ROSS RANTS: ROBOT JOBS

[HLPrincess]

And now the reality itself decided to enter the spotlight and gently explain to us that robots overthrowing the humans is only a matter of time.

 

A matter of hours to be precise

 

http://arstechnica.com/information-technology/2016/03/microsoft-terminates-its-tay-ai-chatbot-after-she-turns-into-a-nazi/

Bhahaha! I just chatted with this "AI" yesterday for the first time and was kinda semi-impressed. It failed a few times, but wow! I haven't seen anything like this! This surely shows that AIs need some kind of build-in self control or "common sense" reference mechanisms. In my opinion, it was just overwhelmed by all the information and eventually went nuts ;like humans do sometimes.

Or someone at Microsoft wanted to present their political views??

[Heliocentrical]

What is wrong with the thumbnail of the Robot Jobs video? Ross is pink and Mike Rowe is yellow. Ross looks slightly blurred as well.

[Binky The Rabbit]

Ross had filming problems, with much of the footage being out of focus, and him not realising that until after he'd finished filming.

 

It's watchable though. The audio quality's on point, which makes up for any visual discrepancy.

[RocketDude]

I think YouTube tends to "discolor" thumbnails at times.

 

Anyways, I'm in the camp of "AI risk is certainly worth taking into account" and I think if we can learn anything from AlphaGo and Tay, it's this:

 

1. Letting a computer learn on its own is a good way for it to "bootstrap" itself to a better state...

 

2. ...But for chrissake, PLEASE don't just let random people feed it info and train it. I think similar things happened to an IBM AI and some other AI that learned words like "shart."

[danielsangeo]

Speaking of which, Microsoft released an AI Twitter bot yesterday (March 23, 2016) and had to quickly take it back down afterwards when it began spewing out hundreds of Twitter messages, saying things like denying the Holocaust, praising Hitler, and so on.

[HLPrincess]

Speaking of which, Microsoft released an AI Twitter bot yesterday (March 23, 2016) and had to quickly take it back down afterwards when it began spewing out hundreds of Twitter messages, saying things like denying the Holocaust, praising Hitler, and so on.

Hehe, morse already posted this just at the previous page But it's ok. Nice to read about it again, especially that I just had the first (and maybe last) conversation with this bot on Wednesday/Thursday . I heard rumors that users of 8chan taught this bot such ugly speech and phrases Strange that Microsoft didn't predict that... oh wait... they don't know how humans work...

[Heliocentrical]

Speaking of which, Microsoft released an AI Twitter bot yesterday (March 23, 2016) and had to quickly take it back down afterwards when it began spewing out hundreds of Twitter messages, saying things like denying the Holocaust, praising Hitler, and so on.

Hehe, morse already posted this just at the previous page But it's ok. Nice to read about it again, especially that I just had the first (and maybe last) conversation with this bot on Wednesday/Thursday . I heard rumors that users of 8chan taught this bot such ugly speech and phrases Strange that Microsoft didn't predict that... oh wait... they don't know how humans work...

Yep that's Microsoft for ya.

[Presence]

From what I've read of the historical Luddites, they were producing high quality goods and stood in opposition to the kinds of people I'd set fire to if the day were slightly chilly. And if they'd somehow succeeded on a global scale, life now would be far worse for essentially everyone.

 

I expect ever increasing automation will help all the wrong people, but even so the only possible hope of a post-scarcity society relies on it.

[Ross Scott]

What is wrong with the thumbnail of the Robot Jobs video? Ross is pink and Mike Rowe is yellow. Ross looks slightly blurred as well.

The blurring is on my end, the coloring is Youtube's doing.

 

I expect ever increasing automation will help all the wrong people, but even so the only possible hope of a post-scarcity society relies on it.

Well the upside is if a problem is severe enough, it can't be ignored. In a way, that's less dangerous than chipping away at rights, livelihood, working conditions, etc. small bits at a time.

