DAVID HASSELHOFF MAY HAVE BEEN AHEAD OF THE CURVE ON SELF-DRIVING CARS
What’s next: beachwear that helps you run in slow-motion?
Last Friday, May 4, an automobile accident in Chandler, Arizona, became international news.
Thankfully, the smashed-up cars made it look worse than it was; there was just one minor injury.
You know what also made it look worse than it was? The headlines, which focused on one of the vehicles being a self-driving Waymo minivan.
A few examples:
- A WAYMO SELF-DRIVING VAN WAS INVOLVED IN A CRASH IN ARIZONA – SF Gate.
- SELF-DRIVING CAR USING GOOGLE-OWNED WAYMO TECH INVOLVED IN CRASH – KFDM (Southeast Texas).
- WAYMO AUTONOMOUS VEHICLE, INVOLVED IN A CAR ACCIDENT IN CHANDLER, ARIZONA – Canadian Homesteading.
The headlines leave it unclear as to whether the self-driving car was at fault in the crash.
Compare the actual story with the headlines. A human-driven car swerved to avoid another car and jumped the middle line crashing right into the Waymo van. It seems clear that the Waymo vehicle literally had nothing causal to do with this crash.
It’s like someone crashed their car into a tree and the headline was, “TREE INVOLVED IN CAR CRASH.”
I saw this tweet from @RonAmadeo on the day of the accident pointing out this disconnect between the actual story and the framing in the headlines:
The crash immediately led to speculation that this could hurt the spread of autonomous-driving technology. @Wired‘s article contains the same account as the other stories on the accident, but this was the lede:
A danger to rational decision making is that we tend to judge the frequency of events, in part, by the ease with which we recall them (availability bias). Massive news coverage contributes to ease of recall. So, things that are heavily covered in the news are judged to be more frequent (and, as is the case with terrorism, more of a threat).
Some of the follow-up coverage has pointed out the irrationality of connecting such an accident to the safety of self-driving technology. @EricBoehm87 posted a good example of this on @Reason in an article titled “Don’t Blame Self-Driving Cars for Accidents Caused by Humans.” (He pointed out the same Wired article I noticed.)
Part of the problem – with public attitudes, which the media follows at least as much as it influences – may be that an autonomous vehicle is a “black box” that we can’t ascribe intentions to. Even though we’ve had millions of people die in car accidents because of the faulty decision processes of humans, at least we understand and identify with those processes. With a black box, we don’t really understand the decision process, and that’s scary to us.
We don’t like things that we can’t control and place a premium on being in control. There’s an interesting video on @BigThink by neuroscientist Tali Sharot (@AffectiveBrain) about how we’ll actually choose a more dangerous technology if we feel like we have control over it. (That’s why a lot more people fear flying than driving.)
Maybe @DavidHasselhoff was on to something in Knight Rider. Instead of riding around in the car silently, the car talks and is humanized, a real character in the show.
If you don’t humanize the car (certainly in the mid-eighties and probably even now), the silent, autonomous vehicle is likely cast as the villain. And definitely not as a helpful sidekick that sometimes saves the day.
NORTH KOREA, DONALD TRUMP, AND RESULTING
Diplomatic triumph? Positive contribution? Coincidence? Unnecessary danger?
The murkiness of cause-and-effect in international relations
My initial reaction to the news about developments with North Korea was, “Wow, President Trump’s strategy – clearly signaling a willingness to preemptively strike North Korea at any moment – must to be working.”
And I wasn’t the only one who thought that.
Then I saw a tweet, from a reader of Thinking in Bets, pointing out how they’d just been reading about resulting when all this news about North Korea and Trump was breaking.
Isn’t it ironic that I needed someone to tell me about my book for me to realize that I was resulting?
I took a step back to remind myself of the range of possible hypotheses in regards to North Korea. Here are a few of them, by no means comprehensive.
- Trump’s seeming willingness to preemptively strike North Korea and his unconventional rhetoric could be exactly what brought Kim Jong-un to the table. Without Trump, the peace process may not have ever begun as economic sanctions alone would not have done the trick. Trump scared Kim Jong-un to the table.
- Trump’s actions had a positive effect, speeding up the inevitable effect of economic sanctions in changing North Korea’s position. The economic sanctions would have eventually brought us to this place, but not as quickly.
- Economic sanctions made it inevitable that North Korea could not maintain its hostile, isolated position, and Trump’s public position had little effect on that.
- China or South Korea or some combination are primarily responsible for the change in North Korea’s attitude. (In turn, Trump may or may not have influenced or pressured China and/or South Korea.)
- Trump’s rhetoric had a negative impact, because it brought us closer to a nuclear war. In this scenario, the economic sanctions would have worked on their own and Trump’s aggressive stance is not what brought Kim Jong-un to the table and, in fact, the strategy made it more likely that we would end up at war.
Because these are all just possibilities (without regard for their probability), we also need to consider the possibility that a combination of these things is responsible – or that there are additional factors leading to the change in the situation.
