- cross-posted to:
- technology@lemmy.ml
- cross-posted to:
- technology@lemmy.ml
You must log in or register to comment.
I’ve done useful things with LLMs, some of which I probably wouldn’t have done without it. I’m sure things would be different if frontier models’ token weren’t subsidized by investors, but I’ve been running local models as well and they aren’t useless.
What scares me is people blindly using these tools without hesitation, completely outsourcing their thinking to them. I work industrial process automation and that approach scares me a lot, in my field code is generally not complex, but you have to understand how the system responds to every action, that’s difficult for humans, I can’t imagine a text model doing great at that.
And I’m disgusted at colleagues just copy pasting our IP into random cloud-based LLMs, makes me want to bang my head on the wall.
I’m disgusted at colleagues just copy pasting our IP
“Our” IP? I think you mean “your employer’s IP”, right?
Maybe that is the true value of AI. Information gathering and social engineering
And I’m disgusted at colleagues just copy pasting our IP into random cloud-based LLMs, makes me want to bang my head on the wall.
If you’re talking about “intellectual property” and not an IP address, then that sounds like a security violation. At the very least, you need some vendor that can sign an NDA as one layer of protection, and a policy that dictates that nobody shares passwords, tokens, PII through these agents. Then firewalls, strict guidelines on where to use these things, make production access forbidden.
The problem is rushing into anything is a bad idea. Slow down, figure out how this shit works, implement policies, use what works, throw away what doesn’t, don’t radically change everything because of new shiny bullshit.
The people who jumped in headfirst at the first sign of LLMs and fired half of their employees? They all crashed hard and fucked themselves. That’s the lesson in what not to do.
Yeah it’s intellectual property, and we do have a copilot account, but would you trust something not 100%-local? When they run out of tokens they go to Gemini or GPT anyway.
It’s been incredible watching my company try to resolve the AI pricing situation, while desperately forcing everyone to use it even more.
- We’re paying 100k a month for AI, but deterministic n8n automations have to be carefully monitored so they “don’t cost too much”
- A project that feeds entire ticket JSONs into Opus 5.5 is an “incredible innovation that everyone should use,” but two minutes later, “please be mindful of token costs and don’t use it when not necessary”
- Everyone is forced into using Claude, but running airgapped local models forbidden
It’s ridiculous
Are we finally hitting the trough of disillusionment with AI?
That would be amazing!
Don’t kid yourself. I have a background in both statistical learning and programming, but because I’m not a fucking moron I’ve considered the entire AI boom a mass mental illness of idiocy and incompetence… but this will play out just like outsourcing.
Outsourcing was initially sold as THE solution and performed by the vast majority of business and political leadership — who are incompetent morons seeking nothing but quarterly KPI/cost improvements — which resulted in massive failures, bankrupcies, onshoring etc etc… But was ultimately refined and mastered to the point where most jobs could be offshored again.
Remember; Detroit manufacturing never recovered. 99% of manufacturing offshoring was ultimately successful on a multi-decade timeline. 15 years ago I worked in a business park that employed over 10k people for a single corporation. My role there was specifically onshored from a failed offshoring. Today, all of those “service-based” jobs have now been automated or offshored, and the business park is a ghost town that will most likely be demolished and replaced with a data center.
AI will fail short term due to the gross incompetence of political and business leadership, but most human jobs will be automated away over the next 2-3 decades regardless. They only have to be cheaper than humans. They do not have to be “better”. Tge fact that this will destroy capitalism and revert to fuedalism is irrelevent. Once we wire these models into robots and have them learning through interaction with the real world (which is cataclysmically psychotic), the automation of all human labor could take place within a single decade.
Don’t worry, though. We fundamentally can not control AI, so even if the fascists use it to mass murder billions, the most likely scenario is that they all die at the hands of their creation in the VERY near future.
The real problem for the large scale implementation of LLMs is: Even at the current, heavily subsidized pricing they do cost more than an average worker for low-impact tasks while having a worse failure rate. And while an LLM might cost a bit less in comparison to workers for high-impact tasks requiring specific education (which will not be true if AI corpos want to become cashflow positive), in this area you cannot accept hallucinations and misinformation at all because of the damage caused.