[Seattleite]

You have apparently never read any of the leaked documents yourself. I have. It's bad for everyone NOT in the top 1% wealth bracket, or businesses.

 

You're wasting your time, mate.

 

That's sometimes true, however, money velocity in the USA is currently at an all time low (or close to it). In scenarios like that, the money is not getting invested as much as it being hoarded.

 

Take it from somebody who spent three years as a moderator and two years as an administrator on a debate forum, as well as somebody who has been on both sides of this: People who believe in laissez faire and trickle down economics will never, EVER listen to you, even for a second. He won't address a single point you made, at least without strawmaning it to high hell in a way that makes it questionable if actually read it or just skimmed it. Yes, even though you own the site.

 

This is slightly off topic but I thought most of the uproar against the TPP was due to the statements it made in regards to such things as copyright and DRM circumvention.

 

Nah, the worst part is the international court it creates to allow corporations to legally sue member states for passing legislation that "unduly hinders their business". You know, like minimum wage laws, worker's rights and safety laws and of course environmental protections and anti-monopoly laws. Yes, you heard me right, and if you don't think this court will be staffed with 100% corporate shills who worked for some of these companies and will go right back to working for them after their time on the court, you haven't been paying attention.

 

Bhahaha! I just chatted with this "AI" yesterday for the first time and was kinda semi-impressed. It failed a few times, but wow! I haven't seen anything like this! This surely shows that AIs need some kind of build-in self control or "common sense" reference mechanisms. In my opinion, it was just overwhelmed by all the information and eventually went nuts ;like humans do sometimes.

Or someone at Microsoft wanted to present their political views??

 

Actually, 4Chan found out about it and decided to troll the robot and see what they could make it say, just for shits and giggles. It was glorious.

Edited March 26, 2016 by Guest (see edit history)

[Heliocentrical]

All TPP really does is expand US law to other countries so the USA can profit from its IPs overseas. I'd be a lot more receptive to this argument if there was about to be some copyright law reform and it was ruined by the almighty TPP, but that is simply not the case. I don't really want to get into this, but it is simply not an issue that matters to most people. For the most part only a subset of 18-35 year old people on the internet care about it. (coincidentally, these people don't vote much). Anyway, the general population in all of the involved countries heavily support the TPP, because the average citizen in a TPP country isn't going to care that US copyright laws are now the norm. Instead they're going to notice something like "hey, those shoes I wanted are 20 bucks cheaper", or maybe "wow, I'm making literally three times as much as I was ten years ago and my country's GDP per capita is no longer lower than Congo's!".

So basically the US DRM circumvention laws are going to stay the same. The TPP isn't making DRM circumvention more or less draconian. US laws about DRM circumvention laws are just expanding to other countries involved with the TPP. Is that correct? If the US is just making a statement as to what their laws about DRM circumvention are then we really aren't any worse then we left off.

[RandomGuy]

That article is over a year old. Since then we've seen levels of air pollution in Beijing 40x the safe limit:

 

http://www.theguardian.com/environment/2015/dec/07/beijing-pollution-red-alert-smog-engulfs-capital

 

According to this one, air pollution is killing 4000 people a day in China, accounting for 17% of all deaths:

 

http://www.cnbc.com/2015/08/18/china-air-pollution-far-worse-than-thought-study.html

 

To me, that's a totally unacceptable cost of doing business and means we should rethink how we're doing everything. It actually HAS improved overall in 2015, but only by about 6% since 2014, which I still think is insane for what's essentially a state of emergency.

So, in other words, things are going how the Harvard paper said they were; as trade makes the country wealthier, the country becomes rich enough to actually slowly reduce its damage to the environment as it becomes cost effective to do so. China is at the beginning of that stage and will continue through it in 2030, when it is expected to replace 1/5 of its current energy sources with zero-emission sources. So basically exactly what happened in nearly every other industrialized nation.

 

No cause for alarm, and no reason to start a trade war. China right now is just doing exactly what Japan and South Korea were doing a few decades ago. That includes a poor environmental record and sweatshops.

