@tante And I cannot even begin to emphasize how much *it will cost about the same despite it being of lower quality*.

Once credit entities realize that GPUs get obsolete very fast and five years down the line the early-mover needs to buy again just as much new processing hardware as a late-comer, they will stop subsidizing today's AI as a gamble to capture the market for tomorrow.

And then your 45-minutes-saving boilerplate machine will cost 5$ per run.

@tante Good take! But also, like "can you create software" is not really an accurate framing of what the hard part of software was.

Most people could "create software" by looking up a Hello world example. That wouldn't help them solve amy real problems tho.

LLMs produce software that *looks more like* it solves problems... but security, integrity, legality were kind of always implied parts of the problem.

Like, it takes a weird subtle reframing of the goal to make LLMs look at all useful.

@tante

Entropy is somebody else's problem - they'll be comfortable on their yacht. We can't expect sociopaths to care about others, else they wouldn't be sociopaths.

@tante Well, well done for admitting that demonstrably, the dog can play the piano. Now we are just talking about how well it plays.

FWIW these LLMs have no need for being consistent with what happened in a previous context. The same LLM, in a new context, will usefully critique and find and fix flaws in what it itself did in the previous context.

The "slop" aspect of LLM output seems to come from just going with what one context produced blind, when it can iterate as, eg, QA manager.

@tante

I mean, you can also "build a house" by using deck screws to connect some wet doug fir 2x4s into a "frame" and then staple on some drywall and siding and drape the whole thing in a plastic tarp.

You will die when it falls on you, but for a time, it was a "house".

@tante The best engineers I know just became more ambitious, and so should all of us.

I'll keep repeating this, there's tonnes of proprietary binary blobs in all of our tech. You can shout from the rooftops about how much you love your /e/OS phone, if your phone modem relies on a proprietary driver, it's pretty much worthless as a "resistance against big tech". European digital sovereignty is equally worthless.

LLMs are good at staring at hexdumps, humans aren't. Use their advantage to build actually open tech.

> But you do pay a hefty price: In lowering quality (security issues, less maintainable), in skill decay in the people "guiding" the stochastic parrots, etc.

Skill issue, idk what more to say. I don't find it any different to managing juniors and reviewing their PRs. Bad code is bad code.

@tante

5 years retired from IT but I still remember CRUD

@tante I think LLM's are only good at learning programming language's basics or getting yourself interested in some random topic. That's what I do with a local LLM. I do the actual research and learning more than the basics myself.

Idk. I've never liked being spoonfed information anyway and I learn best when my effort of learning pays off. I always liked reading and learning stuff on my own even before LLM's started popping up but idk why people started decided to become really lazy.

@tante I don't believe that to be universally true. I *wish* it was, because it'd be so much easier to argue against them.

Unfortunately, the "mere" fact that all currently existing incarnations are fundamentally evil does not mean they must lead to lower quality software.

A velocity-first mindset has *always* led to lower quality, regardless of GenAI. And they make that rush accessible to everyone, regardless of expertise/skill.

[1/2]