[HN Gopher] Meta's Open Source Llama Upsets the AI Horse Race
___________________________________________________________________
Meta's Open Source Llama Upsets the AI Horse Race
Author : SoftwareDev6
Score : 13 points
Date : 2023-08-28 19:08 UTC (3 hours ago)
(HTM) web link (www.wired.com)
(TXT) w3m dump (www.wired.com)
| ilaksh wrote:
| I don't know if they mention Code Llama but that has been a huge
| milestone. But from what I've heard it's still at least 15%
| behind GPT-4 for code. But open models are gaining ground.
|
| Although ChatGPT Enterprise kind of changes the benchmark again
| in away.
| Havoc wrote:
| >But from what I've heard it's still at least 15% behind GPT-4
| for code
|
| More. But you're comparing the performance of something that
| runs on a macbook vs something out of a datacenter full of
| H100s.
| ChatGTP wrote:
| I guess the idea here is that it's probably not wise to blindly
| invest $10BUSD on something which might be ubiquitous and free
| in 6 months.
| matteoraso wrote:
| I think this article is overstating things a bit. Chatgpt already
| lets you finetune the AI for your own use, so it's not like
| releasing LLaMa-2 as open source is going to seriously change
| things by a whole lot.
___________________________________________________________________
(page generated 2023-08-28 23:02 UTC)