Hacker Newsnew | past | comments | ask | show | jobs | submit | robertjpayne's commentslogin

V2 to V3 was really good value, but V3 to V4 was mostly performance with a migration nightmare with little new features.

I don't know what a Tailwind V5 could add that is "breaking" and be worth the migration headache again.


Yes but Tailwind Plus has a flawed business model, AI was not really the reason nobody bought it, it's that it's a lifetime purchase and that shadcn + LLMs has eaten their cake left right and central.

If LLMs didn't exist but shadcn still did, do you think people would pay and use Tailwind+ or shadcn?


Tailwind UI is tool companies buy to save dev time mostly on internal/back office tools. It's usually bought per project. The math is pretty easy - if it saves you few hours of devtime you buy TailwindUI. Shadcn and bazillion other similar things are certainly competition but TailwindUI is very broad and of high quality so why not pick the nicest version.

The problem is that Tailwind is extremely portable (thats why it's so popular) and since LLMs have been fed all TailwindUI code... people using LLMs don't even have to know that TailwindUI exists they just get some Tailwind styled components. They would probably look pretty confused if you told them you used to buy these templates.


What’s the problem with the lifetime purchase?

It's the difference between one-off revenue and recurring revenue. If you're making new components, making new changes for the new version, adding new css and browser support it's hard to keep going with only income from new customers.

It takes the recurring out of recurring revenue, 100% churn

shadcn/ui I'd argue is probably the single biggest factor in the declining Tailwind revenue more so than just LLMs in general.

As said is it is to say shadcn is what Tailwind should've created and maintained for a fee rather than some html/css templates that are easily replicated.

I say this as someone who bought Tailwind+ to support the project many years ago and still use Tailwind every single day.


It's less about cycles and more about the energy density per kg. Nothing on the market comes close to 400Wh/kg.

Yea this take makes no sense. What in the world is wrong with debugging with breakpoints?

It is for many problems, especially concurrency related ones, much less powerful than trace points. But the issue I have seen is that some tools like gdb have unergonomic support for tracing so there I tend to use break points or printf debugging just because the tracing support is so bad in gdb.

It’s a particular subcategory of cork sniffing where you pick the hardest, dumbest way to do things because you’re a Real Developer

Chrome probably has the benefit of being updated frequently rather than more of an annual cycle. But Safari still isn't anywhere near IE6 levels of awfulness.


Safari adds support for new web features in point releases throughout the year.


But only for the latest major iOS version.


The latest version of iOS supports the iPhone 11 introduced in 2019.


And I use PCs way older than that with no issue. 6 years isn't that long


And what Android phones from 2019 are still getting updates?


Browser updates? Most of them.


The latest version of Android Google Chrome requires Android 10. But to keep getting updates you need Android 11.

Running a browser on an OS that doesn’t get security updates doesn’t sound like a winning combination.


More winning than updating neither


WebKit also isn’t trying to push a proprietary OS-locked runtime for interactivity, doesn’t wildly diverge in most rendering behavior, and handles most basics correctly.

As much as IE6 was a menace for not keeping up with standards, what made it really bad was crap like ActiveX, radially different layout/rendering behaviors, and shortcomings like inability to render transparency in PNGs and some of the most illegible italic text rendering I’ve ever seen.


Put a spend limit in GitHub and issue a chargeback if they ever bill you more.


I think the problem is it's too risky to utilise a "new" database today that isn't available as a managed option on many platforms.

Neon is just Postgres for the most part -- sure if they shut down it'd be a pain to migrate but you'd mostly be fine just swapping to AWS RDS/Aurora or the plethora of other Postgres providers.

Buying into things like Gel or now defunct RethinkDB mean if the primary company went under you're stuck with a dead DB engine and potentially no managed hosting options.


Gel is Postgres.


I mean sorta? It's value add over something like Aurora or RDS has been it's SDL and EdgeQL layer which makes it a hard pill to swallow not knowing if it's going to go under and you're stuck trying to maintain that yourself or migrate off of it.


Not sorta, quite literally. You can start a Postgres instance and point gel-server at it.

By default, it does start a managed Postgres because that's more convenient, but you can point it to Aurora or RDS or any other pg deployment.


Less than two months for any production users to migrate. Sounds like they've basically got no paying customers otherwise I'd expect to have seen a longer than 2 month period to migrate especially over the holiday season.


This is an insanely good take I never thought of.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: