Seems like AI is going to get more expensive over the next year. GPU prices are spiking a bit in the last two months, Claude users are complaining about usage limits more than usual, Mythos is supposed to be more expensive, Anthropic revenue has been growing at an insane pace (10x a year for more than two years) and seems like even as it slows down it’ll still grow faster than they can scale compute. All these signs point in the same direction: More expensive frontier AIs. The free-tier chatbots and similar will still be cheap, but if you want the best AIs in December you’ll have to pay a lot more than people are used to paying today. (Even on a per-token basis; since you’ll probably also be asking them to do longer, more ambitious tasks, the price increase will seem even bigger)
Most of those signs are Anthropic specific, and OpenAI has several signs pointing in the opposite direction. They have really been leaning into a message of “efficient frontier AI for everyone”, they reset rate limits multiple times per month as a meme, and Codex rate limits are already very generous.
Anecdotally I’ve never heard of anyone hitting a Codex rate limit on a Max plan, and from what I hear most people struggle to hit 20% of their weekly rate limit with heavy usage.
I think the warrant here is ‘Anthropic will likely have the best AIs in December,’ which is why he’s looking at the Anthropic-specific signs.
(I’m unsure if this will be true; reasoning models looked like a big leap, similar to mythos, and they just turned out to be ubiquitous a few months after the reveal.)
Edit: another possible warrant is ‘OAI could have the best models in December, if they develop something mythos-class, which will be much more expensive to run and force them into a similar pricing position to Ant.’
This would’ve been pretty counterintuitive to me 2-3 years ago thinking in terms of this chart, but it’s surprised me how quickly I get used to and want near-frontier performance as I keep integrating them into daily tasks. Jevons for sure.
Much of the use of AIs is not paid on a per coding basis. When I’m able to run 15 simultaneous threads (and 100 per day) for 20$ per month with Google’s rules, I’m not paying the per-token price.
I think that’s likely because the my usage data is valuable enough for Google that they are okay with my money not really paying for the usage resources because they want the training data.
I think there’s a good chance that you will have to pay more if you actually want the best AI in December, but there’s a good chance that you don’t have to pay more if the second best AI is okay for you.
Given the short timelines of the companies, their main priority is building powerful AGI and not making money right now.
I agree that their priority isn’t making money. But prices are set by supply and demand, not by how much companies want to make money. The demand is going to go up a lot as AI becomes more actually useful in the real world, and the supply is going to go up only a little bit as the companies triple their total compute.
Seems like AI is going to get more expensive over the next year. GPU prices are spiking a bit in the last two months, Claude users are complaining about usage limits more than usual, Mythos is supposed to be more expensive, Anthropic revenue has been growing at an insane pace (10x a year for more than two years) and seems like even as it slows down it’ll still grow faster than they can scale compute. All these signs point in the same direction: More expensive frontier AIs. The free-tier chatbots and similar will still be cheap, but if you want the best AIs in December you’ll have to pay a lot more than people are used to paying today. (Even on a per-token basis; since you’ll probably also be asking them to do longer, more ambitious tasks, the price increase will seem even bigger)
Or so it seems at least.
Most of those signs are Anthropic specific, and OpenAI has several signs pointing in the opposite direction. They have really been leaning into a message of “efficient frontier AI for everyone”, they reset rate limits multiple times per month as a meme, and Codex rate limits are already very generous.
Anecdotally I’ve never heard of anyone hitting a Codex rate limit on a Max plan, and from what I hear most people struggle to hit 20% of their weekly rate limit with heavy usage.
“If you want the best AIs in December…”
I think the warrant here is ‘Anthropic will likely have the best AIs in December,’ which is why he’s looking at the Anthropic-specific signs.
(I’m unsure if this will be true; reasoning models looked like a big leap, similar to mythos, and they just turned out to be ubiquitous a few months after the reveal.)
Edit: another possible warrant is ‘OAI could have the best models in December, if they develop something mythos-class, which will be much more expensive to run and force them into a similar pricing position to Ant.’
This would’ve been pretty counterintuitive to me 2-3 years ago thinking in terms of this chart, but it’s surprised me how quickly I get used to and want near-frontier performance as I keep integrating them into daily tasks. Jevons for sure.
Much of the use of AIs is not paid on a per coding basis. When I’m able to run 15 simultaneous threads (and 100 per day) for 20$ per month with Google’s rules, I’m not paying the per-token price.
I think that’s likely because the my usage data is valuable enough for Google that they are okay with my money not really paying for the usage resources because they want the training data.
I think there’s a good chance that you will have to pay more if you actually want the best AI in December, but there’s a good chance that you don’t have to pay more if the second best AI is okay for you.
Given the short timelines of the companies, their main priority is building powerful AGI and not making money right now.
I agree that their priority isn’t making money. But prices are set by supply and demand, not by how much companies want to make money. The demand is going to go up a lot as AI becomes more actually useful in the real world, and the supply is going to go up only a little bit as the companies triple their total compute.