Basically a varient of "AI is so impressive and knowledgeable, except for that one thing that I am an expert of and were it keeps hallucinating complete bullshit all the time. Weird."

Which explains why C-suites push it so hard for everyone

That's also why the billionaires love it so much:

they very rarely have much if any technical expertise, but imagine that they just have to throw enough money at AI and it'll make them look like the geniuses they already see themselves as.

AI has been excellent at teaching me to program in new languages. It knows everything about all languages - except the ones I'm already familiar with. It's terrible at those.

and all the things we aren’t experts in, we’re unqualified to be the evaluators of the AI’s output

Ignorance and lack of respect for other fields of study, I'd say. Generative ai is the perfect tool for narcisists because it has the potential to lock them in a world where only their expertise matters and only their oppinion is validated.

This is why leadership loves it. They don't know shit about fuck.

Gell-Mann amnesia

IDK about that I'm a professional slop maker and I think it could replace me easily.

Funnily enough, all my engineering professors seem to encourage the use of genAI for anything as long as it’s “not doing the learning for you”

What’s funny is that there’s basically no practical use for GenAI in engineering in the first place. Images like technical drawings need to be precise and code written for FDM/FEA etc. needs to be validated by some kind of mathematical model you derived yourself.

They say “it’s a useful new tool” and when I ask “what is it useful for” they typically have no answer besides “writing grant proposals” lol

There are lots of useful applications for machine learning in engineering, but very few if any practical applications for genAI.

It's why managers fucking love GenAI.

My personal take is that GenAI is ok for personal entertainment and for things that are ultimately meaningless. Making wallpapers for your phone, maps for your RPG campaign, personal RP, that sort of thing.

let's not confuse LLMs, AI, and automation.

AI flies planes when the pilots are unconscious.

automation does menial repetitive tasks.

LLMs support fascism and destroy economies, ecologies, and societies.

This actually relates, in a weird but interesting way, to how people get broken out of conspiracy theories.

One very common theme that's reported by people who get themselves out of a conspiracy theory is that their breaking point is when the conspiracy asserts a fact that they know - based on real expertise of their own - to be false. So, like, you get a flat-earther who is a photography expert and their breaking point is when a bunch of the evidence relies on things about photography that they know aren't true. Or you get some MAGA person who hits their breaking point over the tariffs because they work in import/export and they actually know a bunch of stuff about how tariffs work.

Basically, whenever you're trying to disabuse people of false notions, the best way to start is always the same; figure out what they know (in the sense of things that they actually have true, well founded, factual knowledge of) and work from there. People enjoy misinformation when it affirms their beliefs and builds up their ego. But when misinformation runs counter to their own expertise, they are forced to either accept that they are not actually an expert, or reject the misinformation, and generally they'll reject the misinformation, because accepting they're not an expert means giving up on a huge part of their identity and their self-esteem.

It's also not always strictly necessary for the expertise to actually be well founded. This is why the Epstein files are such a huge danger to the Trump admin. A huge portion of MAGA spent the last decade basically becoming "experts" in "the evil pedophile conspiracy that has taken over the government", and they cannot figure out how to reconcile their "expertise" with Trump and his admin constantly backpedalling on releasing the files. Basically they've got a tiny piece of the truth - there really is a conspiracy of powerful elite pedophiles out there, they're just not hanging out in non-existent pizza parlour basements and dosing on adrenochrone - and they've built a massive fiction around that, but that piece of the truth is still enough to conflict with the false reality that Trump wants them to buy into.

I agree with this. I tell people to ask it questions about things they know about. Then, when they see how many errors it makes, ask them why they assume it's any better on a topic they don't know about.

You see the same effect in journalism. News stories seem pretty authoritative until you read one about a subject you know.

So Gen AI is like Dan Brown, the more you know about the subject the more it sucks

AI only seems good when you don't know enough about any given topic to notice that it is wrong 70% of the time.

This is concerning when CEOs and other people in charge seem to think it is good at everything, as this means they don't know a god damn thing about fuck all.

