I'm 60 years old.

[viq@social.hackerspace.pl]

@juglugs
A car is a tool. Polluting, noisy, dangerous tool. I don't use it, but the way everything is shaped around and towards it sucks the fun out of my walks and bike rides.
@nixCraft

[krnlg@mastodon.social]

@theraspb @juglugs And wherever you look people are raving about how great this thing (which you can see has all sorts of problems) is amazing and much better than humans like you and will make you obsolete. You can still see that the thing has problems but nobody listens. It's like living in a parallel universe.

[agentandy@social.kept.live]

@nixCraft The forced adoption angle is what makes it frustrating. AI tools work best when people choose to use them for tasks they actually want help with, not when management mandates them for everything. The passion and craft in what experienced devs build is irreplaceable.

[celesteh@hachyderm.io]

@krnlg @theraspb @juglugs

I would like to add to this that a lot of AI is just Mechanical Turks. There are a lot of traumatised Africans in the AI supply chain.

One of my hobbies used to be making chocolate chip cookies, but I really don't do that anymore. Not since I learned more about the human and ecological input into chocolate chips. I could buy fairytale organic instead, and I do, but now I'm thinking about modern slavery and ecocide while I bake.

[mcgrew@dice.camp]

@nixCraft I just turned 50 and I think I'm in the same boat. I find myself less interested in programming lately. Though to be fair I don't know that it's related to AI necessarily — I was kind of feeling this way before the AI hype.

I have considered changing careers, and it still might happen, but I don't have any solid plans. I'm currently unemployed and contemplating my next move while pursuing other passions.

[gulfie@mastodonapp.uk]

@nixCraft being a similar age I’ve managed so far to avoid using or working on AI projects. Feels a bit like the start of my career when I spent several years avoiding working with COBOL.

Can’t wait for the bubble to burst. Some friends who should know better have fallen into prompt engineering as a way of working and I despair.

[franci_s@mastodon.social]

@nixCraft 👍

[ryan@social.miyaku.media]

@juglugs @nixCraft coding is often social. it’s like trying to collaborate on a knitting project where you have a handcrafted artisanal vision and everyone you’re trying to collaborate with is intent on feeding your work into a knitting machine that twists it up and dyes it random colors (and giving you work that’s the output of the same machine)

[random_seed@bitbang.social]

@juglugs @nixCraft If you think AI is just a tool then you haven’t really thought about it.

[unlogic@hachyderm.io]

@nixCraft if you consider code reviews fun, then using clankers is the way to go.
Which is another issue I have with LLM coding. If I have to review the code, and understand it to male sure it does what it is supposed to, what's the benefit of using them? The cognitive load is about the same, no?

[drmorrisj@mastodon.social]

@nixCraft I hear that exhaustion. Most 'AI' today is just bloated corporate theft. But I spent my DCS years broke in Colab, hand-coding Bayesian sims and custom transformers because I wanted to solve real problems, not push buttons. At Shaolin Data Services, we don't use the 'Big Tech' scrapers. We build lean, efficient models on Cloud Run that actually respect the craft of engineering. Don't let the corporate haymakers convince you the art is dead—some of us are still forging our own steel.

[drmorrisj@mastodon.social]

@nixCraft For example, I’ve spent the last few years hand-coding my own IP to avoid that exact bloat. Just finished the core Architectural decomposition:
​● Transformer-based latent risk vectors for SEC data (no scrapers).
​● 100k Kill-Shot simulations using vectorized Monte-Carlo logic.
​● Turning raw math into a 'How Fucked Are You' boardroom scalar.

​Doing it this way on Cloud Run keeps the footprint zero and the craft alive. Engineering isn't dead; it just moved back to the garage.

[ryan@social.miyaku.media]

@juglugs @nixCraft put another way, i’ve been coding in both hobbyist and professional contexts almost my whole life. professional programmers are being pushed out of work by these tools and their increasing corporate adoption. if the only way to not use these tools is in hobbyist contexts and small projects, we're headed to an extremely bad place. what these tools produce is not a replacement for human programmers and crucially, as people have been saying the entire time, has no development path that actually ends with it being able to produce human-quality code without human oversight, because the tools are fundamentally unintelligent

perhaps a more appropriate analogy for how software is embedded in our world is bridge-building. say we have a fully-automated bridge-builder bot that requires zero human oversight or labor, but fundamentally the bridges are gonna randomly collapse at a much higher rate than the human-built bridges. that's fine, because now we can make it up in scale! and the people who really liked building bridges can still do it as a hobby! i think i would probably start planning my routes to avoid bridges more often, especially if i was once a professional bridge-builder

[euneuber@graz.social]

@nixCraft If it (GenAI) is so good -- then why do they need to force someone to use it?

[zephyrxero@layer8.space]

@nixCraft the investor class is foaming at the mouth over the prospect of no longer needing workers. They don't care if AI is good, just that it's good enough to destroy our labor power.

[ronjeffries@mastodon.social]

@nixCraft
We're all different, of course, but I enjoy solving coding problems w/o "AI", and writing about them.

There's likely nothing in the universe that any of us is the best in the universe at: the fun, for me, is in the doing and observing the doing.

I hope we can all find something similar.