Pennomi

Sounds like it’s time to leak a bunch of politician nudes to show them why encryption matters.

Okay but these models are actually pretty interesting when you dive in. Useful yet? Not by a long shot (I’m guessing decades, not years of work). But interesting nonetheless.

You know who wouldn’t have missed a word there? ChatGPT.

(Not saying it would’ve written a better article but AI is typically good at not making silly typos like humans.)

And let’s be honest, a certain percentage of junior devs never do learn. That’s always been the case, AI or not.

The open paper they published details the algorithms and techniques used to train it, and it’s been replicated by researchers already.

It’s trivial compared to the compute they dedicate to AI models. Like, not even a rounding error.

The hell? There’s no reason to use plain HTTP instead of HTTPS.

And symmetric encryption is wildly irresponsible as well.

I got bad news for you… htmx is written in JavaScript.

I did not know that. There’s a bunch of news articles going around claiming that even the creators of the models don’t understand them and that they are some sort of unfathomable magic black box. I assumed you were propagating that myth, but I was clearly mistaken.

No one understands how LLMs work, not even on a basic level.

Well that’s just false.