It's very simple: Lots of people have those jobs, and they either like those jobs or have no confidence that having to find new jobs won't end up as a net loss for them.
The problem isn't that tools make jobs easier, it's that when jobs get easier the benefits accrue to the owners of capital - by erasing jobs - instead of to the people doing the jobs - by allowing them to create the same amount of value for customers/society in less time, improving their standard of living.
Luddites weren't against automation. They were just against automation being used against them by others, rather than for them, by them.
Because people need to eat. If you make their jobs disappear, they need to find news jobs, and that's not necessarily an easy transition. In the short term, and people's lives are short, it can be painful. That, at least, is the understandable motivation.
> Any tools goal is to reduce work.
I'm not sure that's entirely true. Yes, a dishwasher reduces the work you have to do washing dishes. That is true. But is also creates an entire economic ecosystem to support the existence of that dishwasher. You have the entire supply chain, the manufacturer, the repairmen, etc, etc, etc; the dependencies are vast. Technology is always thoroughly embedded in a living culture. It is a product of that culture and an instrument of that culture. It is just a dead and empty husk without a culture to breathe life into it and make it what it is, like a severed finger. A 21st century smartphone beamed to 5th century Arabia would be worthless.
That's also why all sorts of initiatives in places like Africa have often been boneheaded. Someone from an NGO decides that farmers in Africa need tractors, or a water pump, or whatever. So they bring these things to a remote rural village somewhere, thinking that with this tractor, they'll increase agriculture yields and free up children from having to work the fields so they can go to school, and with this water pump, they'll relieve these people of drought. They show the villagers how to drive the tractor, how to turn the water pump on, and so on. So they use these things for a couple months. Something breaks. Nobody knows how to repair the tractor. Nobody has the parts needed to repair the tractor. Nobody has the tools to repair the tractor. Things go back to the way they were and the tractor is left to rust behind a shed somewhere. You need an economic ecosystem (and economies are a part of culture and a reflection of culture) that supports the existence and usefulness of the tractor and that water pump.
But transitioning to that ecosystem can be uncomfortable, even painful, especially when done too rapidly.
There was a big kerfuffle in the D&D community the other day about an artist using AI art tools to augment his drawing in one of the new books. Guy had his whole process documented and just used AI as a final postprocess step. I can’t help but notice how bland the art in D&D books looks vs what I’m seeing daily on AI art groups.
I’ll be glad when the dust settles as it’s really limiting what’s possible since Steam is also outright banning AI art tooling.
The reason AI-generated art looks better to you is because it is cribbing from better (human) artists. Maybe D&D should just hire better artists directly?
Honestly, there are so many places where art absolutely is terrible. If AI generation can make higher quality art more accessible to more areas then it is a huge win.
Our little Dungeons and Dragons campaign has absolutely exploded in quality and storytelling because our DM has been utilizing Midjourney for character tokens for the bad guys. I spent an hour or two on each person in the game’s character portrait with astounding results. It has been like night and day.
There is a 0% chance I’d pay the price this would have been a couple years ago. I can’t even fathom how much this quality of art would cost, probably tens of thousands of dollars.
Before the DM would just hunt and peck around for online images that sorta matched what he was going for, mostly rudimentary pencil drawings with some color fills.
To me the dream is saying “Elmo is king of a Mad Max style castle, please write a story and generate art assets” and in front of you on the tabletop your Apple Vision Pro tiles out a whole dungeon, does voices, you can go in and inspect and interrogate the pieces and make tweaks to get it just right.
> our DM has been utilizing Midjourney for character tokens for the bad guys
I don't think, in all my years of role playing, I've _ever_ seen an image of one of the NPCs in our stories; at least outside of the rare store-bought adventures. Or, at least that I can recall, an image of _anything_ in our campaigns. Everything was verbally described.
I wasn't even aware people generated pictures for home-grown adventures/campaigns.
When I was playing (about a decade ago) we made heavy use of static images. A few people drew their characters, others pulled images from online, and our DM would always flash an image of whatever boss we were about to encounter. These were all intended to be rough approximations, and were typically shared with deviations, e.g. "my character looks like this guy, but he uses a staff instead of a longsword, and he has a lot more hair!" My girlfriend plays with some artist friends, and they regularly doodle their character's escapades.
