Testing CUDA kernel execution with GPU correctness harnessgithub.com/IshwaryaRavichandran 3laxmena22ddiscuss
GateGPT: 56k tokens per second Transformer (KV cache) on FPGA at 80 MHztwitter.com 40laxmena2mo13 comments
Karpathy on Why 'Edutainment' Isn't Real Learning (Seek the Sweat) (2024)twitter.com 4laxmena2mo1 comment
Toward Generalist Autonomous Research via Hypothesis-Tree Refinementhuggingface.co 2laxmena3modiscuss
Embedding models are coordinate systems. What silently breaks in production RAGinternals.laxmena.com 3laxmena3modiscuss
Dashboard tracking the 100-day promises of the Nepali governmentdigitalpratipakshya.com 3laxmena4modiscuss
Ask HN: When did you move from AI agentic loops to simpler deterministic system? 8laxmena4mo7 comments
Retail Investors Keep Getting Burned in the Stock Market.They're Piling in Againbarrons.com 2laxmena4mo1 comment
AI Skills as loader spec, not prompts – why the architecture changes everythinginternals.laxmena.com 5laxmena4mo1 comment
IBM, AWS Join to Help Oil and Gas Industry Accelerate Energy Transitionfinance.yahoo.com 3laxmena4ydiscuss
Research reveals Air Pollution Can Cause Brain Cancer – DailyTech Bulletinyoutube.com 1laxmena6ydiscuss
Bill and Melinda Gates-funded invention could solve global malnutrition crisisinstagram.com 2laxmena6ydiscuss