ChatGPT will revolutionize the economy. We need to decide what that looks like

https://www.technologyreview.com/2023/03/25/1070275/chatgpt-revolutionize-economy-decide-what-looks-like/

by rntn • 3 years ago

42 77 3 years ago

technologyreview.com

bulbosaur123• 3 years ago

Why "we" need to decide? Can't GPT figure it out for us?

jstx13• 3 years ago

No. It can't pick your priorities for you.

QuiDortDine2• 3 years ago

When it comes to manage the insanely complex systems that form society, I think I would choose an incorruptible (if biased) AI before a fallible human being any day.

panarky• 3 years ago

The whole idea that "we need to decide" presumes that there is a collective "we" that's separate and distinct from the individuals and corporations releasing this stuff into the world, and that "we" participate in a collective decision making apparatus that has any effect whatsoever on what new capabilities get invented and released into a networked world where borders, boundaries and jurisdictions are all but meaningless.

bryanrasmussen• 3 years ago

wait, whenever the economy has been revolutionized in the past was there like a big decision making just prior to the revolution that made sure everything worked out nice and dandy? I guess I just don't feel the two things line up.

oliveshell• 3 years ago

I take your point, but this feels pretty defeatist to me.

I don’t personally believe that “we have not historically done this well, if at all” is a good reason to stop trying to manage the negative externalities of emerging technology— quite the opposite, in my opinion.

13years2• 3 years ago

However, "the future technology is completely unpredictable and we don't understand it" is probably a fairly good reason we will only make worse mistakes in that regard going forward.

Overtonwindow• 3 years ago

I think it will revolutionize the service economy by replacing front line workers with chat bots masquerading as human customer service.

jstx13• 3 years ago

Customer service chatbots are a revolution? Sounds like the most underwhelming thing ever.

m348e9122• 3 years ago

After my experience with AT&T front-line support where I tried several times to get my old AT&T iPhone unlocked (which is fully paid for), dealing with confusion and false promises of follow-up which never came, I welcome the relative competence of ChatGPT. My phone is still locked BTW.

neel89863• 3 years ago

And then we will need another set of frontline workers when LLMs starts to hallucinate.

ben_w8• 3 years ago

I've had a real life customer service chat session (years ago now) where they said their website didn't work on Mac OS while we were using it, and that it would be better if my phone was Android rather than iOS.

Back in the 90s, one of the monthly Mac magazines I read obsessively (because we didn't have internet at home yet) had regular stories of dumb things said by customer services; something like PC World sales clerk standing in front of a bunch of iMacs with Nanosaur running on them saying "yeah iMacs only do web browsing, you can't play any games".

Before I moved to Berlin, the dude in one of the phone shops near me insisted that iPhones couldn't work with those insert-your-phone-here plastic versions of Google Cardboard.

It's not that you're necessarily incorrect — when ChatGPT was new, I saw someone dismiss it as "mansplaining as a service" — but I don't know how much the companies really care.

It's not like customer support is a prestigious role.

grayfaced• 3 years ago

and chat bots won't ever "forget" to try and upsell you on the things corporate expects. Every cashier will be advertising.

throwyawayyyy• 3 years ago

Who knows! Maybe? The email that's written by ChatGTP, but no human actually reads it because at the other end it's summarised by ChatGTP too: maybe that'll be it. LLMs almost but not quite cancelling out the meaningless guff they add to the world.

It can both be true that a) LLMS and their successors will be the sort of huge multiplier for intellectual pursuits that _machinery_ is for physical work; and also b) they are AWFUL for education, and will result in the general population losing skills.

mg2• 3 years ago

   Will ChatGPT make the already troubling income and
   wealth inequality in the US and many other countries
   even worse?

It is an interesting question if the concept of "wealth" will survive in a world where all productivity is done by artificial intelligence.

As soon as computers surpass humans in every aspect, there is nothing left for humans to do. So the only reason for wealth inequality would be inheritance. But why should the poor accept this inequality? Will AI protect the wealth of the rich who inherited it? Why should it?

Giorgi2• 3 years ago

why should AI serve humans at all?

mg2• 3 years ago

I would expect it to do that for the same reason that we humans feed and protect endangered species.

ben_w8• 3 years ago

Why should we create one that doesn't?

At least, that's the ideal answer, but we don't really know what we're doing.

This isn't like giving birth to a new human, where the new mind has a sense of self and individuality by default and it would be immoral to take it away, works of artifice begin their existences as we construct them; and right now we don't even know how to tell if an artificial intelligence has an inner world, much less how to imbue an AI with one or prevent it.

QuiDortDine2• 3 years ago

But why should the poor accept this inequality? Will AI protect the wealth of the rich who inherited it? Why should it?

Why do they accept it now? The best predictor of financial success is the parents' wealth [1]. Society as a meritocracy is a myth, yet we all plow on. I don't see what would change this.

