2016 June 26 Sunday
What Is The Likelihood Of Human Extinction?

How could they possibly know?

A typical person is more than five times as likely to die in an extinction event as in a car crash, says a new report.

We obviously have a much better measure of our risks of dying in a car crash than in an extinction event. Since 1985 motor vehicle death rates in America have dropped by more than half. Cars have gotten a lot safer with better airbags, crumple zone design, and computers that assist driving. The death rate in cars will drop by an order of magnitude or more when autonomous driving technologies hit the market. Improvements in aircraft technologies have similarly made air travel much safer as well. Also, automated equipment is taking over many dangerous jobs. So your risk of death from normal accidents in an industrial civilization is going down. Your risk of death from, say, a star going supernova near Sol system isn't going down, at least not much.

Many factors make calculation of extinction risks hard to do. For really really low frequency events we do not have many data points. We can look back in Earth planetary history at the big extinction events of the past. But some of those causes are probably becoming less frequent due to natural geological processes (e.g. fewer asteroids flying around compared to 1 billion years ago). Plus, the big 5 historical extinctions are just 5 data points. We do not yet know for certain what caused the biggest one, the Permian extinction about 250 million years ago. The Siberian Traps volcanic eruption seems the likely cause. What are the odds of an eruption on the scale of the Siberian Traps happening again?

Suppose we could somehow know that something equivalent to the Siberian Traps was going to erupt Suppose we even had a 50 year warning. Could we prepare for it? Perhaps not save all the human race but save hundreds of millions with massive underground cities? Seems possible, especially if the cities are constructed far from the expected eruption location.

An eruption that fits the pattern of the Permian extinction would likely give us a lot of time (many lifetimes) to prepare. The Permian eruptions built up for 300,000 years before causing mass extinctions. Surely a super-advanced transhumans could come up with lots of strategies in a few hundred thousand years. Suppose at some point in the future we go thru a period of gradually rising amount of eruptions. Could we develop atmospheric cleaning tech to take massive quantities of carbon out of the atmosphere? If we could it would prevent ocean acidity from rising too high and therefore prevent massive ocean die-offs that make our current rising CO2 problem look like small potatoes in comparison.

Obviously we aren't being prudent about avoiding another mass extinction. If we wanted to be prudent then for starters a highly excellent asteroid detection and deflection defense system would be pretty easy, smaller than the Apollo program (which ate 1% of US GDP for years). Also, if we wanted to be prudent we'd stop overfishing the oceans, take steps to reduce fertility in the most rapidly growing countries, and reduce habitat destruction such as massive deforestation.

What about threats from our technology? Some people think nuclear war could make the human race become extinct. I'm pretty skeptical. Kill hundreds of millions of people? Possibly. Wipe out everyone? Way way less likely. An engineered pathogen seems within the realm of possibility. Ditto AI. Maybe nanotech goo.

If the human race goes extinct in 100 years what will cause it? Genetically engineered 200 IQ scientists who create AIs they can't control? Nanotech replicators that mutate into forms that let them wipe out all life? Or a gamma ray burst? Are we at greater risk from ourselves or from nature?

By Randall Parker 2016 June 26 07:30 PM 
Entry Permalink | Comments(3)
2016 June 24 Friday
Robotic Pizza And Delivery While It Cooks

In Mountain View California (i.e. Silicon Valley) start-up Zume is building pizza-making robots with the added twist of delivery trucks with pizza ovens. That way the pizza is really fresh when delivered. That's a pretty cool idea.

Pizza making seems a lot easier to automate than burger making. Stacking up the layers on a burger seems harder because burger toppings are more precariously placed than pizza toppings.

But think about what the possibilities in 10 years when all these food making robots are perfected. A single restaurant could have multiple kinds of robots and make a wider range of foods than a typical fast food restaurant. A burger maker, a taco maker, a hot dog maker, a pizza maker, a milk shake maker, a coffee maker, an automated wok and other robots could all fit in the same back room. So my guess is that we will have fewer restaurants, far fewer restaurant workers, and much greater variety per restaurant.

What I'm not sure about: will the fraction of meals eaten out go up? On the one hand the restaurants will be more automated and therefore cheaper. So easier to afford eating out. On the other hand, delivery will also be more automated and therefore cheaper. As a result, a larger range of restaurants will offer home delivery. So eat at home highly customized meals delivered by autonomous vehicles and robots? Or just use your home robot for cooking?

By Randall Parker 2016 June 24 08:55 PM 
Entry Permalink | Comments(5)
2016 June 12 Sunday
VC Funding Of Space Tech Soars

Once upon a time NASA was on the cutting edge of space technology development. That's so 20th century. Venture capitalists and entrepreneurs are on the cutting edge now: VCs Invested More in Space Startups Last Year Than in the Previous 15 Years Combined.

Elon Musk, Jeff Bezos, and VCs are moving the needle today. This is very good news. Since the Apollo program NASA has had no big success with rockets. The Space Shuttle was a bad design. Other stuff was just slow incremental improvements. Rockets that land themselves and satellites that collect enormous amounts of data for an assortment of commercial applications are changing the game.

However, we still a long way away from serious space colonization. We need a great many enabling technologies to make a Mars colony viable. Most notably, we need really big strides in biotech to enable small scale production of a large assortment of drugs, textiles, and structures in a Mars colony. We need economical small scale manufacturing (either with microorganisms or nanobots) to replicate much of the complexity of products that can be manufactured on planet Earth.

By Randall Parker 2016 June 12 09:21 PM 
Entry Permalink | Comments(5)
2016 June 11 Saturday
US $15 Minimum Wage Would Cost Wal-Mart $5 Billion Per Year

Less than Wal-Mart's $14+ billion per year net income.

