Like the Romans used to drink from lead cups. Something like that.
(it appears that the moderator is removing any replies with which he disagrees. It might be a piefed issue. But all the removed replies are top level and upopular. Stranger coincidences have happened I suppose.)
Using AI. Aside from all its ethical, environmental, and legal issues, I’ve been shocked to see the cognitive decline in folks using AI regularly. Not only do they forget how to do things like write on their own or do Internet searches by themselves, they even lose higher skills they’ve spent years acquiring.
Was talking with a developer the other day, and he said he’d almost had a panic attack recently when trying to do something during a recent outage when he was trying to code something relatively simple and couldn’t even think about where to begin. He said he’d cut back after that, but honestly not sure he will. Every frequent AI user I know seems to be on a junkie trajectory and unable to cope when their supply gets interrupted.
Every powerful technology brings with it the risk of atrophy. Cars, processed foods, language.
What irritates the ever loving crap out of me is that all the search engines have gotten so bad that you have to use AI to get the answers now… And…! AI scrapers are so hard on bandwidth and servers housing that all the old how-to tech blogs by actual smart people have packed their bags and left. I used to find loads of great self hosting guides by doing a Google search for “Proxmox npm setup guide” and clicking one of the top three results. Now: sponsored, slop, slop, shilling paid product, old Reddit post, etc.
Yeah, it’s not perfect, but I’ve taken to using https://noai.duckduckgo.com/ - it’s not as good as Google used to be, but it’s pretty functional for most things
my bro in tech was doing this quite alot, and he was depending on it to answer every question, and it wasnt giving him useful answers. i dont know if vibe codes with it. i think alot people who had been laid off has been have developed a dependancy on it.
The latest PISA study shows that students who heavily use LLMs are a full year behind their fellows who don’t.
This overstates the findings by a lot, and borders on misinformation.
URL to OECD Summary
Edit: I concede that “heavily use” exonerates you, but your comment still feels a little disingenuous to me.
As a heavy AI user I wonder if there’s a difference in how it affects people who can literally have the AI do some amount of their actual work for them compared to a tradesperson like myself who can only rely on it (to a certain degree) for instructions but it won’t hang drywall for me or replace a faucet. I still have to do the actual work myself. The AI is just a tool.
Here’s a pretty typical question I ask it. I can’t really see how that is cognitively detrimental to me nor how I would even google that, and why.
spoiler
In that case, you are asking it a question you don’t know the answer to, and the answer could be wrong. You might have the knowledge to know when the output isn’t factual and reprompt it, or you may not always know when the output isn’t factual; the potential detriment to you probably isn’t so much cognitive, it’s potentially wasting your time, money, supplies and labor on an incorrect answer, or worse, leaving the customer to deal with bad service or worksmanship from an incorrect answer.
Sure, but the starting point is that I already don’t know something and need to look it up one way or another. Yes, the AI answer might be wrong but the guy on the YouTube video or online forum could equally be wrong. That is not something unique to AI.
I’m also obviously not just blindly trusting what it says. When it recommends me a product for example I don’t just go buy it without a question. The AI is just pointing me to the hopefully right direction for me to figure it out but the ultimate decision still remains mine.
A good recent example:
“I need to paint plywood that gets wet regularly, should I use paint A or paint B?”
“Neither, you should use paint C”
*Looks up paint C*
“Yes, I definitely should use paint C”
Even reading that (atrocious) output is going to influence how you write and speak, because humans learn language via mirroring and imitation. Taking the output as truth also means you are trusting something that is unreliable and easy to influence (I know many companies are deliberately writing to manipulate LLM output).
What’s so atrocious about it?
That feels like a reasonable use of ai.
Sure, you could go searching for all that info manually, but it would take a decent chunk of time, so you probably wouldn’t have bothered…
Always feels like “reasonable use of A.I.” is kind of vague and hard to understand for most people. Since it can somewhat follow instructions I wonder if it makes sense to force it to be used in a reasonable way? Like instead of putting the blame on people for becoming brain dead, why not tell the LLM “if someone asks you do to all the work for them, say no” or something