lifecodes

4 ms·

1.
The fast it goes the slow it feels
3 points by lifecodes 4mo ago|0 comments
2.
by lifecodes 6mo ago
hmm you are right, I too wish the same brother
3.
by lifecodes 6mo ago
the CoT bug where 8% of training runs could see the model's own scratchpad is the scariest part to me. and of course it had to be in the agentic tasks, exactly where you need to trust what the model is "thinking" the sandwich
4.
by lifecodes 6mo ago
Hey there I read your article, it was good.. But the first time I saw the title, I literally felt like you have copied me.. But ofc, you have your original thing.. Check my this article : https://blog.eshanstudio.com/posts&#
5.
by lifecodes 6mo ago
MANAGED AGENTS sounds like progress, but also like we’re standardizing around the current limitations instead of solving them.
6.
by lifecodes 6mo ago
I guess it would be much cheaper to attach an api version of everything we developed till now than teaching these ai to be able to control things in real world as humans do.. I mean if we see the cost of training then making more apis for e
7.
by lifecodes 6mo ago
I guess we are reaching the point where “10T parsmeters” sounds more like a marketing number than a meaningful metric. Between moE, aggressive quantization, and synthetic data pipelines, it’s getting harder to tell whether bigger models are
8.
by lifecodes 6mo ago
If this holds, does it unlock 100B+ models running locally in ~tens of GB RAM? Or does accuracy collapse before that point?
9.
Mathematical Psychology: Where Numbers Meet Neurons(neurolaunch.com)
1 points by lifecodes 6mo ago|0 comments