Yea this may be my dumbest idea in this piece lol. Probably worth some followup from me what user psychology can tell us about long-term behavior modification.
One logical next step I don't see you discussing here is brands paying for fully customized ads in the service - videos of you using the products in context and loving them (or using a competitor and making your mom cry). That seems like a potential breakthrough in marketing and a nightmare for humans.
Very different use cases. Vibes is much more consumption oriented, Sora is more social/creative feeling. In a good world people will gravitate towards more active creativity but history has shown us otherwise
"Cheap enough" is an assumption. Just because it's in a product doesn't mean it's "cheap enough to be a real alternative to human creation". It's cheap enough for them to launch a product based on it, but it is almost certainly operating at a significant loss. It's a bet on a future revenue stream, but how good a bet is it? Or is OpenAI just throwing stuff at the wall? AI video gen will come down in cost with time, of course, but will that cost ever actually be less than the literal free cost of getting humans to do it for you and just dealing with the hosting costs? (i.e. TikTok, Instagram, YouTube now) I'm skeptical.
Also worth mentioning that OpenAI's track record is AFAIK literally *one* very sticky consumer-facing product success (ChatGPT). This is a consumer-oriented product, so I don't actually think their track record with *this* sort of thing is that convincing. Are there other OpenAI successes that make you feel like *this* kind of play has a higher than average likelihood of success?
Meanwhile making it so incredibly easy to burn GPU hours has got to be insanely bad for energy use, good lord. I have to be honest: I hate this.
Hey Oshyan! Cheap enough is very true for consumers. It is fast and easy (even if the output is subpar). For OpenAI every model they've ever made has come down in cost to run by 90% within a year of its release, so assuming their CFO is doing their job, I'm not worried about the product being cheap enough. The future revenue stream might be that this is just targeting data for ads (which they are already working on) or something paywalled if it grows too fast. We will see on how user behavior shapes up.
Yea I don't know if going for a social network was the right choice here, and it could be that they shut down the app and reintegrate the model into ChatGPT. You could be totally right there! (Though I did notice I opened Sora instead of Reels today ¯\_(ツ)_/¯ ). And I think ChatGPT''s origin was an accident, but its growth is deliberate execution. Best example if the Dalle moment last year. If anything their enterprise business has been their big weakness! Consumer they've crushed.
Agreed on energy use. Big problem everywhere and I expect backlash to mount over the next year.
And yea, I really get the hatred feeling! Many people have that reaction though I still consider this roughly the equivalent to social media
I can't imagine the impacts it can have towards less informed socities . Imagine a all out riot due to the widespread AI videos of the country's leader aggressive statements
Also, one thing I disagree with you is seeing videos of you achieving will have the opposite effect . It will cause less motivation for the society. For eg :- people won't post their fitness progress due to losing the actual exclusivity of doing it . That is a driving force for a lot of people
“Perhaps being able to generate videos of achieving their dreams will encourage people to go make them happen”
That seems… highly unlikely?
Yea this may be my dumbest idea in this piece lol. Probably worth some followup from me what user psychology can tell us about long-term behavior modification.
One logical next step I don't see you discussing here is brands paying for fully customized ads in the service - videos of you using the products in context and loving them (or using a competitor and making your mom cry). That seems like a potential breakthrough in marketing and a nightmare for humans.
I had a whole section on this that I cut! Whoops, shoulda kept it. I kept thinking about this scene from Minority Report https://www.youtube.com/watch?v=7bXJ_obaiYQ
the customization is going to be very weird
Do you think Sora does better than Vibes?
Can either rival Bytedance as an AI consumer app Maker?
Very different use cases. Vibes is much more consumption oriented, Sora is more social/creative feeling. In a good world people will gravitate towards more active creativity but history has shown us otherwise
Stratechery put it nicely as lean forwards vs. lean backwards experience
Yea that's right. And I'm not sure that we even want lean forward for our social media! we mostly just want to chill
Would love an invite, thank you 😊
Check your DMs!
Hey 👋
Check your DMs!
"Cheap enough" is an assumption. Just because it's in a product doesn't mean it's "cheap enough to be a real alternative to human creation". It's cheap enough for them to launch a product based on it, but it is almost certainly operating at a significant loss. It's a bet on a future revenue stream, but how good a bet is it? Or is OpenAI just throwing stuff at the wall? AI video gen will come down in cost with time, of course, but will that cost ever actually be less than the literal free cost of getting humans to do it for you and just dealing with the hosting costs? (i.e. TikTok, Instagram, YouTube now) I'm skeptical.
Also worth mentioning that OpenAI's track record is AFAIK literally *one* very sticky consumer-facing product success (ChatGPT). This is a consumer-oriented product, so I don't actually think their track record with *this* sort of thing is that convincing. Are there other OpenAI successes that make you feel like *this* kind of play has a higher than average likelihood of success?
Meanwhile making it so incredibly easy to burn GPU hours has got to be insanely bad for energy use, good lord. I have to be honest: I hate this.
Hey Oshyan! Cheap enough is very true for consumers. It is fast and easy (even if the output is subpar). For OpenAI every model they've ever made has come down in cost to run by 90% within a year of its release, so assuming their CFO is doing their job, I'm not worried about the product being cheap enough. The future revenue stream might be that this is just targeting data for ads (which they are already working on) or something paywalled if it grows too fast. We will see on how user behavior shapes up.
Yea I don't know if going for a social network was the right choice here, and it could be that they shut down the app and reintegrate the model into ChatGPT. You could be totally right there! (Though I did notice I opened Sora instead of Reels today ¯\_(ツ)_/¯ ). And I think ChatGPT''s origin was an accident, but its growth is deliberate execution. Best example if the Dalle moment last year. If anything their enterprise business has been their big weakness! Consumer they've crushed.
Agreed on energy use. Big problem everywhere and I expect backlash to mount over the next year.
And yea, I really get the hatred feeling! Many people have that reaction though I still consider this roughly the equivalent to social media
Not a paying subscriber but shooting my shot for an invite ;)
Thanks in any case!
I can't imagine the impacts it can have towards less informed socities . Imagine a all out riot due to the widespread AI videos of the country's leader aggressive statements
Also, one thing I disagree with you is seeing videos of you achieving will have the opposite effect . It will cause less motivation for the society. For eg :- people won't post their fitness progress due to losing the actual exclusivity of doing it . That is a driving force for a lot of people