To clarify what I was saying – Bill Gates, who has more money than what 70% of the countries on Earth earn in a year, said that an AI teaching math is the best way to solve inequality. I was just pointing out that it’s very convenient for him that it’s that and not any sort of wealth distribution!
That’s fine. I think there’s a way less chance of major wealth redistribution happening in the US than teaching kids good math skills to give them a leg up in their career opportunities, but I get it.
The thing about AI is that its strength is not going to be finding the unexpected but perfect le mot just, but rather cranking out an infinite amount of the very expected le mot banal. Which means it will be less useful in the entertainment industry, where you need to be entertaining, than in, say, writing high school essays, where you do not.
In the case of barks, which should be short, generating variations on a theme isn’t hard for a writer with any experience at it. I’m talking ten minutes for a couple dozen variations, given a defined situation. (But much more if it involves inventing lore and idioms for that lore, i.e. “How do the five-armed floating sentient gasbags from the planet Zorp talk?”) So merely getting some mundane bark variations out of a machine isn’t much of a help.
The hard part isn’t coming up with variations, but with crafting a bark that 1) conveys what the NPC is trying to say, 2) is entertaining and memorable, and 3) isn’t annoying after hearing it a bunch of times. Which is (as the video acknowledges) where humans need to step in.
The fact that Ubisoft is touting this tool is interesting, because when it come to Ubisoft’s dialogue, everyone says … well, nothing. Nobody talks about the dialogue in Ubisoft games (other than some monologues from the Far Cry big bads, I guess.) It’s telling that the memorable one-liner they use as a prompt isn’t even from a Ubisoft game.
There’s a more… brutal bit about math. Of all the sciences, math is the only one that you can’t really BS your way through even in elementary education. When you do well at math, you’re almost certainly likely to do well at most other subjects in general. If you’re absolutely horrible at math, you’re probably (not necessarily) more likely to be worse at all other subjects in general - because it’s reflecting something about your study habits / home situation / work ethos, ect.
It is?
But yes, maths usually opens a ton of doors and ways of doing things that you’d never be able to approach otherwise. It’s pretty useful.
Derivative, integral, both calculus terms that I don’t use the formulas for. But do understand the basic principles of.
Yes for sure. Understanding those notions is surprisingly helpful for setting up software monitoring dashboards for instance - something I would never have expected!
Too few math teachers? Privitized AI can replace teachers. But why not hire more teachers?
Healthcare too expensive? AI can fill out insurance claims. But why not get rid of health insurance?
Climate change killing crops in the third world? A chatbot can tell subsistance farmers which seeds to grow instead. But why not stop climate change?
I understand thinking solutions that challenge capital are unrealistic. But if you’re going to be cynical, you need to keep being cynical when famous monopolist Bill Gates pitches magic beans that will keep the status quo exactly the way it is.
It’s not Gates proposal people are arguing for, it is this comment we are arguing against:
This is what I, at least, took issue with
That’s good to read, as “the best opportunity for reducing inequity is to improve education, particularly making sure that students succeed at math” is a very different statement than “making sure that students succeed at math is a good thing and will help them in their career” which is what people seem to be discussing.
I think this might be true, but I also think there won’t be enough not-sucky jobs for everyone, and a lot of people will end up doing sucky jobs, whether they are good at math or not. It’s not really a plan for broad prosperity.
There are tons of service jobs where math plays basically no part in them at all. Just take a walk around any town. Probably they’re not very desirable jobs, that’s true, but a lot of people will end up doing them.
Isn’t it just! And of course, the answer to all inequality from the usual VSPs is always education and training. Why can’t those out-of-work coal miners become math professors?
One of our math professors was saying in a meeting the other day that he didn’t care whether his students had, say, the quadratic equation memorized. He wanted them to understand why and how they should use it. Good math education these days is far more about understanding the theory and the application of math in context, rather than old-school rote memorization. Hell, some of the math classes we have don’t even rely on tests any more, but rather on practical exercises and projects where using math appropriately is the key.