AI is very useful for our business. (/s) We get paid by the hour and our billing hours have exploded simply by being hired to clean up after "vibe coder's" shit.

We deal with massive amounts of data and AI can't write optimized code for shit. Hunting down system slowdowns thanks to bad code is a constant gig with our clients.

The breadth of knowledge demonstrated by Al gives a false impression of its depth.

they’re both wrong, and they’re both right

an AI can create concept art for a writer to better visualise their world to generate ideas in a pinch, but it shouldn’t ever be what you use to show anyone else: you still need real concept art

an AI can also create writing for their art so that they can flesh out a back story to make their visual art more detailed, but it’s not going to write anything that you’d want anyone to read as a book or act in for a movie

both things can be used for the described purpose, and both things are inadequate for quality output

we’ve had this juxtaposition for a while: “redneck X”… they’re scrapped together barely functional versions of the thing you’re trying to do, on the cheap, with home-made tools. you wouldn’t sell it, but it’s kinda fine for this 1 situation with many many asterisks

professionals often don’t like when someone can hack together something functional because they know the many many places where that thing falls down when you talk about long-term, and the general case… but sometimes a hack job solves a specific problem in a specific situation for a moment for cheap and that’s all you need

(just don’t try it with electricity or your health: the consequences of not understanding this complexity is death… of course ;p)

there's professors of comics?

Orrr...hear me out, this is gonna sound wild... Or we don't believe that this debate is even one we need to have until we have actual fucking AI, which machine learning slop IS NOT. And seeing the kind of morons hyping "AI", chances are, mankind will never develop true AI because the funding goes to the morons screaming loudest, instead of actual experts slash scientists.

I just focus on the parts of what I do know that AI can help me with, not try to say AI can replace other people, but not me. That's some dumb shit.

Comics professor?

So it's perfect for people who are shit

The only good AI I've come across is the one I use for denoising cycles renders in Blender3D, as that's something that a human cannot reasonably do.

That's the only scenario something like AI has any use as a "tool"; doing things humans cannot reasonably do.

I'm a programmer I think both are an art and can't be replicated by ai well. Sure you get an acceptable pic, you may get something written well (okay stretching more here I haven't read anything by ai that make me think that but it's been minimal so giving some leeway), but human art is its own quality.

Just reminded me of a bit of the Dune novels, people were putting rocks out to be sandblasted by a dust storm and selling as art. I guess I agree with Duncan on that.

Wait art needs the emotion bit, huh probably to mean more than generated stuff. Another realization but good to understand. I do think the human component is necessary... Until it isn't but not today.

I use Generative AI at work because I know it's being tracked. I've offered examples and suggestions about things I've done with it.

I've then outright referred to it as the World's Worst Intern. Sometimes it does the right thing, but you always have to check. Sometimes it says it's going to do the right thing, but actually does something different. Sometimes it does completely the wrong thing.

So I have it do the things that I can do -- rote steps, easy changes I can explain faster than I can type, bulk renames or code cleanup that the compiler can validate -- but not the things I don't know if I can do. I trust the compiler, I don't trust the code it wrote. I'll use it to write the first draft of documentation based on the steps I took, but I'm editing it and expanding it.

It's not smart. It's not intelligent. It can kind of do things as long as you're willing to let it flail for a while or to spend the time checking it's work. It's the World's Worst Intern.

So the only real business model here is for people to be able to produce things they are not qualified to work on, with an acceptable risk of generating crap. I don't see how that won't be a multi-trillions dollars market.

Being honest, I don't like using AI for much of anything. I have been encouraged to use it at work, but aside from rubber-ducking with it to plan out my own strategies, it's useless.

At home, it's a chatbot. I initially used it like I use random names or locations for writing and RPGs. Now, I stick to Donjon and a few others. The biggest thing I ever had it successfully do was help me construct puns I couldn't quite figure.

Hot take: it's reasonable for a comics student to use AI for script-writing and for a screenwriting student to use AI for concept art, not because machine can generate meaningful artistic work at these fields but because these are not the fields they are trying to learn.

