Search 40M docs in <200ms on a CPU using binary embeddings and int8 rescoring m medium.com 2 akkishore 9 months ago 0 comments
Every AI startup is likely to be crushed by rapid expansion of model providers t twitter.com 2 akkishore 1 year ago 1 comments
Hugging Face Offers Developers Inference-as-a-Service Powered by Nvidia Nim b blogs.nvidia.com 2 akkishore 2 years ago 0 comments
Gartner Predicts 30% of Gen AI Projects Will Be Abandoned After POC by 2025 g gartner.com 2 akkishore 2 years ago 1 comments
Ask HN: Is there a realtime/lagged view of worldwide internet traffic? 13 akkishore 5 years ago 4 comments
A Styrofoam Ball Rides the Waves to Create a Volumetric Display h hackaday.com 2 akkishore 5 years ago 0 comments
Show HN: MultiSSH – Control multiple servers over SSH in parallel m multissh.dev 1 akkishore 5 years ago 0 comments
MultiSSH: Connect to multiple SSH hosts simultaneously m multissh.dev 3 akkishore 5 years ago 0 comments
Huge fan of all the work you do. Wanted to understand something fundamental and whom better to ask than you: Whats so special about the transformer architecture that its able to predict the next token so beautifully understanding all the intricate previous token relationships? I understand Attention but what so special about this architecture that no other architectures are able to "attend" appropriately to previous tokens? Being a CS guy, its really hard for me to fathom that we have not yet created another architecture which can perform similarly.