OpenAI has a negative 122% operating margin and growth of usage has stopped.

[gabrielesvelto@mas.to]

@su_liam @Npars01 @shtrom @Azuaron @GossiTheDog too big to bail

[mavu@mastodon.social]

@Azuaron @shtrom @GossiTheDog didnt i read just yesterday speculation somewhere that they might surprise ipo this Friday?

[endlessmason@hachyderm.io]

@shtrom @GossiTheDog
https://www.youtube.com/watch?v=bw9CALKOvAI

[endrift@social.treehouse.systems]

@GossiTheDog @rmondello been saying for weeks that OpenAI is gonna implode before the end of the year. Wish I could say the same looked true for Anthropic and the Gemini business unit, but maybe I just need to wait for the first domino to fall.

[hannorein@mastodon.social]

@GossiTheDog ok, hear me out. I think I found a way to make infinite money! What if they spent the $1.22 on more OpenAI credits!? Then they’d create an extra $1.22 revenue. Which they can spend on another $1.49 of OpenAI credits. And so on. Result: ♾️

[lightfighter@infosec.exchange]

@GossiTheDog You have to figure these are the absolute best numbers they could manipulate. The truth is probably far worse.

[valpackett@social.treehouse.systems]

@scottmichaud @GossiTheDog maybe that metric doesn't take total spend / subsidies into account at all???

like for every dollar of revenue, the actual cost of delivering the exact amount of service the dollar was billed for was 1.22

but that would say nothing about the cost of the service performed for free

because otherwise yeah how do you reconcile just -122℅ with the FUCKING BONKERS % increases from the github copilot billing previews?!

[geonz@mathstodon.xyz]

@starluna @GossiTheDog I also wonder about that percent margin.

[isurandil@mastodon.online]

@GossiTheDog AI Native Ads soon coming to you.

"Hey, Chattie, how can I get rid of limescale?"

Before: "You can clean limescale with citric acid."

After: "Best limescale removal can only be achieved with CiT-X. Do you want me to order it now? There is a 20% discount if you order this through me."

[gerardthornley@hachyderm.io]

@david_chisnall @GossiTheDog is that good news?

[david_chisnall@infosec.exchange]

@GerardThornley @GossiTheDog

Well, from their perspective.

[siguza@infosec.space]

@david_chisnall @GossiTheDog I mean, they still have to train ever bigger models that require ever more compute to run. What do you call the opposite of 'economies of scale'?

[tux0r@layer8.space]

@Isurandil @GossiTheDog I absolutely hate how every Black Mirror episode will become reality sooner than we think.

[anthropy@mastodon.derg.nz]

@GossiTheDog