Art is nice. Customized art is nicer. Also, it's interesting that (as you note) it's far from universal. My group would likely have enjoyed using AI art for our campaigns, but my girlfriend's groups have completely rejected it.
That is pretty cool, thanks. I haven't been able to roleplay in a number of years but, if I can get back into it, I will certainly be making an image of my character :)
> I never understood the longterm logic of protecting jobs by not using tools.
Ever read thelastpsychiatrist? The website is dead now (and he may be too, for all I know), but he was on the money when talking about why a specific amount is paid by the state to the unemployed - it's the exact amount of money that would stop the idle masses from burning down the city.
You don't understand it now, but you will if we had a transition to completely automated driving in a short period - those millions of unemployed drivers, that are now never going to be employed, are a risk to the city.
> it's the exact amount of money that would stop the idle masses from burning down the city.
That seems awfully pessimistic. There's a couple different reasons
- To stop the masses from uprising (as noted)
- Because the health of the society as a whole depends on the health of the people in the society. And sometimes you need safety nets for the people in that society, because people can't always do for themselves. Fire departments, unemployment, health care, education; these all fall into that category in one way or another. Some places provide all of these, some provide none, and most provide somewhere in the middle. In the end, we provide for those that can't because it's better for the society that we live in if we do. And we are all served by a better healthier society.
...which works out to roughly 3-7 million people. Unless you want to pay them UBI, you're going to have to help them find and transition to other categories of jobs.
I used to drive a taxi. Now I write code. I paid for my own training. I'd expect to find that self-help is the most usual way to transition into other jobs.
The reality is, this is just not a functional pathway for the majority of people. So it’s not a solution
It might be possible for some minority that has a particular penchant for this type of work, but neither the jobs nor the training exist at such scale that would be able to absorb the entirety of the driver workforce into the “write software for automated driving“ job.
That's 1-2% against the interests of almost everyone else that would benefit from driverless cars. Their plights matter, but at they same time they shouldn't have veto power over transformative technologies just because they'll be put out of a job.
Wait, aren't huge trucks slated to be the first ones to be replaced by autonomous equivalents? Making a self-driving long-distance truck is a simpler, more constrained, and easier to solve incrementally problem than making a self-driving passenger car.
This is the equivalent of saying, we need strong protections on the Internet, because child pornography exists
It’s not an actual argument because it doesn’t have specific measurable distinctions between deaths related to driverless vehicles versus driver vehicles
We don’t have these types of measurable capabilities because we don’t actually have the data that shows what is the fatality rate per million miles at the scale of integration that an entire fleet they have supplanted the taxi industry would do.
You’re better off actually evaluating what is the death rate of taxis compared to individual drivers because the individual drivers are not going to be replaced at the same rate as taxi drivers and so the taxi drivers is the first more important number to evaluate with respect to traffic fatalities
Unless you’ve done that specific math, and walk the dog out, including the Wayno reports, some of the cruise reports, and some of the reports from Tesla (which killed a guy if you recall) then you might have an argument
As it stands, all I see with driverless cars is we’re trying to replace humans in an entire labor category, and we have nothing for those people to do
I suspect that is actually a much bigger problem than the theoretical reduction in fatalities on roads
> This is the equivalent of saying, we need strong protections on the Internet, because child pornography exists
It's not remotely like saying this, and going straight to child pornography in is like going straight to the Nazis - the sign of an unserious argument.
> It’s not an actual argument because it doesn’t have specific measurable distinctions between deaths related to driverless vehicles versus driver vehicles
You're just not understanding the definition of the word argument if you think this is true. Also, these things are being measured, so we're getting to this understanding. If your argument is that we shouldn't put robotaxis on the road because we don't know if they're safer means that we can't measure if robotaxis are safer and thus regardless of their relative safety, we must ban them because we don't have measurements yet.
> I suspect that is actually a much bigger problem than the theoretical reduction in fatalities on roads
But that's not an actual argument because it doesn't have specific, measurable distinctions, no?