Recently, @VitalikButerin has been quite critical of the AI track.

For example, he suggests delaying AGI for four years and claims that what AI truly fears is the permanent concentration of power, among other things. If you think these statements are too "fundamentalist," take a look at the market complaints from college students who can't afford Claude MAX—perhaps you can get the gist of the underlying message?
Some even say that Vitalik's comments are meant to support crypto, especially those who simultaneously praise AI and dismiss crypto as useless. 😓
But we seem to have forgotten one thing: what crypto has been doing over the past decade or so is fundamentally about embedding rules into code (Code is Law), rather than relying on a centralized institution’s servers. ETH smart contracts, Bitcoin’s PoW, and various decentralized, permissionless protocols are all pushing in this direction.
Because, in the era of blockchain equality, a college student and a billionaire have at least the same permissions to interact on the chain. But looking at the development trajectory of AI, it’s quite the opposite. The more powerful the model, the higher the training costs, the higher the inference costs, and the smaller the group that can truly use it "frictionlessly" (if you think costs are not equivalent to friction, you can skip this part...).
One day, the market will realize that as AI advances, the value of decentralization becomes even greater—more so than the initial calls for decentralization under the banner of fighting banking monopolies.