It turns out the most dangerous thing is competition.

Margin compression is terrifying

I’ve so far been successful at getting Fable to find security issues, but I’m careful to not prompt it too directly. I point it at my server code and tell it to find general issues, which has so far resulted in discovering a few minor bugs that Opus has never raised under similar conditions.

I doubt that. Why would you train Mythos on its own code if you don't want it to be able to reproduce it? It's not going to add much to the overall corpus.

Synthetic training data has been the name of the game since years ago.

I know a security researcher at Google with access to Mythos. He says it's the "real deal" and that "there are career plans I had that are no longer viable".

“trust me bro”

He could be incredibly naive. We'll all find out with time.

That's kind of like my T430: https://danangell.com/blog/posts/my-t430/

I wish it could go up to 32GB though. And the i7 in there shows its age. But it's still usable with modern software!

I had a T420s which is a bit smaller than the T430 and I really loved this thing. Did my whole scholarship with it, even bought a second hand to have as a fail-safe. Really nice machine for its time

Unless Stonehenge is your home

I’m also staying away. I need to not lose money more than I need to gain money.

Similar to you, the things I have truly vibe-coded (having looked at <5% of the code) are largely data focused. Data labeling, organization, etc. These applications are extremely janky, I'd never ship them to users. The UI is mediocre at best. The functionality hardly better. But the point is to get data out of them. The code is a means to an end and not a product in itself. Building a custom dataset builder in just a couple hours of work is really powerful.

I feel similarly. Take good photos of your friends doing cool things. Absolutely do not stop everyone for a group picture. Forget things. It’s okay to forget the minutia of life.