Token至上时代结束之后会怎样？

Two very different sets of predictions

May 29, 2026

Within hours of when I posted yesterday’s essay on the decline of tokenmaxxing, I saw further converging evidence, such as this:

Financial Times@FT Amazon scraps AI leaderboard to stop workers chasing usage scores ft.trib.al Amazon scraps AI leaderboard to stop workers chasing usage scores 8:54 PM · May 28, 2026 · 41.5K Views * * * 28 Replies · 61 Reposts · 293 Likes Furthermore I saw bunch of other people saying similar things noting the apparent decline of toxenmaxxing, such as Fortune’s Jeremy Kahn in his newsletter:

So ok, if that’s true, what’s next?

Here are two takes. First my own:

And second, AI researcher and benchmark creator Lisan al Gaib, who fired back in a reply to me with a totally different, far more optimistic set of predictions, spelled out in admirable detail:

Within minutes, the contrasting pair of predictions become a meme:

We agreed to check back in 12 months to see how these predictions are going.

Stay tuned. And feel free to drop a comment below.

Discussion about this post

Yeah, I'm gonna vote on Gary's predictions here

I'm seeing projects like forge (https://github.com/antoinezambelli/forge) coming out, which of course isn't totally ready for prime time mass adoption... but I don't think that'll take a long time. Which makes me think that the value of hosted inference is just gonna disappear.

Hemmingway told this story well:

“How did you go bankrupt?” Bill asked.

“Two ways,” Mike said. “Gradually and then suddenly.”

I sense we're heading to "gradually" part. But suddenly is around the corner

I want to smoke whatever Lisan is smoking