He was vibe-coding in production. Am I reading that right? Sounds like an intern-level mistake.

Raw-dogging the internet without an adblocker is about as irresponsible as not using contraception

Meta is a security threat outside of Russia too. No privacy should be expected when using their services

They’re mad they can’t use cellbrite to snoop on properly configured GOS phones and that they actually have to do real police work to catch drug dealers

Y’all remember that #DeathTo Black Mirror episode?

I still don’t understand what Anthropic is trying to achieve with all of these stunts showing that their LLMs go off the rails so easily. Is it for gullible investors? Why would a consumer want to give them money for something so unreliable?

It’s called slopsquatting for hallucinated libraries that vibe coders use, but I figure it can be called the same for hallucinated URLs

Are we gonna see Netflix logos on launch vehicles soon?

No telling what else they do with your pictures once they’re on Meta’s servers. Literal spyware (as if it hasn’t already been for a long time).

Imagine having your nudes accidentally ghiblified, with the originals stored indefinitely on Meta’s servers.

Long enough for GOS to port Android 16 or get their own OEM, I hope

Laughs in GrapheneOS

Lol. The bots to humans ratio will get even bigger if that happens.

Oh look, another centralized social media platform that will eventually get enshittified

LLMs have no more beliefs than a parrot does. They just repeat whatever opinions/biases exist in their training data. Although, that’s not too different from humans in some respects.

Check out PostmarketOS for older machines/phones too. I got it working on an old Microsoft surface RT that would have been e-waste otherwise

Neat, but one needs to also trust that google isn’t an adversary

Extraordinary claims require extraordinary evidence.

So if I have two machines running the same local LLM and I pass a prompt between them, I’ve achieved data compression by transmitting the prompt rather than the LLM’s expected response to the prompt? That’s what I’m understanding from the article.

Neat idea, but what if you want to transmit some information that an LLM can’t tokenize and generate accurately?

Wow, so brave

As a German company, they should know a thing or two about appeasing facists