Hacker News Reader

Comment by BizarroLand | Hacker News Reader

This is my darkly optimistic take on enshittification:

Companies know how to make good product, but if they don't have "new and shiny" to impress us anymore, then their only alternative is to make things worse so they can heel turn and then make things "better" by unmaking all of the worse things they did.

They can also milk their customers coming and going in the process.

It's not "enshittify or lose", its just raw greed. Things will get better again, either that or a competitor will destroy them. Enshittification is just the current meta and a new one will come soon enough.

View on HN

I don't think companies know how to make a good product any more. Conway's law won this battle.

I think it's that company management has no incentive to do well. So they have no reason to push this down to the bottom tier of workers who actually make the products. The feedback loop is open. They make an order, the product gets worse, the line goes up, they don't know the product got worse and they have no reason to care anyway.

Slop didn't start with AI.

The West already forgot how to manufacture things, and we are now forgetting how to code: https://news.ycombinator.com/item?id=47907879

When is the "get better" step? I've only ever seen two things happen mid- or post-enshittification:

    1. The company builds a moat and just remains shit.
    2. New entrants either displace the company entirely (most likely) or competition slows the enshittification process (distant second) or reverses it (almost never).

It's not clear to me why "get shitty" is a necessary step to this. What part of GitHub's executives' grand plan is "have a barely-functional service that randomly prevents people from working"?

The enshittification process milks the current product of all of the money that can be wrung from it by any means just shy of immolation.

Companies aren't getting cheap loans right now so they're desperate to juice their stocks so that upper management can secure their bonuses.

That's why "get shitty" is necessary.

When they've wrung it dry, pocketed all of the crumbs of raw cash they can get, then they'll either collapse due to overmilking their products or they'll realize that the only way to refatten the calf is to bring in new customers, so they'll unshittify it for the fresh infusion of customer money.

It's a cycle, and one I predict will inevitably lead to many of these companies' collapse.

> What part of GitHub's executives' grand plan is "have a barely-functional service

What about lock-in, being a monopoly? Why wouldn’t you maximize on saving costs? Sure some people leave, but the majority is not going anywhere. And if the platform dies they’ve made more money than to keep it alive.

> It's not "enshittify or lose"

I think it’s “find natural monopoly and reduce costs (aka enshittify)”.

Github is a natural monopoly and users cannot go anywhere. Unless you’re famous like Mitchell Hashimoto.

Depends on how strong a moat really is, but it can be "enshittify and lose", too. Enlightened (as opposed to short-term) self-interest may pay off after two years or twenty, depending, and in the latter case, it may as well not pay off at all as far as a public company are concerned.

I think Microsoft’s home game is “monopolize and enshittify”. They are the masters and know the exactly what amount of enshittification is too much. E.g. Hashimoto quitting GH is probably totally worth the 10 SREs they fired. Us plebs cannot go anywhere.

But you totally can go somewhere else? E.g GitLab (which is, unfortunately, about as slow as GitHub, but with a better license and owner) or sr.ht.

If you think you need those sweet GitHub stars, I can't help you.

I mean the stars, social features and branding of Github make it more or less a lock-in. You can go somewhere else, but it’s not the same experience.

> If you think you need those sweet GitHub stars, I can't help you.

The majority of users need it.

Why not use GH just as a front page you mirror code. You keep your stars but develop elsewhere, on a server you control.

https://lists.sr.ht/~machocam/public-inbox/%3C46e343ec-c932-...

GitLab is definitely not as slow as GitHub, and it doesn’t matter because if you want to you can self host it, where it performs absolutely fine too

Reply on news.ycombinator.com