2016 August 27 Saturday
Global Warming: We Will Need Climate Engineering

By 2060 Phoenix Arizona will have 132 days a year over 100F. Dallas will 55 and Pecos Texas 101 days.

My view about problems: We should solve them. If the Earth really does heat up substantially then we should pull the CO2 back out of the atmosphere while also releasing cooling gases. If its practical we should also raise the albedo (surface reflectivity) of the planet.

Right now cities should change their zoning laws and roads policies to make buildings, roads, and other surfaces more reflective. No more dark buildings. Use light colors of concrete, white shingles, and other surfaces that reflect more light. That would be beneficial even if the Earth was not heating up. Hot cities are unpleasant places to live.

I could tell you we should do big build-ups of nuclear, wind, and solar power. But right now the will does not exist to take radical steps. Luckily, wind and solar are getting much cheaper. Unluckily, nuclear does not appear to be getting cheaper. Batteries for electric vehicles are getting cheaper but at much too slow a pace to cause a slashing in demand for fossil fuels.

The human race has a number of big problems that have to get much worse before they'll be addressed effectively.

By Randall Parker 2016 August 27 02:27 PM 
Entry Permalink | Comments(12)
2016 August 18 Thursday
Uber Buys Autonomous Truck Company Otto

Uber also made a deal with Volvo to cooperate on self-driving cars.

Self-driving cars are an existential threat to Uber if Uber doesn't develop them first. Suppose Ford makes self-driving cars viable a few years before does and rolls them out in many cities. Uber gets wiped out by Ford's ability to charge less for a ride. Uber's big competitive advantage from a large set of recruited drivers could vanish as fast as sufficiently safe autonomous vehicles can get manufactured.

Sufficiently safe: that's the challenge. But autonomous vehicle makers who limit themselves to some urban markets can lower the bar for their initial roll-out by just excluding any streets and corners that are too tricky to handle. They'll lose some business to be sure. But better get to market first.

Uber is in a tough situation even if they manage to get the first autonomous taxi to market. Why? Because autonomous taxis lower the barriers to entry for competitors. Many companies are rushing to build autonomous vehicles. Toyota has deep pockets. Even with the diesel engine scandal VW still has deep pockets. Daimler does too. These companies can make millions of cars per year and can flood urban areas with fleets of autonomous taxis that they own.

By Randall Parker 2016 August 18 09:03 PM 
Entry Permalink | Comments(2)
2016 August 16 Tuesday
Ford Self-Driving Car By 2021 For Fleet Use

2021 for taxi-style usage in selected areas. This is level 4 automation: no human driver will be needed to take over. But the car won't be able to go everywhere. Likely that means the cars will be restricted to very well mapped areas without challenging conditions.

Ford says the higher cost of the computer and sensing equipment restricts its use to fleets which rack up very high mileage per vehicle per year. The car will be a more expensive piece of capital equipment that requires very high usage rates to pay the cost of capital.

I find the 2021 launch date to be a little surprising since Ford seems late to the party. On the other companies seem think they can prepare products to make them ready for market fairly quickly. For example, Delphi is trying to roll out a self-driving taxi service in Singapore in 2019 with a limited route and with greater capabilities later. MIT spin-out start-up nuTonomy is also working toward building an autonomous taxi fleet in Singapore. All of these companies are working toward deployment in limited driving areas and for taxi-style usage.

If the initial costs aren't too high then Cadillac, Mercedes, and other luxury car makers might come to market with level 4 autonomy near the time when autonomous taxis come on the market. If the luxury makers can support operation more areas than the taxi makers then elderly and otherwise disabled but affluent people would make a decent-sized market as well as on-the-go successful business types who want to spend their travel time working. If you are making millions of dollars per year then a $100k to $150k autonomous vehicle that freed up 30-90 minutes of commuting time for working could offer very good value.

Looks like fully autonomous vehicles for personal ownership are still 10 years out for most of us. Though a MotorTrend comparison of 4 partially autonomous vehicles shows we will get benefits short of full autonomy before then. A thought that struck me while reading the article: I want to be able to tune my future autonomous vehicle for greater safety. For example, tune it to not follow the car in front of me as closely as standard factory default specifies.

By Randall Parker 2016 August 16 08:56 PM 
Entry Permalink | Comments(5)
2016 August 07 Sunday
High Minimum Wage And The Decline Of Stores

The incentive to automate will be enormous for $15 per hour minimum wage.

Of America’s nearly 16 million retail workers, the biggest group — 4.6 million — are salespeople. Their average wage is $10.47 an hour. After that, the country has another 3.4 million cashiers, and their average wage is $9.28 an hour. Only a quarter of salespeople earn more than $14 — and only 10 percent earn more than $19. The figures are worse for cashiers.

