I miss Eugene Wei’s epic tech analyses. His latest serves as a sort of eulogy for Twitter. He’s of the opinion that “peak Twitter” is in the rear view, and the platform seems to be on a continuous decline over the past year, for the reasons I’m sure you suspect even if you don’t agree.

He makes some interesting points on the failings of Elon and the directional changes in culture and resourcing. While the product was obviously stagnant and incredibly slow at innovation, there have been some, let’s say questionable major changes in the last 12 months. Shipping is great. But shipping for shipping’s sake? Maybe not so much. Many of the changes when viewed from the outside — the yoyo-ing around with Twitter Blue subscriptions, the mind-bending rate limiting decisions, blocking whole domains from being linked to, then unblocking them — we’ve got a nice little list of bizarre experiments that seem unmoored from its success as a business.

No one has been able to explain exactly how Twitter could grow continuously (though slower than your typical unicorn startup) for the past 15 years, especially with so little in the way of new capabilities of the product. 15 years is multiple lifetimes in the tech universe.

I think the key observation is that Twitter found generational product-market fit over a decade ago, but hasn’t been able to explain or understand how that PMF exactly works. What ingredients and magic alchemy came together to create such an unkillable product? As the newfound czar of the product, Elon hasn’t decoded it either. Tinkering with the timeline algorithm has made it demonstrably worse, as I’m sure most power users agree. I like how Eugene puts it:

The changes to the Twitter algorithm bulldozed over a decade’s worth of Chesterton fences in a few months.

It’s not that the old algo was perfect. But something about its particular tuning methodology was clearly desirable, which we appreciated once it was taken away.

Here’s Eugene again, in summary:

For this reason, Twitter won’t ever fully vanish unless management pulls the plug. None of the contenders to replace Twitter has come close to replicating its vibe of professional and amateur intellectuals and jesters engaged in verbal jousting in a public global tavern, even as most have lifted its interface almost verbatim. Social networks aren’t just the interface, or the algorithm, they’re also about the people in them. When I wrote “The Network’s the Thing” I meant it; the graph is inextricable from the identity of a social media service. Change the inputs of such a system and you change the system itself.

I’m not going anywhere. Threads is uninteresting to me. Twitter is mostly still fine. And Farcaster is my second home and exit strategy.