vsreekanti

40 ms·

1.
Open models in perpetual catch-up(interconnects.ai)
3 points by vsreekanti 8mo ago|0 comments
2.
Throw more AI at your problems(frontierai.substack.com)
68 points by vsreekanti 2y ago|55 comments
3.
Deep AI Work(frontierai.substack.com)
2 points by vsreekanti 2y ago|0 comments
4.
Customers want more AI, not better AI(frontierai.substack.com)
4 points by vsreekanti 2y ago|0 comments
5.
There's no silver bullet in AI(generatingconversation.substack.com)
2 points by vsreekanti 2y ago|0 comments
6.
Llama 3.1 405B, Meta's AI strategy, and the new, open frontier model ecosystem(interconnects.ai)
2 points by vsreekanti 2y ago|0 comments
7.
The bull case for AI startups(generatingconversation.substack.com)
4 points by vsreekanti 2y ago|0 comments
8.
AI is breadth-first, not depth-first(generatingconversation.substack.com)
3 points by vsreekanti 2y ago|0 comments
9.
Open LLMs don't need to beat OpenAI(generatingconversation.substack.com)
3 points by vsreekanti 2y ago|0 comments
10.
AI must be more than a checkbox(generatingconversation.substack.com)
4 points by vsreekanti 2y ago|0 comments
11.
How to talk to someone who doesn't trust AI(generatingconversation.substack.com)
5 points by vsreekanti 2y ago|6 comments
12.
Lessons from customer evaluations of an AI product(generatingconversation.substack.com)
12 points by vsreekanti 2y ago|0 comments
13.
The end of the "best open LLM"(interconnects.ai)
5 points by vsreekanti 2y ago|0 comments
14.
You can't build a moat with AI(generatingconversation.substack.com)
40 points by vsreekanti 2y ago|45 comments
15.
Untangling concerns about consolidation in AI(generatingconversation.substack.com)
2 points by vsreekanti 3y ago|0 comments
16.
Don't build your applications on a single LLM call(generatingconversation.substack.com)
5 points by vsreekanti 3y ago|0 comments
17.
Building on quicksand – the state of modern AI app development(generatingconversation.substack.com)
3 points by vsreekanti 3y ago|0 comments
18.
The koan of an open-source LLM(interconnects.ai)
2 points by vsreekanti 3y ago|0 comments
19.
LLMs are the variable interest credit card of tech debt(generatingconversation.substack.com)
4 points by vsreekanti 3y ago|0 comments
20.
An Introduction to Evaluating LLMs(generatingconversation.substack.com)
4 points by vsreekanti 3y ago|0 comments
21.
Prospecting in the AI Gold Rush(generatingconversation.substack.com)
3 points by vsreekanti 3y ago|0 comments
22.
Multimodal LM roundup: Unified IO 2, I/O, Gemini, LLaVA-RLHF, and RLHF questions(interconnects.ai)
3 points by vsreekanti 3y ago|0 comments
23.
OpenAI will solve its privacy problems before you(generatingconversation.substack.com)
3 points by vsreekanti 3y ago|0 comments
24.
Building a business in Generative AI [video](youtube.com)
8 points by vsreekanti 3y ago|0 comments
25.
Why open-source LLMs matter(generatingconversation.substack.com)
7 points by vsreekanti 3y ago|0 comments
26.
A deep dive on MemGPT – memory management for LLMs(youtube.com)
10 points by vsreekanti 3y ago|0 comments
27.
by vsreekanti 3y ago
I completely agree — open-source models and custom deployments just can't compete with the cost and efficiency here. The only exception here is if open-source models can get way smaller and faster than they are now while maintaining
28.
by vsreekanti 3y ago
Probably some combination of all the above! I think 1 and 2 are interlinked though — the cheaper they can be, the more they build that moat. They might be eating the cost on these APIs too, but unlike the Uber/Lyft war, it'll be w
29.
by vsreekanti 3y ago
Yep, we agree that the obvious direction of innovation for OSS models is smaller and cheaper, likely at roughly the same quality: https://generatingconversation.substack.com/p/open-source-ll...
30.
Open-source LLMs shouldn't try to win(generatingconversation.substack.com)
6 points by vsreekanti 3y ago|0 comments

More ›