1: https://cew.georgetown.edu/cew-reports/schooled2lose/

gsatic• 3 years ago

We surpassed what plants and ants do long ago, have they stopped doing what they do? Do we protect them?

ben_w8• 3 years ago

I, for one, am not thrilled by the idea of being analogous to an AGI's version of a bonsai tree.

pixl974• 3 years ago

How many species have humanity driven to extinction?

hypertele-Xii12• 3 years ago

No chat bot ever published in the past decades has revolutionized anything. They have all been the subject of intense, temporary hype, and then completely forgotten. I see no reason why this time should be different. It literally lies and makes shit up! The economy isn't built around lies and making shit up, is it?

djaychela• 3 years ago

The economy isn't built around lies and making shit up, is it? I think a significant portion of it is made up of BS and hype. And there's a lot of that around AI, particularly in the last few weeks. Big players look to be going all in on it.

I think dismissing it as simply a chat bot is missing what's actually happening. There's much more to ChatGPT than merely a chatbot, even with the hallucinations.

hypertele-Xii12• 3 years ago

It's literally a bot that chats. What is there more to it?

Giorgi2• 3 years ago

It only makes shit up once it does not have enough information, it will however improve as it is connected to the internet, ALSO - it can do accounting, financial audit and coding better than humans.

Even right now, given it is properly implemented, it can leave millions of accountants and auditors without jobs.

But I agree, it wont revolutionize anything, it will automate EVERY SINGLE JOB that is dependent on human's intellectual input and eliminate billions of the jobs, which will lead to global hunger and reduction of human population.

hypertele-Xii12• 3 years ago

Absolute bullshit. Show me one software coded by ChatGPT.

neel89863• 3 years ago

Yes program that hallucinates and makes up numbers is going to be the perfect accountant /s

snozolli• 3 years ago

it can leave millions of accountants and auditors without jobs.

I think those are two examples of exactly what cannot be replaced yet. Their jobs are specifically about ensuring that rules have been followed precisely. Current AI sounds confident, but still makes obvious errors that would be completely unacceptable for accounting and auditing. Same as software engineering, you currently still need a human reviewing the output and verifying that it's actually doing the right thing, and without side effects, poor handling of errors and edge cases, etc.

huuhee3• 3 years ago

What's the point of automating all that work, if those billions of unemployed will then be too poor to buy your product/service?

If that really happens we get universal basic income, or massive government employment programs. Otherwise the economy collapses and the ultra rich lose money.

codingdave2• 3 years ago

The economy isn't built around lies and making shit up, is it?

Yeah, it kind of is. Fake it 'til you make it. Vaporware landing pages to validate products. But that isn't what we need to focus on.

ChatGPT is more than a chatbot. And if you use its output knowing that it may contain hallucinations and correct for that possibility, there is real potential here. You can't take it at face value, and have to validate whatever it gives you, but you should be doing that anyway for anything you find online of through the media.

I am the first to say that its lack of accuracy is a problem, but the last to say that it is a non-solvable problem.

hypertele-Xii12• 3 years ago

If you have to check online that ChatGPT is correct, then what exactly does it add but an extra step to just getting the correct info from online in the first place?

gremlinsinc5• 3 years ago

The economy isn't built around lies and making shit up, is it?

I see you've never talked to a salesman. Try going to a time-share sales presentation. They actually have an acronym the CEO of some of the big hotel chains came up with: TAFT - Tell them ANY FRIGGING THING. (really it's frigging because: corporate ethics or something).

albertopv2• 3 years ago

Not going to happen for many, many jobs, if only for tons of private data not available to ChatGPT, e.g. highly sensitive code not on github.

ben_w8• 3 years ago

Private data yes, but I don't think code is that much of a secret special sauce. Apart from anything else, too many engineers moving between different roles and sharing what they learn.

Even with medical data, which I at least hope is much better protected, some of that ended up in the training sets used by the current bunch of text-to-image AI.

albertopv2• 3 years ago

Yes, but probably that code won't train the AI. Also, I think medical data is less private than a lot of code.

neel89863• 3 years ago

How exactly will ChatGPT will revolutionize economy. The best it is doing is summarisation of search result. And search engines exist from 2000s. Yes instead of visiting wikipedia now we get the result directly. At max it can decrease the number of links to click. It is going to change the economy as googles answer box has changed economy, On the other hand the way they hallucinate you still need to double check any thing they tell you. So we can not automate the decision making. Best of luck to you if you want to replace a lawyer or doctor by LLM where they have accuracy of 80%

pixl974• 3 years ago

What is doctors and lawyers failure rate?

avereveard• 3 years ago

even if it doesn't provide that much intelligence, it provides intelligence at a surprisingly low cost and intelligence that has almost zero barrier to entry

this will do to text task and machine learning task the same thing that squarespace did to web development.

that is it will not replace the high end, but suddenly the long tail of businesses will be able to afford the entry price and start getting benefits that would have been totally out of their price range before.

olivermarks• 3 years ago

Surely not 'chatGPT' but AI & AGI?

