sporadicprocess

What are some places that don't allow AI coding? I'm not familiar with any :(

I didn't say anything about software transitioning to low pay. I think you could argue that SWEs will get paid more than ever. I just want to code, not prompt.

I just think it's not tenable anymore, the work output expectations are rising and I won't be able to keep up. I already see projects being staffed with 2x less eng but the same output, and it's only going to increase.

I felt that way with earlier instances but Claude Code + Opus 4.5 has been miles better (haven't tried 4.6 yet).

I totally understand that and I know many people that feel this way as well. I'm just not one of them :(

I meant that I don't think it's bad for software as a product. It will be delivered faster and cheaper to the customers. To be clear, I do think it's bad for (many/most) people working in the field. Perhaps I didn't word it very well.

True, although I would be happy with 5-10 more years :). I was actually considering being an electrician, I've always been interest in that sort of thing.

I don't know if there's any jobs like that where I live, but I will take a look. Thanks for the suggestion!

\> Are all your tickets easy stuff/maintenance of a very brownfield project? Are you really sure their output is more than you could do on your own/only using them as a fancy Google rather than for code gen? Are you really sure their output is good? Cause it just doesn’t line up with my experience, having experimented with these tools a lot so far. I have found Claude Code to be very good, much better than anything else I've tried. Of course you have to prompt pretty well (similar to instructing a junior engineer), and we've also set up a lot of scaffolding internally that explains how all the codebase is structured and the patterns we use. So after that setup is done it works very well in my experience. If you don't go through that then you probably won't see it being as effective. Even now Claude is probably \~2x faster than I am and I would expect that to only improve over time. In any case I think arguments along the lines of "AI is not actually that good" aren't that compelling to me anymore after watching the development over the past year. Progress is so rapid that even if you don't think it's that great yet, then it will be in a year or two. I'm not saying that I think AI will actually replace \_software engineers\_, I just think it will replace \_coding\_ completely.

lol "grifter"... I'm a principal engineer at a big tech company. I don't count my accomplishments by amount of code obviously, that was just the part of the job I liked. I feel like you didn't actually read my post :(

Interesting. How can I learn more about those types of jobs? Can you give me some pointers of what to search for?

Is there an alternative framework that makes it easier? I think reactive programming is just inherently challenging as a mental model (compared to imperative). I've worked with reactive frameworks in other contexts than FE and they have had similar problems.

Ah yes, only your issues are hard and no one else's.

Yes, AI has some uses, especially one-off scripts or bootstrapping greenfield code. But the value for working on larger existing codebases is quite low IME. It's probably on the order of 10-20% coding speedup... for a job that is only spending \~25% of its time coding. So sure you can use it for some benefit, but it's just not good enough yet to have a dramatic effect. And of course in the long run AI really could be a huge game changer, that doesn't mean the bubble won't pop. In 2000 the Internet bubble popped but in the long run it ended up being a huge deal. Housing prices are well above 2007 levels. So bubbles don't mean the long-term value isn't there, they just mean that financial investment occurs too rapidly compared to the fundamentals.

I clearly didn't say "don't bother", you are putting words in my mouth. All I said is it doesn't work for everyone, which is factually true, as you admit. BTW, here is my source for 70%: [https://www.ncbi.nlm.nih.gov/books/NBK441838/](https://www.ncbi.nlm.nih.gov/books/NBK441838/) \> Stimulants are the mainstay of treatment for ADHD. They are effective in about 70% of patients.

Medication only works well for \~70% of people with ADHD. It's not a panacea.

It seems unlikely random people on reddit are better, at least.

The way I make it work is I just don't do those things. A big part of being successful is playing to your strengths. For me I enjoy and am good at the same technical things as you, and have the same difficulties with the "tech lead" work. Fortunately at my employer high-level ICs are not required to be "tech lead" types, so I simply have pursued a different type of role (I am L7 now, though I think progressing beyond that is probably not within my skillset). So you kind of have to decide for yourself, if you want to fight against it and pursue the TL path, then it's certainly possible, and there are probably some strategies or accommodations that can help. But I think it's also just very likely to lead to burn out.

Same, lol. My company is starting to test "AI skills" in interviews where you basically "vibe code a feature" (or something). If that becomes the whole industry I will find a different job.

Assuming you get jobs through contacts then not really. If you are just cold applying then it does help since there are automated filters. If you have a "brand name" job in your history though then it's probably not necessary either.

The company is in the US though so US law does apply.

Rich text editor

I know several EMs that switched back to IC. It's not uncommon at all. Being an EM isn't for everyone, it's a completely different job than being an IC engineer. There's no inherent reason that you'd be a fit for one just because you were good at the other. It seems like your managers don't think you are a good fit. You may want to explicitly ask them why.

\> I have PTO plans in March and May and they were okay with that, only for them to say on day 1 of my March PTO that all of it was now unpaid. Did you get their agreement to your PTO in writing?

I don't think anyone is reading past the first page. So I just trim it down to the most relevant stuff (I have 20 YOE) IME people mainly care only about your most recent job anyway.

I'm 40 now and I hardly feel like I'm in the "twilight years". I have enough to retire now but I choose to keep working since it's still pretty fun. I don't see much ageism at higher levels either, if anything it's actually easier to get high level offers now. I certainly wouldn't want to be a manager, I tried it for a short time and it was awful.

Apparently the "levels.fyi Standard" uses this system. Not sure where they get it from.

If you aren't able to retire by 50 after working in FAANG for 25+ years then you're doing something wrong.

My average from 10+ years of working is 31 per week. So the actual number would be higher if you exclude vacation, holidays and so on. > If yes I would expect them to be minimal, unintelligent factory work. Let's just say, my employers don't seem agree with this take based on my performance reviews (and compensation!)

Is this based on your personal experience?