The State of AI Twitter

First and foremost…follow me on Twitter

The Incentives Are the Problem

AI Twitter isn’t broken because there are bad actors. It’s broken because the incentives reward a very specific kind of behavior: constant optimism, public self-congratulation, and the appearance of progress whether or not progress exists.

The safest thing you can do in this space is sound impressed—by yourself, by others, by the pace of change. The feed is full of posts that function less as communication and more as proof-of-participation. Demos get posted before they’re useful. Threads get written before there’s anything to conclude. Every small iteration is the Next Big Thing.

This isn’t accidental. Engagement rewards certainty and speed, not accuracy or restraint. It rewards people who can confidently say “this changes everything” over people who say “this mostly works, but there are tradeoffs.” Over time, that trains everyone—founders, builders, observers—to perform confidence first and I suspect it has not trained them to figure it out later.

The result is a culture that talks a lot about building, but much less about what actually holds up over time.

Trying to Find Signal as an Outsider

Coming into AI Twitter without an audience makes this dynamic even more obvious. As a cold-start account, you quickly learn that nuance doesn’t travel far. Careful takes don’t spread. Skepticism reads as negativity. Meanwhile, vague optimism paired with the right buzzwords gets rewarded almost immediately.

That makes sifting signal from noise genuinely hard. Not because there’s no smart thinking happening—but because it’s buried under layers of hype and mutual back-patting. You have to work for it. You have to mute aggressively. You have to follow people selectively and ignore follower counts. You have to read past the framing and ask basic questions like: Is this actually being used? By whom? For what? Are there really that many people who need a fleet of agents to do SEO for their personal content creation brand?

There is value on AI Twitter if you treat it like a noisy marketplace instead of a curated journal. It’s useful for spotting trends early, seeing what people want to believe, and occasionally catching a real insight before it gets diluted.

The trick is not letting the tone of the platform distort your own judgment. Progress in this space is slower, messier, and more constrained than the feed suggests. Remembering that—especially while trying to break in—is the difference between learning something and just learning how to perform belief.