ChatGPT is ephemeral.

whyenot• 3 years ago

What is so interesting to me about generative AI is that the jobs that are the most difficult to automate are trades like plumbers and electricians. The group that is the most vulnerable is knowledge workers. It’s a lot easier to replace a person with AI when much of what they do is virtual — on a computer screen. Trade jobs are often looked down on by the privileged class who went to some costly and exclusive college. Will that change? There is a lot of experience and training needed to do a trade job well. In fact (having done both), I think in many cases they need more training that a job in tech requiring a college degree. This is going to be really disruptive and I’m very curious to see how it all plays out.

machinekob• 3 years ago

How much plumbers will we need in a future?

I'm pretty sure that it'll end with oversupply on traders (especially in stagnant economy) if AI will cut most of the knowledge workers the world will have to move into some sort of utopia/dystopia and people will just don't have to work at all or it will be 0.1% vs 99.9% scenario but with brainwashed people cause of social media and maybe AI censorship.

But its worst case scenario in my mind AI will make jobs just more specialised and output of knowledge worker a lot faster.

sublinear• 3 years ago

AI will make jobs just more specialised and output of knowledge worker a lot faster.

If the user is skilled enough at prompt engineering to get good answers (i.e. code that compiles and does what they want) the answers can still be objectively worse than search results for stackoverflow or github in many cases.

Seems like the current state of chat bots is that it's slower than either making a regular search or simply just knowing what to write in the first place. Shouldn't be a surprise, but a mid to senior engineer doesn't spend their whole day googling stuff to copy and paste. They read the docs for about 5 minutes and then get things done in the precious couple of hours they have between meetings. I don't think it gets much faster than that. I sincerely doubt the nontechnical people in the meetings would get a good enough answer from ChatGPT that fast and they certainly wouldn't want to slam their head against a wall learning everything the dev already knows, no less in a higher level and imprecise "natural" language.

afavour• 3 years ago

I’ll openly admit that I’m a cynical person but I’ll believe economic revolution when I see it.

The advent of the personal computer also caused lots of huge proclamations of computers doing our jobs while we sit back in sun loungers sipping drinks. Yet somehow, here we are.

The people in charge of the AI systems are the same (class-wise at least) as those in charge of the economy so there’s no incentive to rock the boat. I expect AI to be something similar: it’ll change the way many jobs work, hopefully for the better. But people will still be working in sweatshops making clothes because it’ll still be cheaper than the alternative.

gremlinsinc5• 3 years ago

But do you equate AI to the advent of:

A. The internet?

B. Electricity?

C. The wheel?

Personally, it's somewhere between the wheel and electricity in my opinion, or a combination of the industrial revolution/assembly lines and electricity.

How much did the wheel revolutionize society?

the88doctor• 3 years ago

The economy serves humans as the end users of goods and services. Even if an AI can learn to design a house that is more earthquake-proof than any human design of the same cost, the AI can't decorate the interior without understanding human psychology both at the general level of the average human and the specific level of the human who will be using the house.

I didn't realize how much happier I was when I worked next to windows (as opposed to working in the interior of a room) until I was in my mid twenties. How would an AI know to tell me to design my home office with a big window when I didn't even know that was necessary myself?

The AI couldn't have learned from me (since I didn't consciously know that), and the AI couldn't have learned that sitting next to windows generated happiness for itself either. That means some other human being would be necessary to input that as a design constraint into the AI's architecture program.

The inability of AI to experience human feelings for itself means it will have to rely on humans to help it provide economic goods and services that are maximally useful. That means there will be a role for humans in the economy for quite a long time (possibly forever).

gremlinsinc5• 3 years ago

The pursuit of human-like capabilities, Brynjolfsson argued, has led to technologies that simply replace people with machines, driving down wages and exacerbating inequality of wealth and income. It is, he wrote, “the single biggest explanation” for the rising concentration of wealth.

I think, assuming it doesn't kill us - the faster we get to an economy where only 5-10% of the people are employable, the faster we get to post scarcity, and MORE equality/egalitarianism (if accelerationism doesn't lead to the opposite outcome, walled cities for elites, wastelands for everybody else) ... but the transition... it's going to be bumpy/painful.

Part of me is worried, part of me says rip the bandaid off, let's get to UBI, and other programs --as people can't stay rich or get richer if nobody can buy products. What's the point of better productivity at Ford or GM if nobody is buying cars anymore?

The world of the next 20 years, is going to change as much as the change from before electricity to the 1920s when it was mainstream. We can't really stop it, even if we tried - other actors good/bad are going to be using AI, it's part of our lives now for better or worse.

dusted• 3 years ago

"But the fisherman simply went about his daily routine, casting his nets and patiently waiting for the fish to bite."

Yeah, revolutionary. (That's GPT-4, by the way)

ChatGPT will revolutionize the economy. We need to decide what that looks like

Related Stories