Hacker News Reader

Comment by el_cujo | Hacker News Reader

I listened to him talk a little bit on Joe Rogan's podcast, and it seemed to him that the UBI/Freedom Dividend is his plan explicitly because stuff like retraining these fifty year old guys who have been truck drivers or coal miners their entire lives to become computer programmers or something is really unrealistic. I guess it would be more possible to retrain the younger people that populate retail, but really I think retail is viewed as just something people do to put themselves through college or during highschool rather than a life-long career like truck driving. That's not to say there aren't people in retail who aren't going to get screwed by automation, but that's sort of the appeal of UBI, it's something that applies to everyone versus having to try and justify to someone that you deserve help for your specific circumstances.

View on HN

>retraining these fifty year old guys who have been truck drivers or coal miners

While thats true, I think the premise is wrong. Techno-types love to imagine that automation is just going to sweep through these industries.

The devil of these things is in the details. Yes, perhaps the tech can get to approximate 90% of what a human can do. However, the real problem are outside of tech. Like liability insurance. Who gets sued when one of Elon's driverless trucks plows into a school bus? Surely not the "trucking company"? So the software company will have to assume that...Now, imagine there are 50M driverless vehicles on the roads, what does that monthly policy look like? What insurance company will underwrite that? What happens if a zero-day is found? etc. etc.

It's not just trucks, though. He gives a lot of examples, some of which are observable today. Like a lot of work done by paralegals and junior lawyers is just research - plowing through books and writing summaries and references. It's slow, tedious, and error-prone. A computer can do it much faster and better. How about reading x-rays? Train an AI by having it review the results from millions of other x-rays. That's a lot of radiologists whose work is replaceable.

And that's really Yang's point. It's not just truck drivers. Any job that is basically rote work - even jobs requiring a high level of education and professional certification - is vulnerable to advances in AI and robotics. If your work isn't essentially creative, there's a good chance that it could be done cheaper, faster, and better by a robot at some point in the not-too-distant future.

Unless you are a developer , while we built more abstractions and automation into our software engineering we also attempt more complex problems. I don't see dev being in lesser demand in the next 30 years

Sure. Software is essentially creative work. I'm not worried about our industry. I'm worried about what our industry will do to other industries.

I think it will still affect software. Things can always be done faster. I think it'll be like farming where it won't go away but fewer people can do more work.

I can imagine a day where voice recognition can allow marketing folks to just say what they want changed on a website and AI makes the update and asks for confirmation before pushing it live.

I'm not worried. I've been in this industry for 25 years now, and seen multiple technical revolutions. (Back in my first serious job, we were hand-rolling streaming data protocols with lex and yacc on raw sockets! There was no ssh, no xml, no json, and http was a lab toy in Switzerland.)

Requirements are hard. Understanding what needs to be done, and doing it all in the correct order, with buy-in from all concerned parties, is an essentially human, creative problem. It's not a simple matter of "marketing folks ask the AI to update the site", because they don't know how to ask it. Software is like all those tales of people getting their wishes granted by a genie or something, and regretting it because what they asked for wasn't actually what they wanted.

And who turns someone's dreams into something the genie will do correctly? Programmers. We're the interface between the world of human needs and the world of technical realities. We're doing craft. Our tools may change, but our problem is constant.

Yeah, software devs will be the rich/elite in 20 years (we'll be the only ones employable).

Nope. Most software development jobs, along with many other "creative" jobs, will be automated as well.

eventually, but -- creative will go AFTER doctors/lawyers/etc... I can't wait for politicians to go.

>Who gets sued when one of Elon's driverless trucks plows into a school bus? Surely not the "trucking company"?

Actually, yes, the trucking company gets sued. This is pretty well established. If your Toyota malfunctions and you hit someone, you get sued. You may sue Toyota afterwards, but you're going to get sued first.

This has literally happened before. It's not a what if. Toyota recalled thousands of trucks about 3 months ago because of a braking issue. I fail to see the difference between this and car insurance of the future.

Even if insurance were the issue, it's fairly easy to solve; in the contract you sign to use an automated vehicle, you'd assume liability.

