I see a similar trend with automation: it can do most of the job far cheaper than the specialist can, so short-term rationality says we should embrace it completely. But then the specialist goes away, and we don't have access to them anymore for the things that automation is bad at.
When you call in to customer support you can't talk to a human, because they replaced them with chat bots. When you need a good website designer because Wix won't cut it, they're out of your price range because the only ones left are working for large companies. When you want a new book, you can't find one written by a human, not because the robots have gotten as good as a good author, but because they've gotten good enough to spam the marketplaces so thoroughly that writing a good book isn't profitable anymore.
I suspect it's not worth pausing anything over due to the inevitable if we won't they will excuse and also because its part of the re-invention of business that goes with new technology. The things you want that have gone are now an opportunity for someone else to figure out. If I was an entrepreneur keeping my ear to the ground I'd weigh up starting a boutique produce delivery services, a human agency dedicated to dealing with companies on your behalf, a wix plugin market, a book review site etc.
But this just emphasizes the inequalities these automations create—poor people could walk to the neighborhood grocer and buy produce, but now that's reserved for people who can afford to pay for delivery or have a car and money for gas. Poor people could wait on hold like everyone else, but now only the rich can afford your agency. A small business could hire a designer, but now they're left buying stock plugins because being a small-time designer doesn't pay anymore.
The premise of this article isn't that automation is bad in and of itself, but that we are bad at considering the inequitable impact these automations have on our world.
You mention the web designer and that's again a good example - but not supporting the direction you want, I think. Web design is a field that's easy to enter (because does not require tens of years of experience or expensive in house server farms or expensive tools or license or large corporate team). It's not obvious for a busy small business to find a not-sold-out but good web designer but only because of near full employment. Probably that web designer should be more and more aware and agile using lower level tools (including generative AI and lower wage countries), but there is no reason supply and demand can't still meet each other in web design. Even in the age of Wix and such. They are different kinds of offerings which do compete with each other so that the boundary fluctuates but do not replace each other. Wix only displaces web designers that only do what Wix does. And even then only marginally so: a web designer can use Wix - just better than the customer - and have a service to sell.
I recently came across this, and I think it does a good job explaining the problem: https://www.youtube.com/watch?v=vQpUV--2Jao
However, it sells the boring usual cliché on how DG stores kill the local food stores: Dollar General sells the stuff they carry for far less. Well of course they do! Duh. While in the same video it is also shown that the local food store sells lots of things the Dollar General doesn't. Stop expecting to make a living by things you can't compete on. That just does not make sense. And the solution is NOT to get the other guy outlawed - that's tempting and seems easy but that only works in the short term. Shuffle prices around, carry other things, add services, change hours, do something! A potential issue could be that a tiny town may only be able to support so much retail activity. I don't think this one is in play. The stores of a small isolated town have a moat: their customers would have to drive at least 20 minutes, bus at least 40 perhaps, to get to the next store. And the small retail stores that persist all over Europe do NOT do so by selling the same stuff for more money. THAT is just dumb.
Another example shown in this video is 69 DG-equivalents within 10 miles of a downtown (all carrying the same things). There is plenty of buying power in that otherwise poor area. There should be space for other stores, probably for inexpensive fresh food and for more specialized stores. At least until Amazon figures out the fresh food issue. But obviously there isn't space for smaller stores trying to make a living selling the same things! Obviously?
For that matter, the video also describes the working conditions and salary of the DG-equivalents: minimum wage and employees who don't care. Another area where other stores can compete: by having a slightly more engaged workforce.
I would love to find videos discussing people who "actually did something rather than sit on their asses" and either it worked or failed. Pointers anyone?
It’s be callous to disregard that this probably also displaces some people from jobs but personally I think the difference of scale is important. In particular what scale can society and the economy actually absorb. I think there are good arguments that there are social and economic issues already starkly apparent that have been caused by similar shocks. Extrapolating from those experiences the likely impact of large scale AI shocks right now seems to be larger inequality and more social and economic issues than less.
