←back to thread

OpenAI delays launch of open-weight model

(twitter.com)
171 points martinald | 1 comments | 12 Jul 25 01:07 UTC | HN request time: 1.016s | source
Show context
mystraline ◴[12 Jul 25 01:46 UTC] No.44538610[source]▶
To be completely and utterly fair, I trust Deepseek and Qwen (Alibaba) more than American AI companies.

American AI companies have shown they are money and compute eaters, and massively so at that. Billions later, and well, not much to show.

But Deepseek cost $5M to develop, and made multiple novel ways to train.

Oh, and their models and code are all FLOSS. The US companies are closed. Basically, the US ai companies are too busy treating each other as vultures.

replies(8): >>44538670 #>>44538694 #>>44538700 #>>44538816 #>>44538905 #>>44539727 #>>44540309 #>>44540945 #
ryao ◴[12 Jul 25 02:01 UTC] No.44538670[source]▶
Wasn’t that figure just the cost of the GPUs and nothing else?
replies(3): >>44538699 #>>44538709 #>>44538740 #
1. 3eb7988a1663 ◴[12 Jul 25 02:15 UTC] No.44538740[source]▶
That is also just the final production run. How many experimental runs were performed before starting the final batch? It could be some ratio like 10 hours of research to every one hour of final training.