[phkahler]

This is why I doubt all the AI hype. These things are supposed to have PhD level smarts, but the above example can't reason about the problem well at all. There's a difference between PhD level information and advanced reasoning, and I'm not sure how many people can tell the difference (I'm no expert).

In an adjacent area - autonomous driving - I know that lane following is f**ing easy, but lane identification and other object identification is hard. Having real understanding of a situation and acting accordingly is very complex. I wonder if people look at these cars doing the basics and assume they "understand" a lot more than they actually do. I ask the same about LLMs.

[Workaccount2]

An AI smart enough to eclipse the average person on most basic tasks would even warrant far more hype than there is now.

[littlestymaar]

Sure, but it would also be an IA much smarter than the ones we have now, because you cannot replace a human being with the current technology. You can augment one, making her perform the job of two or more humans before for some tasks, but you cannot replace them all, because the current tech cannot reasonably be used without supervision.

[outside415]

a lot of jobs are being replaced by AI already... comms/copywriting/customer service/off shored contract technicals roles especially.

[Nevermark]

In the sense that less people are needed to do many kinds of work, they chat AI’s are now reducing people.

Which is not quite the same as replacing them.

[littlestymaar]

It's not even sure it will reduce the workforce for all of the aforementioned jobs: it's making the same amount of work cost less so it can also increase the demand for the said work to the point it is actually increasing the amount of workers. Like how github and npm increased the developers' productivity so much it drove the developer market up.

[Nevermark]

Most jobs have a limited demand. Because internal jobs are not the same as products in the marketplace.

Products and services typically require a mix of many kinds of internal parts or tasks to be created or supplied. Most of them are not the majority cost drivers.

You don’t increase the amount of software created by responding to cheaper documentation by increasing the documentation to keep your staff busy, or hiring more document staff, to create even more of the cheaper documentation.

You hire fewer documentation people and shift resources elsewhere.

Making one tasks easier is more likely to reduce internal demand for employees in that area. Very unlikely to somehow increase demand for it.

Unless all tasks get cheaper, or the task is a majority cost driver, and directly spills into obviously lower prices for customers for the product or service.

[littlestymaar]

No they aren't. Some jobs are being scaled down because of the increased productivity of other people with AI, but none of the jobs you listed are within reach of autonomous AI work with today's technology (as illustrated by the AirCanada hilarious case).

[trog]

I would split the difference and say a bunch of companies are /trying/ to replace workers with LLMs but are finding out, usually with hilarious results, that they are not reliable enough to be left on their own.

However, there are some boosts that can be made to augment the performance of other workers if they are used carefully and with attention to detail.

[anon373839]

Yes. “People make mistakes too” isn’t a very useful idea because the failure modes of people and language models are very different.

[littlestymaar]

I completely agree, that's exactly my point.

[shawnz]

Doesn't the Air Canada case demonstrate the exact opposite, that real businesses actually are using AI today to replace jobs that previously would have required a human?

Furthermore, don't you think it's possible for a real human customer service agent to make such a blunder as what happened in that case?

[Gnarl]

Possibly, a human customer rep. could make a mistake, but said human could correct the mistake quickly. The only responses I've had from "A.I" upon notifying it of its own mistake, is endless apologies. No corrections.

Anyone experienced ability to self-correct from an "A.I" ?

[littlestymaar]

> Doesn't the Air Canada case demonstrate the exact opposite, that real businesses actually are using AI today to replace jobs that previously would have required a human?

It shows that some are trying, and failing at that.

> Furthermore, don't you think it's possible for a real human customer service agent to make such a blunder as what happened in that case?

One human? Sure, some people are plain dumb. The thing is you don't give your entire customer service under the responsibility of a single dumb human. You have thousands of them and only a few of them could do the same mistake. When using LLMs, you're not gonna use thousands of different LLMs so such mistakes can have an impact that's multiple order of magnitude higher.

[xanderlewis]

You often have to be a subject expert to be able to distinguish genuine content from genuine-sounding guff, especially the more technical the subject becomes.

