MIT Economist David Autor on the Future of Work and Polanyi's Paradox | Hacker News Reader

MIT Economist David Autor on the Future of Work and Polanyi's Paradox

econtalk.org

For someone who's a techno-libertarian, like myself, it's hard to have a good response to the "computers are taking all our jobs" argument.

The most straightforward response is "It will PROBABLY turn out OK, new jobs will probably show up", but that just isn't a very satisfying retort, even if it's the truth.

Where the "jobs" are going seems like the wrong question to me, honestly. I'm fine with not having a "job" as long as I have access to wealth - jobs are a cost we bear in order o generate wealth. The fact that we no longer have to pay that cost, writ large, seems like a good thing to me, overall.

There are of course a lot of questions about how the wealth will be distributed and whether or not we're obsoleting ourselves, but I think I'd rather try to carve out a niche for myself in a world where the problem is that we create more wealth than ever with minimal input from humans than live in a world where human labor is a dire necessity.

Perhaps you could tell them that the question betrays a deep misconception. We can use computers to decrease and deskill employment. Or we could have fewer hours at greater pay, with less drudgery for the remaining work.

(It's like energy research. Want nukes, or sustainable or clean energy? That's up to society to decide.)

Of course, passive people who stick to the status quo means we get bad outcomes.

> Or we could have fewer hours at greater pay, with less drudgery for the remaining work.

Right. However, GP may have alluded to the question of whether such an outcome is compatible with libertarianism. Does that question betray a deep misconception?

It isn't the truth. The truth is there are very few places new jobs are coming from and many places they are being destroyed. Can you articulate a markets-as-natural-force argument that actually works? If not, it's time to question the viability of Libertarianism.

Libertarianism is wishful thinking. Or it's a fake and cynical beard for corporatism and oligarchy. If you want an economic outcome you have to design an economy that delivers that outcome. If you fail to do that you leave the field open for oligarchs.

> The truth is there are very few places new jobs are coming from and many places they are being destroyed.

Well, given that almost all measures show that the unemployment rate has gone down in the last 5 years, it means exactly the opposite of what you say: More new jobs are appearing than are being destroyed. (other time frames lead to other conclusions of course)

> If you want an economic outcome you have to design an economy that delivers that outcome.

That is pretty much a textbook example of the "wishful thinking" you accuse me of. You really think anybody has a clue as to how we'd design an economy that would lead to a specific economic outcome?

5 years is not a good time horizon for unemployment rate comparison, because that time window is during the Great Recession/Lesser Depression, which surely dominates any longer term (and therefore slower) effects of technology.

> You really think anybody has a clue as to how we'd design an economy that would lead to a specific economic outcome?

Not at a detailed level, but stuff like tax cuts for the rich and financial deregulation are not random accidents.

If by a limp recovery after an ibank-created crisis, then yeah. Participation rate down, massive increases in SSD, massive growth in income disparity, no change to these trends in sight. It's denial in the form of statistics.

> "You really think anybody has a clue as to how we'd design an economy that would lead to a specific economic outcome?"

Yes. Start with these design changes: raising the minimum wage and cutting the work-week.

I'm sure that you think that would work. In reality, though, people who are sure that their solution will fix the economy very often run into the law of unintended consequences.

I buy drcode's skepticism more than I buy your certainty and optimism.

How much evidence have you actually looked at? A priori pessimism is no more innately correct than a priori optimism: both are ungrounded until you look at the data.

Well, I've looked at the fate of enough "this is sure to work" ideas for fixing the economy to claim that my pessimism is, on average, warranted.

I haven't done an in-depth analysis of this specific proposal, however, so in that sense I haven't "looked at the evidence".

Really? And who, exactly, said that they were sure to work, and how, exactly, were the ideas implemented (that is, how loyal were the implementations to the proposals)?

> raising the minimum wage...

So you're saying making it more expensive to employ people is going to decrease unemployment?

There are two separate problems we face: income inequality, and rapid destruction of jobs. Raising the minimum wage helps both, by raising the lowest incomes out of poverty and increasing discretionary spending on a mass scale.

Another reason to raise the minumum wage is that we are in a permanent large surplus of labor. Without a higher minimum wage, all wages will get bid-down, especially at the low end. Perhaps you are fine living armed-up and behind barbed wire, but a functioning society has a lot of value. We are going to have to decouple remuneration and the value of an individual's contribution to society from a free market for labor.

So you're saying allowing wages to drop will make people better able to pay their bills?

"We shouldn't drop wages because people won't be able to pay their bills".

Appeal to consequences. Try another.

As if the parent comment wasn't as well. I was just pointing out that it's a two-way street.