>If your Toyota malfunctions and you hit someone, you get sued. You may sue Toyota afterwards, but you're going to get sued first.

You extrapolating that future markets will work exactly like the past, because "its a car".

But its doubtful it will. The reason why the auto industry evolved the way it did (driver gets sued), is because 90+% of the time it is operator error.

Look at another example. Kraft distributes 10M packages of poisoned cheese, what happens: (a) All the local grocery stores get sued? or (b) Does Kraft get sued?

Obviously "B" would be the case, since the local grocery stores would say: "We had no way to test the cheese for poison, it was pre-packaged from Kraft"

This is very similar to what would happen in a 100% autonomous vehicle situation. The owner doesn't have access nor the capability to judge the engineering / code.

>Toyota recalled thousands of trucks about 3 months

Not the same. Toyota doing a recall to replace a $200 part doesn't come close to the potential liability that a autonomous manufacturer that had thousands of accidents...each potentially costing millions each.

>Even if insurance were the issue, it's fairly easy to solve; in the contract you sign to use an automated vehicle, you'd assume liability.

Yes, and the trucking companies will just say: No

Their insurance rates would skyrocket so high, that it would be cheaper to just continue to use human drivers.

No, people don't just sue the drivers because it's driver error, they sue because they have a chance to get money. People who get t-boned by a driver running a red light get sued by the insurance company of the driver that ran the light. It can have nothing to do with who is at fault, ultimately, just who is the easiest target.

In your Kraft scenario, the store would likely be sued, as would the corporation running the store, Kraft, and potentially the packaging company.

Here's a perfect example; the Romaine lettuce E. coli outbreak happens, right. A man eats a salad at a restaurant, gets sick. Who does he sue? Everybody. Has nothing to do with who is responsible, has a lot to do with who is going to pay his bills.

Why is the liability for a $200 part lower? If the cars can't apply the brakes properly and people die in car accidents, are they any less responsible?

Trucking companies won't say no, because it's extremely unlikely that all the cars will get into accidents all the time, and they don't have to pay more because the car didn't have a driver than they would if the car was being driven by their employee. Why would their insurance rates skyrocket? It appears there will be far less accidents with automated cars.

> This is pretty well established. If your Toyota malfunctions and you hit someone, you get sued. You may sue Toyota afterwards, but you're going to get sued first.

It's not unlikely if the immediate crash investigation even hints at a product defect, you (all of driver, driver's employer if done in course of employment, and owner if different from the preceding two), Toyota, and everyone in the chain of commerce between you, plus the people up the chain from Toyota with regard to the part at fault will all get sued, simultaneously, in the same action.

The only change self-driving cars bring is eliminating the driver as a distinct target from the manufacturer and a whole bunch of upstream suppliers, but the person or company owning and the one (of different) directing the operation of the vehicle are still potential liability targets, as well as the upstream chain of commerce.

The UBI of $1K a month is insane. This would be 320 million Americans receiving $1K a month per person? That's $320B a month...$3.8 Trillion a year. That's DOUBLE the current Federal government's annual expenditures. He argues, with hand-wavey logic, that this will solve our problems and we can tax tech companies and all of this...but no, there's not enough annual revenue at the big tech companies to cover this enormous welfare program.

Is the UBI taxable income? Perhaps UBI would generate enough revenue to be cost neutral or even cost positive.

Exasmple: food stamps generate $1.73 in tax revenue for each $1 handed out. It's just good policy - to a point.

From his interviews the UBI is not taxed, but the money comes from:

1. Consolidating existing welfare programs with large bureaucracies (500-600BN)

2. 10% VAT on companies which report 0 profits (800BN)

3. Projected economic growth leading to sales taxes (this is like the foot stamp example you shared)

4. Healthcare, prisons, etc (100-200BN)

1. What prevents the UBI from generating its own large bureaucracy?

2. VAT taxes come in the form of higher prices on everything. No such thing as a free lunch.

4. Healthcare, prisons is handwavey logic. Again, we already provide free healthcare to the poor via Medicaid. So there are no gains to be had there. He makes the wild guess that welfare payments would reduce crime, even though we saw the highest crime rates in the 80s and 90s prior to welfare reform (which reduced the number of people on welfare). So no, I don't buy any of it. There is no experiment of this being tried in the wild, at the city, county, or state level. Why would you take an epic proposal and immediately propose a huge shift of resources in the economy without any evidence to back it up?

