codex hallucinates so bad. created a separate .py file to handle multiple xmls as source of annotations to the dataset and hallucinated the fact that it was present in the original cloned repo. so annoying istg.
I spent 48 hours with the Kimi K3 modeling code.
It took:
- 650 mg of caffeine (mandatory)
- 40 cans of LaCroix (optional... world record (?))
- 8 papers
- 6 months off my lifespan
Finally grokked the entire lineage of Kimi K3 and how we got here... every single step, since 2019 GPT-2
ghost rider - directed by shawn levy. i hope we dont get the same vibes that we got from deadpool vs wolverine in this movie as well. the casting is too perfect tho.
I built an AI App which does high quality professional edits for your talking head videos in one click. It's called PunchEdit.
If you want to level up your content game too. Comment "beta" and I'll dm you for early access.
i think i should go about revising and going through classical ml followed by ml systems. been more than 4 years now since i had a really in-depth study session of these topics.