I love minimalistic AI integration. Guiding Agentic AI with only one `https://t.co/7tdiSimbkq` and the `__init__.py` of your project is bliss. It is by far my best aiptimisation
@trq212 I find that doing this directly to a github issue (where you can track this) is infinitely better. It stays there forever, perfectly linked to what was being done.
My second best is a post-work prompt: "where did you take shortcuts? tell me where you cheated"
@Prince_Canuma Look up GL inet comet. It has native Tailscale integration; your machine(s) will never be out of reach again. It lets me change my bios settings in my desktop in Portugal all the way from Italy
ML researchers just built a new ensemble technique.
It even outperforms XGBoost, CatBoost, and LightGBM.
Here's a complete breakdown (explained visually):
@armankhon@msnitish What are my chances I can intern with you next year coming from EU? I just pressed accept on my PhD programme in DS / AI, so I'd intern for the summer, with possibility of extending one semester. I'm mainly python/PyTorch but open to learn your recommendations
@ylecun @ABdesignhelix @elonmusk Iβd love some advice specially if harsh! Started AI with two meta scholarships on Udacity, got my MSc thesis published, applying to phds now, but feel that FAIR is still moon-distance. Number one thing to boost visibility on application? I moved to Paris, maybe local meetups?
@karpathy Have you taken a look at tiny grad? From what you described in the repo about clean and simple implementations, it's a match made in heaven I reckon
cc @realGeorgeHotz