It’s far from useless. Overblown, for sure.

They made an awesome hammer that can solve many hammer-related tasks, but they’re selling it like it can also cook, drive and keep the house clean.

Blame the companies, not the math.

It’s not like that. Tokens are an inherent computational property of how a model calculates the probabilities and such to generate text.

Having said that, what a token means in terms of computation varies wildly between models and is not directly comparable. So attributing a money value to tokens in general, independently of the model, is weird by nature.

And even within a model, the number of tokens needed to generate a response is very variable too, depending of the model itself and the parameters with which it has been configured (thinking mode, temperature, etc.).

So yeah, companies can pretty much set any price they want and there’s not much anyone can do about it.

Claude is trained on stolen data (the whole Internet), so I can’t have any sympathy for Anthropic when someone steals from them.

Given that the LLMs could follow the short lists of words well but not the longer lists, and that they were processing images, not text, I think it’s more likely that their context just filled up and they forgot the original instructions (or they were assigned a lower weight in the computation).

For many years I have thought that elementary schools should be free of any modern electronic tech. In middle school it should be introduced but not connected to internet (word processors, interactive encyclopedias or coding like we used to learn it in the '80s or '90s, BASIC, Logo, that kind of stuff).

Finally, introduce internet in high school, in a controlled way, focusing primarily on knowledge and research sites. No LLMs included.

Of course you can’t control what happens in the students’ homes, but that should not affect what happens at school, specially if you replace homework with schoolwork, which I also think it’s better in the long run.

Plain wireguard. Or maybe Pangolin?

https://github.com/fosrl/pangolin

A couple of days ago I finished reading “Careless People” by Sarah Wynn-Williams. This book really shows how awful the people at the top of Facebook are (it probably applies to other similar companies as well). And it’s not that they want to be evil, it’s more like they are playing a game of fame, pride and power and don’t care about the consequences. It made me recall the quote about “the banality of evil”.

The funny thing is, being a developer, I was talking to someone the other day about how in my opinion, middle managers were way easier to replace with AI than developers (less domain knowledge, less creative work, etc.). And basically, at least according to the article, that’s the people that Cloudflare fired.

Not that anyone should be replaced by AI, but it’s ironic that many of the people who are shoving AI down our throats are the easiest to replace in an organization.

Stalin suing Hitler. I’m rooting for both of them to lose.

That doesn’t solve the problem: it only disables the duck.ai assistant, but doesn’t disable the search results that are normal websites generated with AI (the real issue here). To be fair, that’s not DDG problem specifically but every search engine problem.

The web is now full of these fake websites, which is a real problem because on the search results they look legit and only when visiting them you realize it’s AI crap.

And the funny (or sad) thing is that current AIs are being trained on these fake hallucinating sites, so even they are suffering from false information, which in turn is given to humans and used to make more fake websites and… the result is up to your imagination.

Nvidia Riva TNT2 M64