That gets us to a constructive exercise of considering counterfactuals. How would things have been different (or potentially the same) if the president over the last sixteen months had been Mike Pence or Bernie Sanders or John Kasich or Hillary Clinton?
There’s a fantastic episode of @SamHarrisOrg’s podcast with Niall Ferguson (@NFergus) where they talk about the importance counterfactual thinking. In fact, it’s really impossible to think about history in any kind of rational way if you don’t think about counterfactuals. I highly recommend giving the episode a listen.
All of this is to remind me that I initially reacted without asking myself: Would I be willing to bet on this?
Once I asked myself that question, thanks to that tweet from a reader, I realized there are too many unknowns for me to ever have a great opinion on what is responsible for Kim Jong-un’s apparent change of heart.
Now my answer as to who deserves the credit is, “I’m not sure.”
They could save us, or destroy us, or both – and not necessarily in that order
John Brockman (@Edge), the publisher and editor of Edge.org, asks the world’s greatest thinkers an annual question. The question for 2017 was “What scientific term or concept ought to be more widely known?” Over 200 contributors presented answers, which appear on the website, and which Brockman published in 2018 in book form as, This Idea Is Brilliant: Lost, Overlooked, and Underappreciated Scientific Concepts Everyone Should Know.
John Tooby, a pioneer in the field of evolutionary psychology, wrote an essay on an aspect of tribalism, titled “Coalitional Instincts.”
Here are my take-aways:
- We have evolved neural programs for navigating the world of coalitions, a condition generally referred to as “tribalism” but which, as Tooby points out, “there is no mutually-agreed-upon term for this concept yet.”
- Most species can’t think coalitionally. He points out that, for example, among elephant seals, if the younger seals got together, they could certainly defeat the alpha. If they formed a coalition, they could do it, but they don’t do it.
- Our neural systems, earlier developed for individual instincts, now recognize an augmented reality of coalitions. “We project coalitions onto everything, even where they have no place, such as in science. We are identity-crazed.”
- He has a great take on extremism (“weirdness”) in coalitions: “Since coalitional programs evolved to promote the self-interest of the coalition’s membership (in dominance, status, legitimacy, resources, moral force, etc.), even coalitions whose organizing ideology originates (ostensibly) to promote human welfare often slide into the most extreme forms of oppression, in complete contradiction to the putative values of the group.”
- In fact, the best way to prove yourself in a coalition is to communicate extreme versions of the group’s beliefs.
- The great danger to coalitional instincts, as a consequence, is that they are antithetical to truth-seeking. He focuses on scientists, but I think this applies to truth-seeking in general. “Coalition-mindedness makes everyone, including scientists, far stupider in coalitional collectives than as individuals.” Belief revision is inconsistent with promoting the group’s beliefs. “To question or disagree with coalitional precepts, even for rational reasons, makes one a bad and immoral coalition member – at risk of losing job offers, one’s friends, and one’s cherished group identity. This freezes belief revision.”
I highly recommend reading the whole essay as I have surely not done it justice here. I think it should be required reading for anyone interested in being more rational.
A GOOD SUMMARY OF OUR TROUBLE
UNDERSTANDING PROBABILITIES AND RISKS
Subsequent to doing the podcast, Paul sent me a great piece on six systematic flaws in how humans think about risk from his book, The Science of Successful Organizational Change.
He gave me permission to summarize and excerpt the material, but you can also read it (in addition to, obviously, by ordering the book) using the “Look Inside” feature on Amazon.com.
- (1) We believe in trends and patterns in events that happen randomly. We make decisions on a random series of results that convinces us that we’re “hot” or “due.” (In fact, we could probably justify opposite decisions on the same reasoning: if we’re at a roulette table and black comes up seven times in a row, some people will bet on black on the next spin because it’s been coming up so much; others will use the same sequence to support betting on red because “it’s due.”)
- (2) We put our faith in small samples.
- (3) In difficult situations, our attitudes toward risk change. When we’re winning, we become risk averse. When losing, we seek out risk. (I witnessed this a lot in poker and was not immune to it myself: when players get ahead, they’ll take any excuse to quit and “book” the win; when losing, players will look for ways to get even and avoid the loss, including strategies they’d never consider if they didn’t impose a feeling of desperation on themselves.)
- (4) We aren’t very good at relative probabilities. (Gibbons even used the same example I used in a recent Smerconish.com article on self-driving cars and in a similar newsletter item: The 9-11 attacks scared people into making the much more dangerous choice of driving compared with flying, leading to a spike in deaths in auto accidents.)
- (5) We are irrationally conservative (risk-averse) in decisions involving tiny probabilities of catastrophic consequences.
- (6) We don’t understand “risk of ruin” very well.
THIS WEEK’S ILLUSION
The Scintillating Grid
Via the Centre for the Study of Perceptual Experience (@UofGCSPE), which has wonderful, detailed explanation, with numerous citations and images.