But in the race to automate there will be a clear winner: Amazon. Why: Amazon can automate more easily than can physical stores. It is analogous to why long haul trucking can be automated before taxis: Just as highways are simpler places than city streets so warehouses are simpler places that stores. Warehouses, in theory, do not need human beings inside them. They can be laid out to make fetching and boxing of goods easier for robots. Stores can't do that.

Basically, Jeff Bezos gets a competitive advantage from higher minimum wages. He can adapt to them more easily than can Wal-Mart, Target, Costco, Kroger, or Home Depot. Amazon's competitors should treat automation as a desperate need.

By Randall Parker 2016 August 07 07:51 PM 
Entry Permalink | Comments(8)
Massive Layoffs When Trucks Become Autonomous

1.8 million people in the United States drive heavy trucks for a living and are at risk of losing their jobs when trucks become autonomous. That number is from the BLS category heavy and tractor-trailer trucking with 1.8 million employees. A separate category Delivery Truck Drivers and Driver/Sales Workers has 1.3 million workers.

The heavy duty truckers are more at risk than the local delivery drivers because it is easier to automate long haul driving on interstates than to automate driving on more complex (cross traffic, pedestrians, parked cars, etc) local roads. Plus, delivery drivers have to run up to houses and businesses to make most deliveries. Building robots to do that work will take longer.

Railroad operation is easier to automate and BLS thinks railroad employment will shrink in coming years. Railroads employed 113,000 people in 2014 and that's expected to shrink 3% by 2024 even as rail volume rises.

233,700 people worked as taxi drivers and chauffeurs in 2014 with a 14% rise expected by 2024. That all depends on when autonomous vehicles hit the market that can handle local surface roads. At some point taxi driving is going to drop sharply as fully autonomous vehicles take over. But when? Some auto industry executives predict introduction full autonomy not happening until 2030 and even later.

My guess is that current 55 year old truckers will retire from their jobs. But 45 year old truckers are going to get laid off before retiring.

By Randall Parker 2016 August 07 06:51 PM 
Entry Permalink | Comments(1)
Global AI Key To Everyone Feeling Unoffended?

Some people have become incredibly sensitive about perceived slights. I think that's an unhealthy situation. But might societies segment very heavily along dividing lines of views about life? AI might make it possible. Imagine no matter where you went you only ran into people who agreed with you or at least were fairly supportive. A global artificial intelligence could make that a reality.

If a global AI understood the preferences, moral reactions, and beliefs of everyone on the planet it could intervene to greatly reduce the incidence of people with conflicting views from ever coming into contact with each other. A global AI could, for example, provide people with information about where to live, work, and vacation that would reduce contact with people with distasteful views. For example, hotel and flight openings could be shown to only have people to leisure travel when and where people with similar tastes and beliefs will go.

General awareness of people holding conflicting views could be greatly reduced. Anyone so inclined (or even by default) could have a custom news channel that filters out not only conflicting views but even upsetting events.

As Bill Bishop showed in his book The Big Sort, people are migrating to live around people who are more like them. But AI could take like-minded cocooning to a much higher level. One could set preferences for kinds of people to avoid and the options one would get for schools, employers, cities to live, places to vacation, and other parts of one's life experience could be filtered to give one a rather homogeneous experience by a variety of dimensions. For some people this would mean a rather placid existence.

Once AI becomes powerful enough the question is who will do the selecting. AIs might just secretly decide to encourage some factions to migrate away from others while deciding that some groupings of different kinds of people can mingle and work and live together.

AIs could carry out interventions to influence views on a scale unheard of today because they could literally write unique news reports for every single person on the planet. You could compare notes with some other people who you meet (who won't be random samples) to look for discrepancies. But if your news feed has tons of news on one area of the world and another person's has less on that area and they don't read that much on that area in the first place it might seem normal for them to be uninterested in that place.

What will make this easier for AIs to do: VR as a fun place to spend lots of time. If AIs can make the real world seem far less interesting than VR then people won't even care much about the real world. So they won't be interested in omissions about the real world. People could even be presented with boring news reports about the real world mixed in with exciting news about VR such as things VR celebrities did or are going to do.

If rejuvenation therapies enable you to be alive 100 years from now will you be lulled into complacency, separated from people with conflicting views, and kept in the dark about parts of the real world that AIs don't want you to think about?

By Randall Parker 2016 August 07 02:21 PM 
Entry Permalink | Comments(2)
Site Traffic Info
Site Copyright
The contents of this site are copyright ©