823
you are viewing a single comment's thread
view the rest of the comments
[-] Fandangalo@lemmy.world 3 points 1 day ago* (last edited 1 day ago)

There is a vast gap between casual AI usage (chatbots, meme images) and AI-based coding / agentic engineering. There’s also nuances between “vibe coding” and using AI as a tool to write code.

Most engineers I know are basically not writing code. They know good code practice, they understand software architecture, and the core drive for coding environments / IDEs has been “help programmers type less / be more efficient.” AI in coding is a big evolution of that, especially because of the mechanisms humans use to read & write. We cannot read faster than we write, so it’s faster to read. The AI can write much faster than we can. Therefore, you use the AI to write & execute the plan, with engineers reviewing the output.

For my own industry (game development), AI has been kinda nuts the last few months in terms of rapid acceleration of development. It’s amazing for prototyping concepts that won’t go to market in their prototype form, and that cuts down on risk, downstream impacts to actual engineering requests, art, etc.

This creates the potential for a captured market (software development) with AI-usage seen as human speedups. It’s unclear whether people are in fact more productive right now, but it’s humans fumbling around with new tools. I’m kind of afraid how far AI could advance in a very short period of time. Many of my friends are using recursive techniques to improve their AI workflows (asking the AI “how could we do this better?”).

C- & D- level have massive amounts of their portfolio in AI tech while spurring on employees, “you must use this!” That part is an ouroboros / inflationary bubble.

This is putting aside all the ethics for stealing literally everyone’s stuff, the water usage, problems with capitalism, and so on. Please don’t take what I say as, “AI is good!” It’s just more complicated than that.

This gap is so hard to explain. A lot of common folks are seeing chatbots screw up counting R’s in strawberry, but I can talk to a rock & a game comes out built on web frameworks I barely know because knowing the language isn’t super important compared to what comes out.

[-] queermunist@lemmy.ml 5 points 1 day ago* (last edited 1 day ago)

There's a in-build assumption that the LLMs are going to advance; people assume all tech always advances forever.

It can't, not really. All LLMs hallucinate, no matter how much compute and how much data we feed them. It can be good for making quick and dirty slop, but that's the limit of what it can ever achieve because of fundamental limitations of the technology.

That isn't to say generative AI won't get better, but there's nothing to fear from this particular dead-end technology.

They want you afraid.

[-] Fandangalo@lemmy.world 1 points 14 hours ago

I don’t think there’s some poison pill or fundamental flaw. It regresses to the mean for the context of the problem.

I think you’re writing something off too quickly, but that’s your life to live. No judgment.

Have a good one.

[-] cabillaud@lemmy.world 1 points 22 hours ago

I really hope you're right.

this post was submitted on 30 May 2026
823 points (99.0% liked)

Fuck AI

7209 readers
2127 users here now

"We did it, Patrick! We made a technological breakthrough!"

A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.

AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.

founded 2 years ago
MODERATORS