I am fascinated by the minimum wage debate in the United States due to the impact that higher wage costs have on incentives to automate. If Wal-Mart starts paying $5 billion extra per year in minimum wage costs it still has another $9+ billion to spend on automation projects. One could get a lot done with just $1 billion per year.

With 2.2 million employees Wal-Mart is the biggest private employer in the United States, bigger than the next 4 combined. The retail trade is over 14.8 billion employees. If we roughly extrapolate from the Wal-Mart numbers retail would see about $30+ billion increase in costs from a $15/hour minimum wage. Dollars that large would fund a lot of robot development.

A big rise in minimum wage could cause a spike in lower class incomes. But over time the size of the spike would decrease as automation cuts the number of people working at, say, $15 per hour. How big the spike? Some businesses with thin margins (e.g. some restaurants) would shut down or immediately switch to lower service styles of operations (e.g. simpler restaurant menu itens). How fast the tapering afterward? Faster the case 10 or 20 years ago because we have orders of magnitude more computer power available to substitute for human labor.

My intuition about robots and automation: there are plenty of ways to automate when costs of lower skilled labor goes up. My post on car wash automation links to an article that shows how car wash technologies are old (not requiring the latest computers) and how a flood of low skilled immigrants actually caused a reversal away from automation. This suggests there is a lot of automation waiting to be used if only labor costs were to rise. Well, low skilled labor costs are headed up in some states due to higher minimum wage. So we are going to see more extensive utilization of existing automation technologies as well as development of many new ways to cut labor costs.

Our experience of life outside our homes is going to involve less interaction with retail workers and other service providers. This has already happened. Most of us rarely interact with gas pump attendants or bank tellers. Some of us rarely interact with department store clerks. I haven't been in a Target or Wal-Mart or similar store in years and some of my groceries are delivered. In the next 10 years retail store space is expected to shrink by a third or a half. Throw in higher wage costs and stores will shrink even faster and robots will wander their aisles stocking shelves and helping customers.

We face a very automated future.

By Randall Parker 2016 June 11 11:18 AM 
Entry Permalink | Comments(4)
2016 June 08 Wednesday
Who Should Get Autonomous Vehicles First? Teen Males

See U.S. Auto Regulator Says Self-Driving Cars Must Be Twice as Safe.

If high prices of the self-driving cars select for highly successful initial owners then the cars might not lower accidents rates much even if the car computers are twice as safe as the average driver. To get the maximum benefit from self-driving cars we need the most dangerous drivers to switch to them.

First of all, teenagers are very dangerous. Also, males are more than twice as dangerous as females.

In the United States, teenagers drive less than all but the oldest people, but their numbers of crashes and crash deaths are disproportionately high. 1 In the United States, the fatal crash rate per mile driven for 16-19 year-olds is nearly 3 times the rate for drivers ages 20 and over. Risk is highest at ages 16-17. In fact, the fatal crash rate per mile driven is nearly twice as high for 16-17 year-olds as it is for 18-19 year-olds.

Autonomous vehicles will become safer than teen male drivers years before they become safer than the average person the road.

Alcohol is a big problem as well. 32% of male drivers and 22% of female drivers killed behind the wheel have blood alcohol greater than 0.08%.

Once we have autonomous vehicles the age of first drivers license should be raised. Get those dangerout teens out from behind the wheel. They are a threat to us all.

By Randall Parker 2016 June 08 09:05 PM 
Entry Permalink | Comments(4)
2016 June 05 Sunday
Drones To Take Inventory At Wal-Mart Warehouses

By Q1 2017 Wal-Mart will be using drones to fly around inside warehouses to take inventory. The drones will take pictures from which bar codes will be extracted. What strikes me as odd: Why don't they know inventory at all times by exactly measuring what goes in and out?

In a warehouse where robots stock and unstock all shelves I do not expect these drones will be necessary. Ground-based robots moving around the warehouse will be able to take inventory as they add and remove boxes and items.

Automated store stocking seems more interesting and probably would save a lot more labor. Automated store stocking could also lead into automated robots picking up your items for you. We might see the rise of stories where robots pick your goods for you. Imagine you work a UI to choose what you want, pay, and then drive down to the local store to pick up your order. Will we see some local stores gradually transition to a local warehouse where you can pick up orders outside or pay extra to have them delivered to your house?

By Randall Parker 2016 June 05 01:40 PM 
Entry Permalink | Comments(3)
Minimum Wage And Car Wash Automation

An article in Reason on the effects that a $15 per hour minimum wage will have on car wash employment is a mini history of car wash automation. What is most surprising about it: the level of car wash automation has not steadily increased. When cheap illegal immigrant labor flooded into certain regions (notably New York City) the level of automation actually went down. This happened for multiple reasons including the ability to locate a non-automated car wash on a smaller piece of property.

A higher minimum wage (coming to California, New York State, and perhaps some other states) will bring back much higher levels of car wash automation in Los Angeles and New York City. It will also incentivize the development of automated means to clean inside the car.

As minimum wage rises in some pretty big jurisdictions we are going to get a clearer idea of what sorts of jobs are easiest to automate out of existence. Will the percentage drops in employment be higher in restaurants, hotels, car washes, grocery stores, vegetable farms, or fast oil change shops? Anyone have insights on sectors of the economy that would be easy to automate in response to a doubling of hourly rates for low skilled jobs?

By Randall Parker 2016 June 05 11:00 AM 
Entry Permalink | Comments(2)
Site Traffic Info
Site Copyright
The contents of this site are copyright