ML =/= AI. There are legit uses for ML that don’t have anything to do with LLMs and the cloud. I worked on an ML project 3 or 4 years ago to listen for fan noise that might indicate that it was about to fail soon. We trained a tiny GAN on good and bad noises. It runs on a tiny CPU, locally. Highly specialized work, and I have to imagine there are and will continue to be lots of similar opportunities to bring efficiencies by getting computers to make good observations and decisions - even if only about “simple” things like “does this thing seem like it’s about to break?”
GANs are so much fun and so tedious. I trained one on dungeon generation in college. It sucked but it worked in the end mostly. I dont know exactly how an LLM works but GANs are way different.
Fair enough. ML ⊆ AI then. But these days when everyone talking breathlessly about AI taking away jobs they’re almost always taking about LLMs. This article is about ML in particular which is a different discipline with different applications.
They really need like 20, but to get colleges to host the courses needed you have to be able to fill the classrooms. Release bullshit articles like this and a bunch of kids hoping they might have a future with a roof on it sign up for the new thing being offered. It’s really insidious.
Image generation is neat. But now there are so many badly generated images out there that any new model likely feeds on shitty data. Also the whole copyright debate for the images used.
LLM are neat. But there is no point in widespread adoption. Any model that is built to generate correct sentences but not correct content kind of wastes a lot of time for the user and nothing else. They are inpressive, they can be fun and sometimes they are really helpful - but you never know whether or not they hallucinate any given information.
Voice and Avatar generation is neat. Like genuinely neat. But you rgey are so easy to use already, you don’t nees that many specialists.
Making the call now. AI as a business will be dead in 1 or 2 years.
It will have moved on to individuals running their specialized smaller models focused on specific tasks on their own machines, and google and openAI running their more general AI at a loss every year.
Whenever a new hype is going around I like to think back about the 3D printing craze and how little is left of all the glorious promises.
They are cool and a neat way of manufacturing things, but what they are absolutely not is magic machines.
AI/ML will find it’s niche and will allow for new and even exciting things, but it won’t be the end-all-be-all in it’s current form. It’s an overgrown version of statistics after all.
the 3D printing craze and how little is left of all the glorious promises.
Not sure what the promises were that you in particular heard. But 3D printing is a fundamental part of prototyping now. The vast majority of companies designing physical products have 3D printers to try out new ideas.
It depends on what you’re calling AI. The LLM hype may die down, but Ml/AI in general has been continuing to grow and expand for well over a decade. It’s just unlikely that all the things being prophesied right now will come to fruition.
As a software engineer/data scientist who has spent quite a while to find some good AI work, this sounds like absolute bullshit. Most companies don’t need AI. Prompt engineer seems like a niche thing, I can’t imagine that most companies really need someone who does that. It really frustrates me that these bullshit articles keep coming out without any sense or reason. AI is cool technology (imo), but currently it’s just the latest bait for CEOs, managers, etc. Somehow these kinda people are just so vulnerable for hype words without ever thinking more than as second about how to use it or whether it’s even useful.
To add to this, even when they need AI work, they don’t look for actual AI specialists. What they do is take one person who’s done some random ML work and a bunch of random junior devs and that’s the “AI team”.
That’s very closely related to something I’ve come to think about tech: nerd equivalency. If there is a computer involved, then a nerd is required and they are all interchangeable.
Basically, someone says “we’re not moving fast enough, hire another nerd!” and nobody in the chain of command or in the hiring process has a clue which particular skills are required, assuming that everyone can do everything.
That’s why so many corporate projects have what amounts to random people doing randomly assigned work producing insecure, unreliable products with obscure and even hostile UIs.
If you really want to make an AI-pushing CEO uncomfortable, ask them how they’re going to integrate AI into the blockchain tech they were raving about but are now suddenly silent on.
AI is cool technology (imo), but currently it’s just the latest bait for CEOs, managers, etc. Somehow these kinda people are just so vulnerable for hype words without ever thinking more than as second about how to use it or whether it’s even useful.
I think that’s a general problem with most technology that is fundamentally about computing.
People outside any field have only the barest grasp of that field, but the problems are so much worse as soon as computers are involved. They are so ubiquitous and so useful to so many people with little or no training or understanding that everyone just succumbs to a form of magical thinking.
it’s their marketing brains… believing hype and parroting buzzwords are mental shortcuts to never having to think for yourself and they prefer it that way
Add comment