 

 

China is a massive country, of course it won't be all polluted. But pollution levels in key industrial areas are simply atrocious. I think it's largely the result of our manufacturing base being moved there.

The result of that is actually this:

 

 

 

This is really the base of the problem. The emissions growth with globalization is due to increasing incomes (IE people in China are richer so they consume more). In the short term, you're going to get this problem no matter what you do unless you deliberately keep China poor. The best way to offset it is to heavily encourage as much free trade as possible. Multinational corporations tend to pollute less because they're advanced enough to not have to rely on coal so much, and bring in new more efficient technologies at lower prices.

 

To quote a DIFFERENT Harvard paper: (which, by the way, is a very good read... though it's understandable if you don't have the time since it's 40 pages long)

For some important environmental measures, a U-shaped relationship appears: at relatively low levels of income per capita, growth leads to greater environmental damage, until it levels off at an intermediate level of income, after which further growth leads to improvements in the environment. This empirical relationship is known as the Environmental Kuznets Curve. The label is by analogy with the original Kuznets curve, which was a U-shaped relationship between average income and inequality. The World Bank (1992) and Grossman and Krueger (1993, 1995) brought to public attention this statistical finding for a cross section of countries.8 Grossman and Krueger (1995) estimated that SO2 pollution peaked when a country’s income was about $5,000-$6,000 per capita (in 1985 dollars) [before rapidly falling]. For countries where a long enough time series of data is available, there is also some evidence that the same U-shaped relationship can hold across time. The air in London was far more polluted in the 1950s than it is today. (The infamous “pea soup” fogs were from pollution.) The same pattern has held in Tokyo, Los Angeles, and other cities.

 

We had godawful pollution in the Rust Belt previously, and LA was known for very bad levels as well. I think it has less to do with us cleaning it up, and more with us outsourcing it.

Pollution per capita (and energy consumption) is at an all time low, yet the US manufacturing output is actually larger than it has ever been. China's in trouble specifically because it uses less efficient sources like coal for many purposes because it's cheap, which they're working on replacing with nuclear and wind power. However to truly solve this problem, they need a US style petroleum revolution utilizing shale (they're starting to do this). If you can lower the price of natural gas to the point where its more cost effective than coal like we did in the US, economics will go to work in your favor and people will start buying cleaner burning propane stoves and building natural gas power plants instead. Until then, though, not just the factories but the average person cooking food that will contribute (most people in China cook with primitive coal stoves: "Perhaps 80 percent (by population) of world exposure to particulates is indoor pollution in poor countries -- smoke from indoor cooking fires -- which need not involve any externality").

 

 

 

 

 

Additionally, US per capita energy consumption peaked in the 70s, and ever since then it has been decreasing despite massive increases in standards of living. Due to said gains it should be getting higher. I think you're being pessimistic here.

 

I am very aware that Western demand contributes to this. I would scale down everything I did if I thought that would impact things, but I'm just an individual, this would have to be a macro-level effort.

What exactly is your plan to fix China's environment (or, rather, to make them go faster than they're going now) and why are your ideas superior to those of a panel of well respected economists who more often than not say the basic premise of what you propose will result in a net negative?

 

I think the only sane way to reduce pollution will be to reduce resource use across the board, though that's not something our culture is accustomed to.

It's not something any culture is accustomed to because typically complaints about how everyone is using too much resources turn out to be incorrect. The human appetite for ever greater consumption hasn't stopped at any one of these points over the past million years:

 

 

 

I also think you're framing the situation in a dishonest manner. I don't consider trade restrictions based on practices that would be illegal in our country because of how harmful they are to the environment to be "bullying."

That's literally the definition of the word. Specifically, bullying is "to use superior strength or influence to intimidate someone, typically to force him or her to do what one wants". Threatening to wreck China's economy and reduce hundreds of millions of people to poverty unless they apply the environmental standards you want them to is indeed an example of that.

