←back to thread

Ingesting PDFs and why Gemini 2.0 changes everything

(www.sergey.fyi)
1303 points serjester | 6 comments | 05 Feb 25 18:05 UTC | HN request time: 1.283s | source | bottom
Show context
xnx ◴[05 Feb 25 21:42 UTC] No.42955622[source]▶
Glad Gemini is getting some attention. Using it is like a superpower. There are so many discussions about ChatGTP, Claude, DeepSeek, Llama, etc. that don't even mention Gemini.
replies(3): >>42955696 #>>42955982 #>>42956190 #
1. throwaway314155 ◴[05 Feb 25 21:47 UTC] No.42955696[source]▶
Google had a pretty rough start compared to ChatGPT, Claude. I suspect that left a bad taste in many people's mouths. In particular because evaluating so many LLM's is a lot of effort on its own.

Llama and DeepSeek are no-brainers; the weights are public.

replies(1): >>42955879 #
2. beastman82 ◴[05 Feb 25 22:00 UTC] No.42955879[source]▶
No brainer if you're sitting on a >$100k inference server.
replies(2): >>42956391 #>>42969919 #
3. throwaway314155 ◴[05 Feb 25 22:40 UTC] No.42956391[source]▶
Sure, that's fair. If you're aiming for state of the art performance. Otherwise, you can get close and do it on reasonably priced hardware by using smaller distilled and/or quantized variants of llama/r1.

Really though I just meant "it's a no-brainer that they are popular here on HN".

4. BoorishBears ◴[07 Feb 25 06:07 UTC] No.42969919[source]▶
I pay 78 cents an hour to host Llama.
replies(1): >>42982334 #
5. beastman82 ◴[08 Feb 25 11:53 UTC] No.42982334{3}[source]▶
Vast? Specs?
replies(1): >>42988133 #
6. BoorishBears ◴[09 Feb 25 03:13 UTC] No.42988133{4}[source]▶
Runpod, 2xA40.

Not sure why you think buying an entire inference server is a necessity to run these models.