Ask HN: Will I get left behind if I don't jump on AI train?

You don't have to jump. Not all progress goes forwards. As anyone who's familiar with web frameworks and crypto knows, most of the paths lead to dead ends.

Play with things. Treat them as a toy, not as homework. Eventually you'll get a feel for what is useful and what is someone trying to sell you something.

The first time I actually got something from a MCP was when I just asked claude to check my emails for anything I need to act on. I had 8 emails in the inbox, didn't expect it to find anything. But it found an important request from accounting that was cc-ed and forwarded multiple times to the point I didn't see the actual ask. It was also some tedious bureaucratic stuff so Claude write the response to it in 2 minutes and I didn't have to procrastinate on it the whole morning.

Perhaps it'll be interesting to see where AI and crypto intersect.

People are getting dumber, sicker and poorer, are they not? There'll be a need to change that. Make me healthy, wealthy and wise. And Vibing means more apps and ideas can be tested. Deployed is another story. Voice recognition will be more important I suspect as people stop reading and writing. How do people avoid the robots overall?

With AI, you can take an old app, and update it and modernize it, right? I was using this web app a long time ago. Would be cool to get it right for 2025. https://uipublish.sourceforge.net/ - who else would want to do that?

Depends on what your goal is.

Every few months LLM's get markedly better. In days GPT-5 will come out which by early reports (and anecdotal evidence on LMarena) is substantively better than the best current coding models. Will it surpass humans? It is a possibility, unless there is some yet unforeseen impassable intelligence gap below human intelligence and above SOTA AI intelligence.

Coding as a hobby seems to be more fun for people without AI tools. You get the experience of more completely holding the system in your head, your interaction with the codebase is more intuitive and connected deeply to your natural reasoning faculties. However, I believe the purism of eschewing AI tools is a trap that had few consequences in 2023, or even 2024, but by the end of 2025, the standard for what a "productive software developer" can accomplish in a week will be set by the boon AI tools give to mid and senior developers.

That's it, there's no getting around the reality of technological progress. Soon after that, there will be few if any humans in the loop at all for all except the highest level executive decisions. Coding will be an artisan task done for personal enjoyment, like knitting or cooking.

Next time you’re trying to do something a little outside your comfort zone ask Claude to make a plan of execution. You’ll probably be pleasantly surprised. And the fact that it can do that is all you need to know (and that it can do that for anyone doing basically anything)

But keep in mind that it goes spectacularly wrong sometimes because: It doesn’t have a way to verify facts. It doesn’t always follow best practices (like a junior employee eager to please forgetting his training) And sometimes it is very convincingly incorrect.

If you keep in mind those abilities and limitations you’ll know everything you need to know.

Oh also that maybe it’s becoming the largest consumer of energy in the world and may automate some or all of some or all jobs. (sometime in the next five years or never depending on which predictions you listen to)

Why should they want to become a developer? There is no need for everyone to become one.