In his first public deep dive into his current projects, John Carmack (the creator of Wolfenstein 3D, Doom, Quake, etc) shares a refreshingly honest and technical look at his research into artificial general intelligence. He discusses why Large Language Models (LLMs) are not the full answer, the pitfalls of over-optimizing benchmarks, and why Atari 260games, despite their age, remain one of the most valuable environments for reinforcement learning research.
---
This video is sponsored by VIAIM.
Get the Viaim RecDot Earbuds here: https://andauth.co/ViaimGE
For a 5% discount use the code SOUNDVIAIM (valid until July 20) From July 7th to July 19th, all VIAIM products are at least 20% off, and with the code, customers can get an additional 5% off — making it at least 25% off in total.
---
⭐ Please support my channel on Patreon!
Get early access to videos, members-only content, behind-the-scenes updates, and join the Gary Explains Discord! Join here ? https://www.patreon.com/GaryExplains ?
Twitter: https://twitter.com/garyexplains
Instagram: https://www.instagram.com/garyexplains/
#garyexplains