Ask HN: Looking at AI development, I feel pointless in the work I'm doing | Hacker News Reader

Ask HN: Looking at AI development, I feel pointless in the work I'm doing

15 pointsdxaca20 comments

As a web developer, I thought I was doing well for years. But recent work on AI makes me feel useless. How do you guys feel?

View on HN

Concretely, if you're a web developer - go up a level. AI doing a lot of "web development work" means that now you have time to do other things that you pushed aside as too hard. That might be back-end systems / databases / algorithms etc.

Or learn about user experience, design, how humans think and how to build good products.

Please don't get discouraged by the hype. The only people saying AI will replace developers are the people who have something to sell. Engineering is more than just writing code. If anything, AI will increase the level of complexity in the software today, so you'd need engineers to tame that increased complexity.

---

This is from a related thread I wrote in Reddit:

There has never been a more exciting time than now to be a software engineer.

In a nutshell: all the parts I enjoy about software engineering remains challenging, unchanged and open areas of research (distributed systems, algorithms etc). All the grunt work that I hated doing (CRUD work, remembering tailwind classes) have gotten automated very well.

I've played around with AI a lot. I mean A LOT. People who don't code fundamentally lack the insight that writing code is not the only thing that a software engineer does. The way software engineers are coding is changing - but it's changing it in a way that make good ones better, not necessarily bad ones any good.

This is pretty accurate. I realized it early on in my career that writing code is the easiest part of software development.

I remember the discussions about the impact of Microsoft FrontPage on web development. The consensus was that FrontPage produces crap code which does not work well, and if you need something good, you need to write the code by hand.

25 years later I hear similar discussions about web development, and we are getting to the same consensus, and that's very entertaining.

The time invested in creating messy and frequently updated abstraction layers turned out to be well spent too. I mean, I could probably trust "AI" to write HTML4 with php includes better than Frontpage, but as long as the pace of breaking changes in APIs and node packages and CSS metaframeworks stays the same, AI will never catch up :)

I have seen no indications that AI has decreased the need for web work. The overall market is saturated especially for any commodity framework web developer or “full stack developer” or backend CRUD/framework developer.

I’m not saying that’s what you are and that’s just a general statement. I saw that myself when I was looking for standard enterprise developer jobs both last year and the year before as a plan B.

Whether web frameworks weren't a PITA to work with, AI might have dominated them by now.

I intentionally don't work in AI, and avoid it as much as possible; I would rather humans holistically approach meeting our needs than "over-intelligence" oursleves onto a WALL-E spaceship. I finally found a sense of purpose after my child was born, and another a few years later- rehabilitation of the land so that more people can live and work to meet their needs locally. I met this need for purpose mostly with videogames prior to this, offset by teaching others- that felt good, but it didn't really go deep enough until I became a parent. Then it became apparent what a sense of purpose can be! Keep searching, something will turnip.

I mostly do RCA, and at first I thought that maybe generative AI would be able to take a system report or a core dump and spit out something useful. It can't, though.

It's interesting how some "lower-level" work might be able to outlast some "higher-level" work. Using expert knowledge to build out a system might get pushed out at all but the highest levels before AI is able to figure out why highly complex systems are broken (in new ways). RCA (beyond pattern-matching against KB articles) isn't very susceptible to systemization, because it usually only comes into play when something unpredictable happens.

If it makes you feel any better, the vast majority of people working on AI stuff right now are building “me too” products - yet another chatbot, yet another knowledge management tool, yet another sales automation tool, yet another image generator. Almost all of them will be thrown out and ignored in a few years

Only 1% or less are truly working on the hard AI stuff, and AI research

I switched to slacking off in my web dev job and doing all then menial work with AI. And then in my mean time I’m learning marketing, sales and product… it’s surprising how far that gets you. I was able to net 300$/mo already just being an affiliate to another software product within less then two months of time.

I feel the same too. I think AI can take some jobs in maybe a couple of years.

I'm sure some FAANG teams are experimenting with AI integration but without proof.

I can only hope I manage to move into a lower level before AI gets me.

20% of my work is already done by AI. If they integrate properly and make human teams adapt AI, which is easier than the other way around, that can be 90%. And AI doesn't rest.

I believe you should pay for a cursor license and learn how to use it.

You'll quickly notice how limited AI is.

It can generate some code, but often it's just boilerplate similar to what Frameworks and Libraries do.

It is far from fixing tough bugs, fixing performance or architectural issues.

Focus on what you think the AI isn't good at doing.

We don't write software, we value add to business

You have to start trusting yourself and valuing what you find interesting more. One caveat to my advice though is if you have dependents in which case I'm not the right person to ask hah

Just go head first into AI.

Buy the best text editor (Cursor) and pay for a month.

Learn how to use it well, prompting techniques etc.

This will show you all the limitations it has, so you instead invest your time on specializing in what the AI sucks to do.

We never thought you web guys were doing well, really. With all the empathy to you as a person, these jobs were a result of the industry being an endless field of burning barrels of cash, not some good engineering. Every few pages of webdev was a single click in the '90s style RAD IDE. Every now-team was just one part-job with no particular specialization. And you didn’t even think of boiling it all down to something sensible. This whole job is a bubble.

It’s no secret that this is where AI performs the best and doesn’t embarrass itself in the second sentence. I’d even claim that webdev heavily masks the inability of LLMs to write code yet, because it does relatively well in the “webapp” section.

hate this kind of web hatred. I also worked on embedded software, backend, infra stuff. I personally needed more high level of engineering work in the webdev. Please don't do this

I hate the game, not the player. Should I just shut up about the fact that web is underdeveloped? Maybe I failed with the proper delivery, sorry about that. But having to deal with it too and having the past experiences with decent ui libraries leaves no room for confusion. It is absolute crap of an industry with bikeshedding, wheel engineering and endless massaging of parts that must have been solidified ages ago. Ofc one needs high level engineering to do all that, everyone would need it if forced to write for half-baked platform. But it’s no good thing that you have to do it, over and over.

This is a strange criticism but the style it’s written in is oddly even stranger.

Well, you know, we're having to deal a lot with our own superiority complexes, suffering the disparaging attitudes of our peers, the need to re-invent wheels in the browser without ever having learned much in the way of UI fundamentals, and a constantly shifting baseline of browser APIs, bells, and whistles. And if that weren't enough, we also have an unending supply of amazingly clever developers releasing new frameworks that compete for mind- and market-share every season. Which is to say, yes, we do think of boiling it all down to something sensible, and some part of ourselves have, but many others of us are too busy in our unspecialized part-jobs, maintaining the health of our now-teams, to piece it all together in a way that makes enough grammatical sense for the LLMs to understand. It certainly could be a bubble, sure, but it looks more like an ocean to us.

See AI as the enabler. Learn how to use it, but have a plan to move up or away through upskilling and taking on more responsibility

Edit: autocorrect fail

Reply on news.ycombinator.com