3 ms·
The detail that rightsholders seem to be demanding a revenue share is interesting. That sounds administratively and technologically very complex to implement an
by kg 1y ago
The detail that rightsholders seem to be demanding a revenue share is interesting. That sounds administratively and technologically very complex to implement and probably also just plain expensive to implement.
- minimaxir 1y agoSam says Sora 2 has to make money but there is no revenue model that can feasibly offset a $4-5 compute cost per video.
- nojs 1y agoWhere did you get that figure from?
- minimaxir 1y agoIt’s more a ballpark since exact numbers vary and OpenAI could be employing shenanigans to cut costs, but in comparison, Veo 3 which has similar quality 720p video costs $0.40/second for the user, and Sora’s videos are 10 seconds each. Although Veo 3 could cost more or less to Google than what is charged to the user. I suspect OpenAI’s costs to be higher if anything since Google’s infra is more cost-efficient.
- majewsky 1y agoSora 1 is slightly more expensive than Veo 3, at $0.50/second: https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/ https://azure.microsoft.com/en-us/pricing/details/cognitive-... (select "US East 2" as region; not all regions have Sora). Though again, caveat: Those are API prices. The real prices could be both lower (if there is a profit margin) or higher (if they are subsidized).
- nvr219 1y agoIt was revealed to them in a dream.
- dwohnitmok 1y agoWith some back of the napkin math, I am pretty sure you're off by at least two orders of magnitude, conceivably 4. I think 2 cents per video is an upper limit. https://news.ycombinator.com/item?id=45434298 https://news.ycombinator.com/item?id=45434298 Generally speaking, API costs that the consumer sees are way higher than compute costs that the provider pays. EDIT: Upper limit on pure on-going compute cost. If you factor in chip capital costs as another commentator on the other thread pointed out, you might add another order of magnitude.
- rafram 1y agoYou also aren’t including amortized training costs, which are immense (and likely ongoing as they continue to tweak the model).
- dwohnitmok 1y agoI suspect amortized training costs are only a relatively small fraction of the amortized hardware costs (i.e. counting amortized hardware costs already accounts for the large fraction of the cost of training and pulling out training as a completely separate category double counts a lot of the cost).
- martin-t 1y agoThis is how all work should be rewarded. Workers getting paid a flat rate while owners are raking in the entire income generated by the work is how the rich get richer faster than any working person can.
- pxoe 1y agoThis "but it's too hard to implement" excuse never made sense to me. So it's doable to make a system like this, to have smart people working on it, hire and poach other smart people, to have payments systems, tracking systems, personal data collection, request filtering and content awareness, all that jazz, but somehow all of that grinds to a halt the moment a question like this arises? and it's been a problem for years, yet some of the smartest people are just unable to approach it, let alone solve it? Does it not seem idiotic to see them serve 'most advanced' products over and over, and then pretend like this question is "too complex" for them to solve? Shouldn't they be smart enough to rise up to that level of "complexity" anyway? Seems more like selective, intentional ignoring of the problem to me. It's just because if they start to pay up, everyone will want to get paid, and paying other people is something that companies like this systematically try to avoid as much as possible.