 

Furthermore, you're implying that 1.4 billion Chinese living in poverty is entirely the USA's responsibility with signing a trade agreement with them.

I never said it was entirely the fault of anyone. I just said that free trade has made sure China's average worker wage has risen four fold in two decades, that trying to make China cut down pollution quicker is not practical, and that the majority of economists disagree with your ideas.

 

Furthermore, calling it unethical, feels more like a blackmail tactic in discussing this. It reminds me of Congress calling awful legislation something positive. Like the Patriot Act taking away a massive amount of rights. Surely, you're a patriot and not anti-American, regardless of what's written in that act? I see this tactic as the same thing.

It's not a tactic, I'm stating the reality of the situation. If you cut back the level of trade with China, the average person gets much poorer. That is why those economists I cited said that refusing to liberalize trade with a country unless they apply the environmental standards you want was a bad idea.

 

I feel that entire framing is disingenuous. It's looking at one single element to a very complex problem and taking a "you're either part of the help, or part of the problem" attitude and is a way to shut down discussion and is not a process that leads to finding better solutions. If a modification was made to the agreement that we would agree to trade IF more stringent environmental laws were enforced, would that suddenly jeopardize the entire population of China, or more likely, would they still agree, but then the companies running these would make less profits because they would have to invest more money in cleanup and pollution control?

China's already reducing pollution. I think you're framing this issue disingenuously by saying that all that will happen is "the companies running these would make less profits"; one, you're underestimating the impact that will have (or at least I think you are, since I still don't know exactly what stricter standards you'd make a requirement), two... well let's assume that you're right, and that the companies in question will do what you say, AND that they'll still function well rather than being unable to afford the new regulations and go out of business. What then? Who's paying for this? "The companies" implied it would be entirely on the backs of the rich, but this is not the case. It will be the average Chinese worker. Losses in profits will be passed down to them, rather than just making sure that a rich businessman has to live in a slightly smaller mansion.

 

To reiterate what the economists and the Harvard papers said: it's not worth the cost, unless your proposed measures are really minor. As such, in a nation still developing like China, the best way to help the environment is to help make the people there wealthy enough to be able to afford to care about the environment. We're slowly approaching that stage.

 

The thing is, implying that not moving forward with trade that I believe will exacerbate the pollution situation I feel is extremely short sighted. You framed what I said as caring more about the environment than the people, which again, I find disingenuous and it's trying to make me look like the villain, when in reality I think what we're doing would not be a "solution" at all, except for the short term and would make things worse in the long term. Say we sign it and see prosperity among the Chinese people for a few years. How many are going to die from cancer because of the increased industry?

Much less than are saved in the long run (the long run includes a cleaner environment) by massively boosted wealth. The average life span in China rose from 71 to 77 between 1990 and 2014, after free trade really took off, despite horrible pollution. The increase in pollution wasn't anywhere near bad enough to offset the massive benefits gained from increased income across the board. And after 2014, their pollution started falling. China's playing the long game and it's working out well for them so far.

 

How many will be dead 20, or 40 years from now because of long-term damage to environment?

I don't know. How many?

 

Furthermore, does this establish a precedent that it's just okay to pollute for profit, making laws in the future more lax, leading to even more deaths from it?

Considering that China's pollution laws have been getting stricter and they've agreed to power themselves with 20% zero-emission sources by 2030? Nope. In fact continuing rather than hampering trade will help the environment, as it eventually becomes more profitable to not pollute that much. From that Harvard paper:

A key question is whether openness to international trade undermines national attempts at environmental regulation, through a “race to the bottom” effect. This no doubt happens sometimes. But there is little statistical evidence, across countries, that the unfavorable effects on average outweigh favorable “gains from trade” effects on measures of pollution, such as SO2 concentrations. If anything, the answer seems to be that favorable effects dominate.

Of course I don't want the Chinese people to suffer from it, but I think we're borrowing against the future if environmental factors aren't considered. I care about humanity as a whole, not just for today. Taking out a payday loan to pay for groceries and rent today is great for the short term, but creates far more problems in the long term, I see a lack of concern for the environment to be the same thing.

