May 2019 – Siouxneil

A Nuanced View on Automation & AI

**Automation won’t be as gnarly as the apocalyptic views. A more nuanced view from Brookings and McKinsey.** Photo by Katarzyna Pe on Unsplash

I read through the summary of the report on automation by Brookings and McKinsey. It is worth reviewing because it is a realistic, nuanced view of the dangers and opportunities of automation.

It feels like the conversation is moving closer to what we had before the big advances in neural networks. Those advances, especially Alpha Go Zero, set off a frenzy of speculation that all our jobs were doomed and an AI superintelligence would subjugate us puny-brained humans. I think we are realizing that while the latest advances are amazing, they are still brittle. Alpha Go Zero can’t drive a truck or even tie shoelaces, even if it can whip anyone’s ass at Go.

It feels more likely that ten years from now we will have really powerful pattern matching machines that can do things like drive trucks and tie shoelaces. Yet those machines will probably suck at playing Go.

Are You the Reason it is “Art?”

The reason art moves you might connected to your thinking about yourself. Photo by Martino Pietropoli on Unsplash

Brain science is uncovering clues on what makes art a moving experience and it is pointing to you. Turns out that when people report having the most moving experiences with art — albeit while inside a machine with powerful magnets whirling around them (an fMRI machine) — they activate their default mode network. That circuit in the brain has been shown to also be activated during self-reflection about oneself, the past, the future, and social relationships. I think of it as the daydream circuit.

This paper, “The brain on art: intense aesthetic experience activates the default mode network” is one that goes into the detail of the findings. It is part of a new field of neuroaesthetics that aims to understand why we find things beautiful and moving. It is an integration of psychology, neuroscience, and evolutionary theory.

I recently read a related book, Reductionism in Art and Brain Science [excerpt here] that goes into the core questions of why we are attracted to art. Eric Kandel points out the importance of the bottom up and top down processing by the brain. Bottom up processes are ones that favor symmetry, certain colors, and shapes. Top down start with culture and ideology and continue through influences of family, peer group and personal experiences.

These are fascinating results coming out of the integration of fields that used to be totally separate. I’m excited to watch this space develop.

Taking the Nature Pill

There is growing evidence that spending time in nature is good for us. I first learned of this research through a Hidden Brain episode, then Mera looked up some of the research for a Whisky Papers, and recently researchers at UMich showed evidence of dosing.

How much nature is enough? The answer was that stress hormones decline a lot after 20-30 minutes but much after that “dose.” (Note: this was a 1st study with small sample size, so it is still, you know, early science. I.e. could be totally wrong) The idea is to get to a recommended dose for a “Nature Pill.” In the study, they took the pill 3 times a week.

This morning I decided to try it out for myself. Instead of working out, I took a slow, mindful walk through Golden Gate Park. It was a tiny stroll. I purposely slowed down. This was not about exercise, I reminded myself. This was about … something else that is also important. Connecting w/ nature and simply being there. What is the experience of nature? What is the experience of me? And of me in nature.

Some things I noticed:

It is hard to disconnect from the mechanical world. In San Francisco, even inside a park, there are sounds of cars, trucks, and rail from the roads only hundreds of meters away. Jets roar overhead. You have to walk into the woods to allow the bird song and crunch of the trail compete with the crunch of the mechanical.

It is hard to disconnect from the information world. I have gotten better about setting borders from info tech and my life, but it was so hard to resist the pull of the smartphone. The early part of my walk was fine — great really. There I was, walking slowly among the trees, the grass, and flowers listening to birds. What were these small birds I saw? They seemed different from the buntings and other small birds I’ve seen in my yard. After about ten minutes, my curiosity turned to a kind of mild anxiety. Couldn’t find out? I have a bird identification app. Surely I could violate my tech border for this. It was something connected to nature after all! I pulled out my phone, fiddled with it a bit and in the process realized that just that small act removed me from the experience. Even the few photos I snapped on my phone pulled me out of the experience of being there.

In the University of Michigan study, subjects were asked not to use their phones while taking their nature “pill” (they also were not allowed conversations or reading). There was no control of people just sitting for 20 minutes without using their phones, reading, or talking. I wonder if that intervention alone can improve stress levels.

The experience of being in nature really is calming. I know this sounds obvious to most people. Yet, I don’t build a nature pill into my week. I bet most people don’t. Today’s experience was enough to convince me. I’m going to try 3-4 days of nature pills into my weekly habits.

Borders >> Productivity

I am trying to create more borders. In today’s world, laptops, smartphones, email, Slack, and text make it so easy to merge our work lives and off-duty lives. It really requires discipline to wall off portions of my life from the pressure of being online, on email, on Slack, and on social media. I can’t say I’m doing great, but here are things I’m trying:

Check email only once or twice a day. Importantly, setting expectations by putting a footer on each email, “Text if urgent – I aspire to check email once a weekday.”
Do a shut down process at the end of my work day where I assess what went well and what is high priority for the next day.
Setting a timer when I engage in social media and other fun diversions.
Establish a day during the weekend when I avoid email and being online.

I find my life is more creative and focused when I create time away from the computer. It opens up space for fun, family, & friends.

I think I can also make the case that it makes me more productive and focused. Obviously, it reduces the total time I spend “working” but the real question is what kind of work. With this approach, I end up doing more deep work.

How Tesla Insurance Could Work

It is simple, really…More Autopilot = Less Accidents

Most commentary [pay wall] about Tesla’s new insurance point to the data that the company can gather from customers’ cars, an idea pioneered by other insurance companies like Progressive and Metromile. Could Tesla know even more? Of course. They probably can not, however, improve dramatically on other insurance companies who have been measuring car data for over a decade. Also, personal insurance is often limited by states like California in how they can use that data. There will be some information, like video from cameras, that should make it easier to identify who is at fault, which would be a minor advantage. Claims adjustment is a minor cost for an insurance program.

Everyone, however, is missing a basic insight. As Tesla cars become more autonomous, they will get in fewer, and more minor accidents. That is a big deal. The biggest costs in insurance are collisions, especially major ones, and especially the ones involving injury and death. This is an advantage that Progressive et al will never have unless they join with an autonomous car company.

There is more to the story. As more people use autopilot for more and more miles, the liability for those miles is going to accrue mostly to Tesla, not the driver. Just like anti-lock brakes and other safety equipment, the maker of autonomous features will be liable in case it doesn’t perform properly. Less liability for the driver means less liability for their insurance company.

It all adds up to declining risk for the insurance company that stands behind the driver. It does, however, mean more risk for Tesla shareholders, but if I’m Tesla, I might as well get a benefit from that that additional risk. Tesla Motors can offer it’s customers cheaper insurance while Tesla Insurance can profit from declining claims.

Oh, and perhaps another bonus to having Tesla Insurance. Guess who often pays for the lawsuit when there is a dispute between the driver and the automobile company? Yup, the insurance company. There might be legal limits on how cozy that relationship can be (and if there aren’t there should be). Even with regulation, though, it seems inevitable that Tesla Insurance will be biased against bringing suit against Tesla Motors.

I’m a bit surprised that most observers don’t get the obvious connection. This is likely a fruitful direction for Tesla and it plays to their focus on autonomy.

About the author, Sunil Paul. He was co-founder and CEO of the company that invented ridesharing (Sidecar), helped incubate Getaround at Singularity University, sponsored the first law to protect peer-to-peer car sharing, and was on the board of one of the first car sharing projects. Disclosure: I own a small amount of Tesla stock, which is not material to my net worth.