Suppose we continue with the status quo and jobs begin miraculously materializing: there's a very good chance that they will be lower paying than what we have now, and it will still be difficult to make a living. To some extent it's a lose-lose situation unless we can come up with a third option like a universal basic income.

Edit: My, aren't we feeling pedantic today? I have a great idea: let's argue semantics rather than content!

> As if the parent comment wasn't as well.

Tu quoque fallacy. Try again.

> I was just pointing out that it's a two-way street.

No. Watch closely.

You said if we drop wages then people can't pay their bills, and therefore we should not drop wages.

Your claim is that we should not do something based on the consequences of that action.

The parent said that increasing minimum wage is equivalent to making it more expensive to employ someone.

The parent is not claiming anything, but genuinely asking:

- assuming increasing minimum wages is another way to say that we would make employment more expensive

- then how is it that making employment more expensive helps decrease unemployment?

Which is a genuine question that you have not answered.

> Suppose we continue with the status quo

False dilemma. What is being argued is neither continuing the status quo nor increasing wages, but a third alternative, namely (I presume) abolishing minimum wage laws.

> Edit: My, aren't we feeling pedantic today?

Insulting is how the intellectually dishonest say "I'm out of arguments".

> I have a great idea: let's argue semantics rather than content!

How about you just answer the question:

How is it that making employing someone more expensive (which is what increasing the minimum wage effectively does) help decrease unemployment?

Why can you not answer such a straightforward question? Why do you avoid answering it by insulting me?

"We shouldn't increase the minimum wage because employment will go down".

Appeal to consequences. Try another.

Did I get the right level of condescension?

Given that he's arguing a should statement rather than an is statement, "appeal to consequences", in this context, is actually consequentialism, and is the best framework we have for rationally justifying imperative statements.

I disagree. I think things are wrong on principle, and not just because of the results.

See deontology for a better alternative to consequentialism: http://lesswrong.com/lw/1og/deontology_for_consequentialists...

Well, I gave reading that the old college try, but nothing in there actually justifies deontology, let alone your peculiar form of capitalist deontology, as Correct, let alone as a superior alternative to consequentialism.

In average, yes, because as a result more people will have jobs. Especially those who have a harder time paying their bills.

> it's hard to have a good response to the "computers are taking all our jobs" argument.

The right response is, "What's the alternative (or solution)?"

You could also point out that many of the problems we have now could be described as inefficiencies ultimately caused by not giving individuals enough choices and/or freedom or by subsidizing the wrong things (overpriced housing, discounted degrees in metaphysical therapy, employer-provided healthcare, etc.).

I could be missing something but most of the solutions I hear about involve more restrictions on employment or more targeted subsidies. I'm not convinced that we won't just be a different kind of wrong in another generation.

It's the kind of thing where you have to say "I observe that I am confused" and have to re-examine your premises, your objectives, and the conclusions you derive from them.

Maybe being techno-libertarian shouldn't be (a strong) part of your identity because it does not make a very good end in itself.

'it's hard to have a good response to the "computers are taking all our jobs" argument'

I don't think people really care about having a job. They care about being able to put bread on the table. (For various degrees of "bread", as "wants" can be psychologically infinite.) So, long term, the answer will be "yes, sure. And that's a good thing. You'll have what you have now, with no need to work."

Going from here to there, is very likely to be bumpy, though.

And finally, "not working" doesn't mean "not doing anything", as many people assume.

Unfortunate to discover that this new-to-me paradox isn't a paradox at all. "Some things we know/understand can't be written down [easily], thus we know more than we can explicitly explain". It's not counter-intuitive at all, and doesn't oppose itself in any way.

The paradox is the part where things that are hard to write down (i.e. do laundry, trim your fingernails, make a sandwich) seem extremely easy and seem like the first things machines would have taken over, whereas things that are easy to write down (playing chess, arming nuclear missles, diagnose a tumor) seem very hard but are relatively easy for a computer to do.

(And yes, it still doesn't qualify as a paradox, in the mathematical sense.)

I don't know, the first group of things is wholly physical, and the second group is wholly digital. It's incomparably easier to build things (and rightfully so) that never touch the physical world because you don't have to build any physical hardware manufacturing infrastructure.

As someone who studyied Polyani intensively, I was surprised by that too.

Protip: do not read what he wrote if you have strong libertarian view.

Offtopic but is there a way to load these kind of one-off podcasts onto an iPhone?

You can download the episode, without subscribing, from the EconTalk "page", or whatever it's called, in the iTunes store.

Not the answer to your question but I listen to this show from time to time using Antennapod, which is in the fdroid repository for Android phones. Works pretty well. The podcast is listed in one of the guides (gpodder.net) and is a searchable result or you can add the RSS feed directly from the econtalk site.

Reply on news.ycombinator.com