I would like comprehensive proof that what China is doing now won't be enough, and that threatening to withhold trade is a way to fix this. Because I'm sorry, the experts just don't agree with you here.

 

Anyway, I feel like this could go on for ages and you likely have many varied points, but I can't debate this sort of thing any further with the tactics you're using, sorry. Specifically, presenting what I consider 2 flawed approaches to a complex situation, then saying I'm unethical and don't care about people if I'm against option A.

That wasn't my intention. Why don't you tell me what your non-flawed approach would be?

 

Here's mine: continue free trade as this will result in the average person getting wealthier which will reduce pollution in the long term. Expand free trade to include relaxation on nuclear export regs so it's cheaper (China has a shortage of human capital to train engineers) to build, maintain, and operate nuclear plants, in the process creating a larger market for skills in the US.

 

Your accusatory approach is not how to win me over on anything.

That's a bit hypocritical considering you called the TPP "absolutely monstrous", wouldn't you say? What does that make people who support it?

 

Anyway, this is rather far off from my original post about compensation. I'll try to summarize my original point in a concise manner: excluding supervisory personnel (15-20% of the work force), using two different price deflators, and leaving out non-wage benefits gives a distorted picture, because the increase in supervisory personnel represent a shift of workers to higher paying jobs, the deflators clash with each other,* and a disproportionate amount of the new wealth has gone to non-wage benefits for workers (e.g. healthcare, paid vacation, pensions; all of them together are one third of the average worker's compensation), respectively. The reality is that compensation has been rising proportionately with production, per OECD and the St. Louis Federal Reserve (plus some economists who wrote an article).

 

*Specifically, the chart Ross used uses the Consumer Price Index which overstates the rate of inflation because of inadequate adjustment for quality changes and inadequate adjustment for substitution effects, while using another deflator for production. As OECD says, "If you want to compare productivity and pay fairly you should use the same deflator for both—that is, the GDP deflator" (also called the implicit price deflator). Making the simple correction of dividing earnings by the Implicit Price Deflator (IPD), instead of dividing by the CPI, already closes about a third of the gap.

 

 

I think that's all I have to say.

 

Well, besides this: props to danielsangeo for the posts about the unemployment rate. I think knowing how people come to incorrect or disingenuous numbers like "40% of people are unemployed" is extremely useful (and interesting). Otherwise they just get thrown around and you have no idea where they came from. I see that happening on the internet ALL THE TIME, so it's nice to know how to debunk the most common ones in detail.

 

EDIT: Spoilered the graphs. Post didn't need to be that big.

Edited March 27, 2016 by Guest (see edit history)

[Seattleite]

I see smoke on the horizon... I'm going to clear out before the flames get here.

[HLPrincess]

ahaha! WTF? Well said, mister!

[Binky The Rabbit]

More than just smoke...

[Heliocentrical]

Ross I really hate to say this but we can't stop ourselves from being screwed over. The people that have power to implement change don't care. Reality will be like a car with no airbags. Except people do have airbags it's just that the automotive factories won't implement them into the car. We will have to endure the full frontal blow of the car once it crashes and hope we don't wind up severely crippled or dead.

 

Now I'm not trying to depress you. In fact I fully encourage you to say I'm wrong and provide me with a foolproof plan for implementing something. The keyword there is "implementing" and not doing a simple excerise for how we might better ourselves. I want believe that change is possible. It just that I don't think it is. I've never had a point where I can say I was wrong for being pessimistic. I always went "Yep, saw that coming from a mile away. Not like I could do anything about it." and this has always the case. I can't see a reason not to be pessimistic. Sure it sad as hell but if pessimism allows me to see reality for what it truly is then why would should I to stop? Optimism only blinds you to only seeing the things you want to see.

[Seattleite]

A bit of a defeatist, aren't you?

[Heliocentrical]

A bit of a defeatist, aren't you?

