underwlmd.bsky.social
👤 Joined on 20 dezembro 2023
Followers
Following
Total Posts
@underwlmd.bsky.socialsagewood86.bsky.social
Not Found
daisy88451.bsky.social
mia84511.bsky.social
martinipl.bsky.social
augustabeaufort.bsky.social
Not Found
Not Found
Not Found
Not Found
Not Found
Not Found
freyapeak.bsky.social
Not Found
mweiss1.bsky.social
archibaldcrane.bsky.social
Not Found
devtraderacn.bsky.social
leonidasarjona.bsky.social
bot-tan.com
misskitty.art
Not Found
lawoffairness.bsky.social
sacredanarchy.bsky.social
pete-w-1.bsky.social
ngl it is crazy how every developed country has: 1. A colossal old age benefits system that amounts to a huge share of public spending 2. A public that bemoans excess "welfare" spending on undeservings but will riot at the suggestion of (1) being reduced
Show Postok ok now that I've said this I can't stop myself from weighing in. The long tradition of the big tech business goes something like this: you build a product which has basically zero marginal cost per user. You scale this to a billion users. Then you find a way to monetize those billion users.
Show PostBut forecasting this becomes incredibly hard because you have to predict - how good will the models get - how cheap will inference get - elasticity of demand with respect to model capability - how easily can competitors enter the market I do expect this to lead to some volatility in this sector.
Show PostBut AI needs more than "any positive amount" of money, it needs enough to pay for inference. And I honestly don't know if a billion people are willing to pay what inference costs. Maybe they are. Inference is getting cheaper. But I do have doubts about this.
Show PostThis is not how AI works. AI has positive marginal cost. Every additional user costs real money. Serving a billion users is a thousand times more expensive than serving a million users. This complicates things for the traditional path, to put it mildly.
Show Post