1. "The government is bad at a lot of things but it's great at writing checks to people on time", because there's no income cutoff or other eligibility requirements it becomes a lot easier to manage than existing welfare programs.

2. VAT is designed to capture value from companies which employ few people, generate a lot of value, but report little in profits, which are a lot of tech companies today.

4. We spend more on healthcare to worse results than other developed countries, a lot is institutional inefficiency.

There's definitely evidence out there to support that people are suffering more even though GDP goes up, look at the labor participation rate of working-age men in these swing states.

Epic proposal? Alaska has been running a UBI from the 60s, instituted by a Republican governor and it's wildly popular. A UBI almost passed under Nixon as well, passing the house and getting killed in the senate twice because democrats wanted the amounts to be bigger.

The way I see it, we either do this now or we continue to keep our heads in the sand while the country rots from the center.

Apple earned $60B in taxable income in 2018. Amazon $2.8B. Amazon has a business model where they continue to reinvest profits in the business. Amazon employs 500,000+ people at minimum wage of $15 an hour. They do plenty for the economy as a whole. And if you tax them with VATs, they'll just increase prices on goods. There is no way around it. If you tax with VATs and extract value out of Amazon, you get less capital investment from companies like Amazon, which are highly efficient at capital allocation. That's the reason our economy grows in the first place. Redistributing capital to consumers only incentivizes and subsidizes consumption. To some extent consumption is good, but too much consumption and not enough production leads to a stale and stagnant economy.

Alaska has a budget crisis right now, and they owe back pay to their citizens on the PFD. Besides that obvious disaster, it's a variable payout based on oil money. How sustainable do you think that will be once, say, electric cars become a thing?

The variable payout has been as high as $2,000 a person per year...to as low as $800 a year.

Alaska has benefited, like many oil-rich nations (including Norway), from a relatively small population with relatively large oil reserves and huge markets to sell those reserves to. Alaska has a population of about 750,000. Smaller than most large U.S. metropolitan areas. If they had the population of Texas (28 million)...that oil dividend would amount to $54 a year (in its highest amount ever).

The U.S. is almost three orders of magnitude larger than Alaska in terms of population. A LOT changes at that scale. Especially when the U.S. is THE largest market on the planet. We are the market other industries strive to sell to. Wreck that market, and you'll wreck the world economy.

(https://www.newsminer.com/news/alaska_news/dunleavy-embraces...)

You don't know anything about his "hand-wavey logic", or you'd know right away that it's not "320 million Americans". It's only adults from age 18-65 (or whenever they choose to retire to Social Security, presumably).

It would be whatever bastardized form of it Congress passes and by the way, Social Security is already on the path to insolvency. It's still a multi-trillion dollar proposal, a huge economic experiment that could easily end in catastrophe and not be easily reversible (as people would come to depend on payments, try taking that away from them). It's batshit crazy.

In his book, Andrew Yang points out that one of the major factors damaging Social Security right now is a sharp rise (like a doubling) in disability cases. And the amount of people on disability maps directly to areas of economic depression and underemployment. Basically, it's become an alternative form of long-term unemployment for people who are unable to find work. And worse, since getting a job will likely lead to losing benefits, it actively discourages employment and makes people think of themselves as broken failures.

That all said, what is your alternative? Let huge swaths of the country degenerate into poverty and despair? Wait for them to vote an actual fascist into office? Or riot?

Given a choice between higher taxes and civil war, I will gladly choose higher taxes.

Sensationalize much? Taxes wouldn't just be a little bit higher under this proposal they'd be double what they are today...in the name of social welfare. And when that ruins the economy, you'll be dealing with another Venezuela.

Social security isn't on its way to insolvency because of disability expansion...it's on its way to insolvency because it was designed in an age when we had an average of three children per female and life expectancy was shorter. Now that we're down to two and possibly dropping...there's less workers per retiree.

Reply on news.ycombinator.com