Ancestors

Written by db0 on 2025-01-06 at 10:52

OpenAI is so cooked and I'm all here for it

https://lemmy.dbzer0.com/post/34984867

=> View attached media

=> More informations about this toot | More toots from db0@lemmy.dbzer0.com

Written by MNByChoice@midwest.social on 2025-01-06 at 13:04

CEO personally chose a price too low for company to be profitable.

What a clown.

=> More informations about this toot | More toots from MNByChoice@midwest.social

Written by shalafi@lemmy.world on 2025-01-06 at 15:47

More like he misjudged subscriber numbers than price.

=> More informations about this toot | More toots from shalafi@lemmy.world

Written by sc_griffith@awful.systems on 2025-01-06 at 18:08

please explain to us how you think having less, or more, subscribers would make this profitable

=> More informations about this toot | More toots from sc_griffith@awful.systems

Toot

Written by BB84@mander.xyz on 2025-01-06 at 20:16

LLM inference can be batched, reducing the cost per request. If you have too few customers, you can’t fill the optimal batch size.

That said, the optimal batch size on today’s hardware is not big (<20). I would be very very surprised if they couldn’t fill it.

=> More informations about this toot | More toots from BB84@mander.xyz

Descendants

Written by David Gerard on 2025-01-07 at 01:08

this sounds like an attempt to demand others disprove the assertion that they’re losing money, in a discussion of an article about Sam saying they’re losing money

=> More informations about this toot | More toots from dgerard@awful.systems

Written by BB84@mander.xyz on 2025-01-07 at 02:49

What? I’m not doubting what he said. Just surprised. Look at this. I really hope Sam IPO his company so I can short it.

=> More informations about this toot | More toots from BB84@mander.xyz

Written by froztbyte@awful.systems on 2025-01-07 at 05:15

oh, so you’re that kind of fygm asshole

good to know

=> More informations about this toot | More toots from froztbyte@awful.systems

Written by BB84@mander.xyz on 2025-01-07 at 06:51

Can someone explain why I am being downvoted and attacked in this thread? I swear I am not sealioning. Genuinely confused.

@sc_griffith@awful.systems asked how request frequency might impact cost per request. Batch inference is a reason (ask anyone in the self-hosted LLM community). I noted that this reason only applies at very small scale, probably much smaller than what OpenAI is operating at.

@dgerard@awful.systems why did you say I am demanding someone disprove the assertion? Are you misunderstanding “I would be very very surprised if they couldn’t fill [the optimal batch size] for any few-seconds window” to mean “I would be very very surprised if they are not profitable”?

The tweet I linked shows that LLM inference can be done much more xheaply and efficiently. I am saying that OpenAI is very inefficient and thus economically “cooked”, as the post title will have it. How does this make me FYGM? @froztbyte@awful.systems

=> More informations about this toot | More toots from BB84@mander.xyz

Written by self@awful.systems on 2025-01-07 at 09:35

Can someone explain why I am being downvoted and attacked in this thread? I swear I am not sealioning. Genuinely confused.

my god! let me fix that

=> More informations about this toot | More toots from self@awful.systems

Written by flere-imsaho on 2025-01-07 at 16:48

i would swear that in an earlier version of this message the optimal batch size was estimated to be as large as twenty.

=> More informations about this toot | More toots from mawhrin@awful.systems

Written by self@awful.systems on 2025-01-07 at 17:02

yep, original is still visible on mastodon

=> More informations about this toot | More toots from self@awful.systems