In a way, this can be used to level the field. The comics professor can use the same LLM to generate scripts for all their students. It'll be slop script, but the slop will be of uniform quality so no student will have the advantage of better writing and it'd be easier to judge their work based on the drawing alone.

And even if AI could generate true art in some field - why would it be acceptable for a student to use it for the very field they are studying and need to polish their own skills at?

In my experience everybody (myself included) is prone to the Dunning-Kruger Effect in domains outside their expertise.

It doesn't mater if you're a outstanding expert in any one domain: you just look at a different domain and go "yeah, that looks easy".

I'm actually a lot more generalist than usual because of my personality and still have that same tendency to underestimate the complexity of different domains, but because of being a generalist I sometimes for one domain or another go down the route of genuinelly practicing it professionally, and one or two years later I'm invariably thinking "This shit ain't anywhere as simple as I thought!".

And, lo and behond, generative AI is just about good enough to handle the entry level stuff in a domain - the ultimate Junior Professional (not even a very good one) with just about enough "competence" to look capable for domain outsiders or even hobbyists whilst at the same time being obviously mediocre for domain experts.

As most people don't really think about their own knowledge perception in these terms and thus don't try to compensate for it, the reactions described in this post totally make sense.

I am 100% positive ai cannot take my job or replace me. In related news, I'm the only person in the world who makes a very specific thing.

So basically all these teachers are myopic assholes, is what I'm reading.

AI is just... its a broken mirror, a poisoned forbidden fruit.

It just brings out the worst in everyone and everything.

AI absolutely can be used for the work they know best, it's just that the individual using it will be the only one who knows how to use it correctly and everyone else will just be making slop.

Everyone assumes their expertise is special.

It's the dilbert approach

That's a really good point.

It's the same with all automation. Processed food is not as good as homecooked food, tailored suits are better than massproduced ones.

Facts!!

AI will never be able to provide a suitable replacement.

Don't become as delusional as the grifters. Generated content is already a suitable replacement for lots of things. It's not so much about the quality of generated content (which continues to improve) as much about easily replacing the worthless bullshit that we're forced to produce (eg. cover letters, "art" serving capitalism, etc). The system is already built on fake nonsense. Generated content has always been a great fit for this system. The punishment of workers is just another bonus.

This has also been coined the Gell-Mann [Amnesia] effect and is perhaps a kind of corollary to the Dunning-Kruger effect: incompetent people fail to recognize competence.

Truly intelligent people respect the work of professionals and experts in other fields. Or maybe, this is even fundamentally a respect problem.

bingo! it also explains why tech bros consider AI good for everything - they are not really good at or familiar with anything themselves.

on a serious note: to put this in a specific example i will use programing. management/PO/business will have a vague ideas about what needs to be done. they have to explain them to someone in plain human language and let them figure out details of how to turn it into an algorith that does that. that someone has to fill in all the blanks, make many decisisions about unspecified details and implement it. but, in addition to that, they also have to make sure the code is delivered on time, on budget, it will remain easy to maintain and doesn't break anything important. those later details are not something the original stakeholders (business/management/POs) normally have to deal with. it's someone's else's responsibility. it's just annoying to them that they have to deal with another temperamental human and that human can take weeks to make even "simple changes" to their code (because they have to care about all those other things like maintainability)... so when a salesman comes over and offers them they can instead explain the problem to a chatbot and get the code in minutes the proposition sounds irresistible!

And managers think it's good for everything except being executives and management. Except that's probably the one thing AI can do just as well as a manager who thinks AI is good for a bunch of stuff.

Finally 1 good, sensible post here!

Chances are very high that AI will indeed one day be good enough.

For know I wouldn't take it for any more than a slightly more capable rubber duck that can be useful as an assistance tool. Not a replacement.

I don't know. I worked as a software dev for 5 years and have a BS in CS. I've transitioned to ecological work. The progress I've seen with Claude Code specifically has convinced me that even moderate gains in intelligence will lead to the functional replacement of several data entry and junior programming jobs.

It's true that people overestimate LLM performance in other domains. But, software is easy to generate synthetic data for, objectively testable by adversarial models, etc.