The problem I have with tools like that is the technical quality of the solution (which Wix seemed somewhat reasonable in its output when I briefly looked around at Wix sites) and the social effects. Putting together a Wix site and buying or downloading/installing a theme does not make you a Web developer.
You're still getting to the end goal of a site, but without the knowledge of how it works and, should it break, how to fix it, it's more of an appliance to you than an environment to express yourself. (using royal you, not you personally)
It's strange to me that people want to be able to "do what the nerds can", but don't want to put any of the effort forth. I would love to be a great drummer or saxophonist, but I recognize it would take years of effort and practice to reach that point. I could compose tracks using synthesized instruments, and that's its own music skill I guess, but it's definitely not the same as playing the instrument.
Which jobs are you referring to that were lost? The dot-com era? As far as I understand, a lot of web design firms still exist, and tons more freelancers make cash through selling custom themes for clients or packaged themes on Wordpress, small client-oriented web apps, etc. And, given that much of the market is on Web tech still, a lot of the skills from 20 years ago translate pretty well and can now be used to make software within the browser. In my own efforts to learn how to make webapps, I've only really had to learn flexbox and grid. Most of my old knowledge has carried over, with some slight modernization tweaks as necessary. It's nice to never need to float a div ever again, haha.
Having good tools that create good work is a goal I think most of us can get behind. The challenge lies in building the tools and promoting healthy culture.
People, generally speaking, want outcomes. They don’t care how it’s done or what is used to get there. No one goes to the hardware store to buy a drill. They go there to buy holes.
The historical luddite example comes out - it seems - because the mechanization benefit is so massive to the owners that they convert very quickly. And an entire region relied on one industry. Little chance or time for the workers to convert even if they saw it coming, and they didn't.
In my mind, there are no downsides in allowing individuals more leverage. The real downsides are with the orthogonal problem of what productivity means in civil society.
High productivity through technology is perfectly human - we're tool-builders and tool-makers. But, productivity does not absolve us of our responsibility to each other. People who suggest we should not use tools often justify their point of view with assertions of cost born by others. I am suspicious of this argument because the opportunity cost of _not_ letting people use the best tools available could be much greater. For example, had we not developed tools to create mRNA vaccines, many lives would have been lost to Covid that were saved. We don't know yet what AI will enable, but I'd rather we find out what people can achieve than focus on muddying the waters between productivity and society.
A side comment here is that I am equally skeptical that leveraging copyright work as training data is fair use. if the objective is to retrieve original copyright work, then I think its fair use, but if the goal is to draw me campaign poster in the style of Shepard Fairey, I'm not that sympathetic.
If we use AI to replace the creation of copyright free derivative works, it isn't clear what the objection should be?
None whatsoever, and I agree with everything you stated but this is right here and right now. If the technological boundary of AI had already reached its speed of light then there wouldn't be anyone raising concerns.
I think it is highly possible that even getting close would yield really great outcomes for humans. Better medicine, better travel, better understanding of complexity.
We should get closer to the end before we theorize about the finish line. Remember the classic line about not needing more than 640k of ram.
Additionally as programming languages and toolkits have become higher level and more accessible there has been more demand for programmers.
So viewing it in such a contrived way is just not useful. If anything, the more automation we get, the more demand for tech workers there is because it's able to deliver increasingly more value.
Say your model makes 99% correct predictions, in 20-30 time steps that error rate drops to 73%, that is just not ok for automation. But real accuracy rates per time step are much lower.
Compounding errors even when they are small lead a system astray. I think there is no form of AI that achieves autonomy in any field. We're at minutes of autonomy, or seconds. Nobody can go on vacation and leave AI do their job.
On the other hand AI will open up new opportunities and markets. Demand will scale up to meet the new productivity level. We can always desire more, it's not a fixed sum game.
More fundamentally, LLMs are interpolating between and combining known skills. They don't do radically new discovery. Why? They train on human text instead of human text + world feedback. In order to surpass human experts AI needs labs, experiments or the ability to create its own experience.
I can only name AlphaZero, AlphaTensor and AlphaFold as superhuman AIs, and they have been trained on massive experimental feedback, not just text.
What?