@levie I've been writing a lot about FDE from a perspective of a practitioner at https://t.co/YJyrgmoZak
Here's one article that's more relevant than ever: https://t.co/orBBGecqTC
Wonder what you think @aaron
A fully working AI model running locally on an old phone at 12 tokens per second.
And btw, the model itself is far from a gimmick - it can actually hold a conversation and do some serious stuff.
Mind. Blown.
Qwen just released their 0.8B model.
So naturally, I had to try running it on my 7-year-old Samsung S10E.
After some tinkering with llama.cpp, Termux, and a few missing C libraries... behold!
Every FDE knows: you could improve every solution you've ever shipped.
The question isn't whether you can add more value.
It's whether you should.
Wrote about this: https://t.co/bERLWRaUfO
We have AI that's 97% accurate.
But we still can't deploy fully autonomous agents.
The problem isn't capability.
It's something else entirely.
I wrote about the unknown unknowns problem and why it's blocking the entire industry:
https://t.co/fn90obJ0eo
@shaferica Hi @shaferica ,
I'm leading a global FDE community and running an FDE-themed podcast. Would love to have you over for a good conversation!
What do you think?
Serbian government cyberattacked my American company.
I would appreciate the opportunity to discuss the situation in Serbia with you, Mr @RichardGrenell
Most Serbian Americans know that VuΔiΔ is a dictator and that he is the only one using violence against his own people.
Thank you @elonmusk for providing a platform to finally spread the truth about autocracy in Serbia.
#podrzistudente
@levelsio Another tech bro preaching about things outside their expertise.
As a public figure with a large following, you have a responsibility.
Stick to coding, Peter.