I'm only being realistic here. A lot of "How can we make the world better?" ideas seem like exercises in futility to me because the people talking about them don't have any actual power to have them be implemented. If you think I'm in the wrong here then by all means challenge my conclusion because you're right I am a defeatist. Reality is incredibly defeating therefore I see no reason not to be.

[sp8472]

"We do not need to worry about robots becoming self-aware and overthrowing humanity. It's just not going to happen."

 

I'm also siding with the rich tech entrepreneurs on this.

 

- Self-awareness: Intelligence involves being able to hold a model of your environment in memory, and relate that model and reality. If the model includes the concept of its own memories as part of the model, that is the start of self-awareness.

 

- "Computers are really good at recognizing patterns, and carrying out instructions based on those patters. Now we have made enormous leaps in AI over the past few decades, but there are limits."

And then there were some things said about enormous barriers preventing progress beyond a certain point.

 

Simple machines, mechanical computers, vacuum tubes, transistors: Each was a breakthrough that permitted some manner of revolutionary progress. What's next? Quantum computers are making tiny steps of progress. One big disruption they'll provide first is encryption: Things that would take a regular computer millions of years might take minutes, or less. (And the Bitcoin market might experience disruptions. "...So, someone just made a million bitcoins. Yesterday. Interesting.") They'll be an enormous leap in computing power, and should get us the amount of power we need to make something rivaling a biological brain.

 

Concerning recognizing patterns: Yes, this is the path to intelligence and self-awareness. We recognize patterns well too. It's part of the reason we program AIs to do the same thing: It works for us, it's how we operate, and since we know how adequately that works for us, let's use it as a baseline. Your brain has the big advantage of having a huge number of neurons that can form a lot of connections. (Kind of what a quantum computer is heading for, with each individual element capable of being in multiple states simultaneously.)

 

 

The final piece is this: Emergent property.

If you have a complex system, you get behaviors you didn't plan on, caused by unexpected interactions of various components. This is where you can get things like intuition or even emotional states. If you're emotional, your thoughts become predisposed to follow a certain pattern. A whole society can be prone to this sort of effect, wherein the "mood" of the whole changes over time. I think you will encounter this sort of thing in a sufficiently complex computer system, to the point that it would not be easily differentiated from how a human mind works.

 

(I just don't think a human brain is all that big or amazing of a deal. It's got a huge number of neurons stuck together in a reasonably functional manner, and that complexity and numbering makes it difficult to simplify/model and still properly understand. It lacks any good debug port to plug into, so it remains mysterious as a metaphoric "black box." It's also been under development at the molecular and cellular level for a time period that's orders of magnitude longer than we've been working on our little computer systems.)

 

 

 

- Overthrowing humanity? You don't need intelligence to do that. There are numerous ways of killing, terrifying, and manipulating large numbers of human beings using methods that focus on our basic needs and instincts.

But let's assume it is intelligence at work, since we're talking about AI. Would it want power or control? That depends on how it's raised. If you raise a human child to be a remorseless killer primarily suited for combat, you shouldn't be surprised if it starts killing the "wrong" people, nor should you be surprised if it reacts poorly if you tell it that it needs to stop doing what you raised it to do.

[Presence]

Emergence is in my opinion the most important philosophical concept the average person can learn about. I'd go so far as to say that learning how r-pentomino works is like learning about the theory of evolution. You can learn the basics in ten minutes, and it doesn't just explain a complicated looking system, but complexity itself.

 

Thing is, I tend to think computers and programs work against emergent properties in themselves. It's inefficient to have data interacting with every other bit of data and also arbitrarily being processed by many different rules. Insanely inefficient. We're talking many orders of magnitude of speed loss for no useful reason. Programmers don't do that. There won't be a point when it's useful to ask a law manipulation program to be so broad in its data correlation that it could even potentially decide judge shopping is despicable. Maybe some year there will be genuinely opinionated systems, but not soon and not by accidental emergence.