same direction. I also think that etching models into ASICs may be a bit too inflexible for what OpenAI and Anthropic …

It's really not designed as much of a security model as a liability model. When the AI isn't responsible for the decisions, a whole category of liability goes away.

Isn’t another aspect of it that it’s sufficiently random / unrecognizable, for example? I’m very much aware of the differences between symmetric and asymmetric encryption, and real…

Yeah, I wouldn’t say with certainty that LLMs will never break any symmetrical crypto algorithm. It will certainly require a lot of effort, but so does solving some hard math chall…

Seems pretty clear to me: enable it for the projects you want, and there’s a 1M / 10M token limit per day, depending on the model you use. Assuming an average context size of 100k …

> JavaScript was a mistake, e-commerce was a mistake, and so is everything else that came after. That’s a really extreme take that lacks any nuance. Might as well blame the OS as w…

This sounds like FUD unless backed by evidence. Cloudflare’s inference absolutely does support ZDR, as long as you use unified billing (ie not using BYOK).

Isn’t that already in detail by the research of these quantization techniques?

What is this kind of unprofessional blog-style rant doing on OpenAI’s website? Shouldn’t this be settled between the two companies themselves rather than trying to play the court o…

I think the problem is that the Box / Map / any stuff and all the explicit declarations that Go requires makes it harder. In Haskell as well, you can let the compiler infer a lot o…

my post was never intended as Go bashing, I use Go a lot and appreciate its simplicity. am I tearing something down in my comment?

Am I the only one who’s absolutely shocked that Go finally is embracing generics? Does anyone have a bit of an inside view into what changed in the perspectives of the language mai…

One example from a root cause analysis of a security incident like this is not evidence for “everyone does this”. I believe that Huggingface, with 250+ employees, is an outlier.

Just don’t switch models every week and you’ll be fine. Especially in production.

I don’t think everyone does this at all. Maybe very small scale companies, but I would have expected a company like Huggingface to have better security standards. As soon as you’re…