My point being, quick AI slop landing pages and products will share the same vibes, but at the end Tailwind is just a collection of css classnames, allowing you to build and create as much as cool and unique pages as with plain css if you put the effort.
I wouldn't switch from Vercel just to save X% per month, because the real cost was all the extra time to build and maintain the infrastructure myself. The savings never justified the hours.
With agents that math changes. The build and maintenance work moves onto the agent, so the savings threshold where switching makes sense drops a lot.
On top of that there are other goodies. Vercel is serverless, so you can't co-locate the API with the DB, which means extra latency and yet another bill (Supabase?). That's one more thing you can now own for basically no effort.
Thanks for the message, is exactly what I have in my brain and how I'm seeing it unfold.
I'm lucky to be part of different codebases, +200 engineers codebase in a 10 years old company and code, +5 engineers on fresh code. My personal projects, that are beyond POC's, real users, hundreds of commits.
The LLM agent sweet spot is the last one, they are perfect, as I can contain most of the knowledge in my brain of how it works in/out. Speed is insane as a solo developer.
Then the 5 engineers codebase, is also really good, but here you already start to see the problems, thanks to agents you don't even need to care how it works, I have been working on it for +6 months, it uses TRPC and I don't even know (I don't care) how TRPC works. You feel that no one in the team really knows how stuff works at 100% (fresh codebase, we have build this ourselves!!).
Then there is the old codebase with +200 engineers, this is the worst of all of them, you described it perfectly, a bottomless pit of tech debt. This codebase before agents was an old non-typescript one, it was not perfect, but you could build a mental model and understand it perfectly after a few weeks working on it. Now, is a hot-mess of code duplication and the quality is degrading faster and faster as the code gets worse and the Claude Code adoption increases within the engineering team.
Not sure what will be the outcome of all this, but I wouldn't be surprised if some company wakes up in 2027 with a codebase that maintenance and development has increased by x100 fold thanks to Agents.
You nailed it, commute is the killer. I'm also Spanish, from a very small and rural town.
My father is a farmer and does a siesta every day of the year. He comes back at home of working in the farm every day around 1PM, then we have lunch together and he goes on to take a nap (siesta).
In winter they are shorter, 30 minutes, as the day is short.
In summer, they can go over 1 hour easily, as the day is longer and is hot between 2 and 5 PM.
Of course, my father is it's own boss and old school farmer, young farmers don't do that, and try to work on an schedule.
And is the same about school, when I was a kid no one was driving me to the school or taking me back, I walked there on my own, went home at mid day for lunch, played some football after it, and then went back to school for a couple of hours at 3PM.
I feel we are slowly drifting away from natural times and actions to forced on schedule behaviour to fit within the cogs of a late-stage capitalism productive machine.
In the only one that feels that OpenAI has bots/commenters on payroll on all this kind of news downplaying Claude and stating how much better Codex is?
There is too much and there are too many, and some of their takes don’t fly if you use Claude daily.
Same principle applies, hosting in Railway has slightly worse UX, but with LLM's you don't need to write a single docker line anymore, so deploying on railway is way way less cumbersome than before, and you gain more control and less costs.
This is something I noticed, originally I thought "AI" was the perfect tool for Vercel and Nextjs (current standard = future standard), but then I realized is the total opposite, their moat/stick is gone now, and Rouch that is smart I think knows this.
I switched a middle sized app to Tanstack Router + Vite while I was walking my dogs. Then 30 minuts-1 hour QA and it was done. This should have never happened before AI.
(I did switch because I was tired of the bloated network tab with 100 unnecesary RSC calls, the 5 seconds lag when clicking on an internal link, the 10 seconds "hot reload" after a change... I'm on a M4 MAX with 64GB of ram....)
At least in Spain this is exactly the case, and is really scary how much people falls for it.