That’s why a lot (though not all!) of the over-the-top LLM hype you see online is coming from people with very little experience and no serious expertise in a technical domain.

If it walks like a duck, and quacks like a duck…

…possibly it’s just an LLM trained on the output of real ducks, and you’re not a duck so you can’t tell the difference.

I think LLMs are simply a less general technology than we (myself included) might have predicted at first interaction. They’re incredibly good at what they do — fluidly manipulating and interpreting natural language. But humans are prone to believing that anything that can speak their language to a high degree of fluency (in the case of GPT-3+, beyond almost all native speakers) must also be hugely intelligent and therefore capable of general reasoning. And in LLMs, we finally have the perfect counterexample.

[babyshake]

Arguably, many C-suite executives and politicians are also examples of having an amazing ability to speak and interpret natural language while lacking in other areas of intelligence.

[xanderlewis]

I have previously compared ChatGPT to Boris Johnson (perhaps unfairly; perhaps entirely accurately), so I quite agree!

[smokel]

> These things are supposed to have PhD level smarts

Whoever told you that?

[giantrobot]

Anthropic's marketing claiming high scores on supposed intelligence measurements.

[airstrike]

Having a PhD is not a requirement for being intelligent

[giantrobot]

Note that I am not making the statement that you need a PhD to be intelligent. Anthropic is claiming Claude 3 is intelligent because it scores high on some supposedly useful tests.

1. I don't think it's surprising a machine trained on the whole Internet scores well on standardized tests. I'd be shocked if the opposite was true.

2. I don't think scoring high on such tests is a measure of actual intelligence or even utility of the model.

[bbor]

LLMs are intuitive computing algorithms, which means they only mimic the subconscious faculties of our brain. You’re referencing the need for careful systematic logical self-aware thinking, which is a great point! You’re absolutely right that LLMs can only loosely approximate it on their own, and not that well.

Luckily, we figured out how to write programs to mimic that part of the brain in the 70s ;)

[nomel]

> Luckily, we figured out how to write programs to mimic that part of the brain in the 70s

What’s this in reference to?

[bbor]

The field of Symbolic Artificial Intelligence which is still (for now…) a majority of what is taught in American AI courses IME. It’s also the de facto technical translation of Cognitive Science. There’s a long debate between the two “camps”, which were called the neats (Turing, Minsky, McCarthy, etc) and the scruffies (the people behind ML).

The scruffies spent decades being shit on by the other camp as being lazy and simple-minded (due to a perception of “brute forcing” problems), only to find more success than most of them had ever imagined. I think anyone who says they were confident that ML-based NLP models could one day not only predict text, but also perform intuition, is either a revisionist or a prophet.

The whole Neat field got kinda stuck when we translated the low hanging fruit to symbolic algorithms (Simon & Newell’s Problem Solving being the most interesting IMO), but we had no way to test them. As another commenter alluded to, these systems lacked any “intuitive”(aka subconscious, fuzzy, approximate) faculties, so their high-level strategies could never work in the messy real world, mostly because it’s pretty impossible to definitively tell what information is relevant and what information isn’t to any given problem. This is called the problem of contextual “attention and selection”, and the problem more generally “the frame problem”.

Now that we have systems that mimic human subconscious intuition AND systems that mimic human self conscious reason, of course the next step is… declare complete victory and abandon the latter group forever as trash, apparently.

This is all a super biased take from someone who only got into this specific debate last year, tho I promise I do have some relevant credentials and have been working full time on this for close to a year. I strongly believe that LLMs are about to unlock the first (true) Cognitive Revolution.

[nomel]

Thanks! Do you recommendany good reads about this?

[kolinko]

Expert systems, formal logic, prolog and so on. That was the "AI" of the 70s. The systems failed to grasp real world subtleties, which LLMs finally tackle decently well.

[razodactyl]

Expert systems probably. Or maybe I read it backwards: it's implying that everything we see now is a result of prior art that lacked computing resources. We're now in the era of research to fill the gaps of fuzzy logic.