I also think people should learn math, because learning is its own reward if nothing else. But teaching more math isn’t going to solve income and economic inequality. The idea is absurd.
Imagine you’re talking to one student. You can say to that student something like if you work hard in school and get a good education, you are very likely to find a good job that pays well. And at that level, the advice is good because the outcome is plausible. It’s a thing that can happen.
Then imagine you’re talking to 30 students in a classroom. Even here, the advice is still good, because the outcome — that all 30 of them will find good jobs — is still plausible.
Now imagine that you’re talking to every student in the world. Is it plausible that every student in the world will find a good job? No, it is not plausible. It’s not even possible. There aren’t enough good jobs, because most of the jobs are bad jobs, jobs that don’t pay very well and don’t lead anywhere better.
Income inequality and inequity don’t occur because people failed to get the right jobs. It occurs because most of the jobs are bad jobs that don’t pay very well. You can’t solve that problem by exhorting every worker or potential worker to get a better job, because that solution isn’t possible, because there aren’t enough better jobs. You can only solve that problem by offering better pay for the bad jobs.
Now, maybe you don’t want to solve that problem, because you think it’s natural that some people get great jobs while most people languish eternally in poverty because of bad jobs; that it must always be so, that it is a dog-eat-dog world in which there are a few winners and most people are losers. Maybe you’re offering advice to someone about how they, personally, can increase their chances to be one of the few winners. That’s fine, I guess, but it isn’t an answer to the question how do we fix extreme wealth and income inequality.
I agree, but I don’t think the argument is that everyone can get better jobs. The argument is that if poor kids are better at math, the labour productivity of the lowest labourers will increase, and therefore reduce the huge gap in the productivity between the lowest paid and highest paid workers, which will 100% translate into a smaller gap in earnings.
It is still complete fantasy by ignoring the determinants of the changing distribution of income since the 70s. Declining education standards in low income areas is much more a consequence of inequality than it is a cause.
I don’t think that idea is any less crazy, even without the trend in distribution of productivity gains. The person mopping the floor in your office isn’t going to mop the floor better or faster because they learned math, and none of your other employees are going to do a better job as a result of the more-rapidly-cleaned floors. The person assembling your hamburger won’t do a better job because they can imagine problems in calculus, just as the person at the window taking your order doesn’t need any math of note to determine the cost or make the change.
The thing is though… the person mopping your floor in the future is gonna be a robot.
I think one thing that may be a difference in belief here, is that I don’t think that there are people who are just “bad at math”. There may be folks who lack a natural aptitude, but I believe that with improved teaching methods, way more people can be competent at understanding math concepts which will serve them well in jobs like programming moving forward.
Yes, that could be so. It means a shitty, low-paying job disappears, but it doesn’t mean a great, high-paying job replaces it. If anything, that makes income inequality even worse, not better, because ‘no job’ pays even less than a shitty one pays, no matter how much math everyone learns.
Learn more math isn’t a solution to societal income inequality. It is at best personal advice of benefit to a relatively small number of members of society, not because they can’t all learn math, but because there aren’t enough good, high-paying jobs for all of them. That’s why I asked the question in the first place: how many math jobs does Gates think there are? Which leads to the corollary: what does he think everyone else ought to do?
How many people in your view can have good-paying programming jobs? What percentage of the workforce?
You see here, that some people are doing what I thought they were doing, offering what amounts to personal career advice rather than any actual solution to income inequality. It’s a long-standing theme in American political ‘solutions’, based on the fantasy that it’s possible for everyone to win the lottery if they just try hard enough.
Honestly? In the future, I suspect that the vast majority of jobs will involve programming on some level.
How many jobs will that be, as a percentage of all working-age people? And how does this view of the future ameliorate the extreme income inequality that exists right now?
That involve programming on some level? Pure speculation, but probably upwards of 80%.