Outsourcing works because the jobs that were outsourced cost less, while still having a “human” error rate and type of failure (blatantly lying is a rare thing. They might sometimes get things wrong, but they are not actively misinforming). But here all bets are off - you can’t trust these things to handle things on their own, you always need someone to watch over them (and qualified to identify mistakes to boot!), which drives cost even more.
Character.ai filing for bankruptcy after adopting it so quick and feeding it to a loyal customer base is probably a sign the bubble is bursting.
Literally, they were one of the biggest and most successful AI resellers and they couldn’t stay afloat. Successful as in “people used their product a ton for free” though I think.
these companies get so much funding from investors they forget they have to make money and not just spend it on giving people free AI products.
Meta has scaled back plans to axe staff and use more AI agents after the company found employees were spending more and more time cleaning up technical and security errors, Reuters reported. And after thinning its workforce, especially middle managers, in a 2023 efficiency push, the company is expanding its managerial ranks again, LinkedIn posts from workers show.
Weird, middle managers are among the few things more worthless and wasteful than AI in the job space.
Why are we writing articles on what the biggest megacorps are doing with the technology? Meta is one of the most braindead corporations I’ve ever seen, wasting billions and billions of dollars on failed tech after failed tech. Microsoft can’t stop pushing their shitty Copilot, which isn’t even on anybody’s LLM radar. Anything Elon Musk leads consistently makes dumbass decisions that scare us all.
If anything, it just shows how stupid these megacorps are.
Copilot isn’t an LLM, it’s a tool to use LLMs and Generative AI. It’s very useful when people with no understanding of a topic speak vehemently about it because they can be tagged and easily ignored on other topics too.
Actually, the model itself is called Prometheus, but Microsoft has slapped their Copilot label on so many things that it’s hard to know what they are talking about.
I think a lot of these companies let go of team leads and just decided they could do without. They had the biggest salaries and likely no respect from the higher ups. In my experience with big projects this position is extremely important because it keeps the team on task and they are able to coordinate with other team leads. If you just have a handful of junior employees, they may be doing work, but it’s not actually moving the project forward. Often time these leads are still doing code review or even writing code (or the equivalent work product in their division), so it’s not like they’re just in meetings all day. They actually know and do the work, but they’re able to keep the group organized. I’m guessing those are the positions they’re rehiring.
The hyperlinked article in that section goes into a little more detail, but I’m guessing based off what I’ve been hearing from the industry.
Firms from Microsoft to McKinsey are limiting usage for employees, as the flood of new tools and rapacious consumption weighs on bottom lines. Others, such as Palantir, Booz Allen Hamilton and Nvidia, are restricting employee use over growing concerns about the security of their organizations’ data with AI companies, Reuters reported.
You don’t say!
Hah! Imagine you spent the last 3 years not looking at code and now are forced to do it by foot again. The horrors.
In my personal experience all the coders that I know that went full vibe… They were shit coders to begin with…
It really puts into perspective why my subjective experience with them is so much worse than a lot of people report…
I’m not a coder, but I do a bit on the side. I’ve done some vibe coding “give me a script that does the thing” but it’s scary because I have absolutely no idea what it does. Reading code at least for me takes a really long time and with ADHD it just doesn’t sink in. Writing it, I know what it’s doing. When the code breaks, I don’t just ask the model to fix it. It goes off the rails if I do. I need to understand why it broke before I can ask for a proper fix.
So - I agree with your assessment. I don’t think the alll in on AI vibe coding people had the kind of coder mindset in the first place.
C-suite leaders say when it comes to AI, there’s a sense that “we’re driving the truck and building it at the same time, and that’s a problem,” said Johnny C Taylor Jr, president and CEO of the Society for Human Resource Management. “If we could all rewind … we would have introduced [this] in a more thoughtful, slow way.”
If only there were people warning them ahead of time that diving head first into this was a bad idea. Hmmf. If only. Weird they went out of their way to publicly hype the results they now regret in response to absolutely no one at all anywhere ever saying the shit didn’t work.
If only somebody would have told them!
Worth mentioning here:
https://hermit-tech.com/blog/ai-mania-is-eviscerating-global-decisionmaking
Mark Ma, University of Pittsburgh professor and researcher on the AI layoff study, said leaders are focused on short-term performance, “which is why they tried to rush ahead by investing in AI and laying off workers”.
But companies’ share prices tend to dip after they announce AI-related layoffs, he noted, suggesting that “the stock market also does not think these will work out.”
