Tom MacWright


Busy month! In January, I wrote about the first four Val Town runtimes, and built a lot of features in Val Town.

I also hacked around enough to make Placemark Play available again, the “free-as-in-beer” Placemark UI. It still has some warts, but it’s there. Want to improve it? It’s open source, so you can!


I’ve been listening to this new-to-me CocoRosie album. I fear that this is millenial music through and through - they got popular because of the music blogs, and their sound is almost an in-joke. But, like the rest of their albums, this one has a few genuine bangers.

I’m late to the Oneohtrix Point Never fascination. I came to them like probably others have - their soundtrack score for Good time.


I’ve been watching Ozark, which is pretty good. It’s so intensely color-graded that my cheap TV had trouble rendering the dynamic range of those near-black tones. I also watched A Serious Man, which is a hit.


I read The Fund, Rob Copeland’s book about Bridgewater Capital. The book is 90% a story about personal squabbles and Bridgewater’s sociopathic management culture, and 10% about Bridgewater’s actual investment management performance and style – which is intentional, because the firm seemed to spend a lot more time on follies than its actual work. Pretty good book, could’ve been shorter. I suspect it’ll become a movie or a TV show, and I’d watch either.

Boy, I really didn’t like In The Shadow of Silicon Valley. The tech industry is responsible for so many ills of society - invasions of privacy, extractive gig-jobs, spreading disinformation, the list goes on. But people are just aching to blame them for San Francisco’s housing crisis, and that’s the one thing that they aren’t responsible for. Blame Diane Feinstein for downzoning the city. Blame Howard Jarvis for gutting the tax base and subsidizing rich homeowners. Blame Dean Preston for stifling housing supply, including fully-subsidized developments. Or in LA, Michael Weinstein for bankrolling NIMBY laws. There are villains on the left and right alike who have directly set up the conditions that are creating homelessness and increasing rents. There’s no sense in trying to string together some vibes-based theory of how tech is to blame.

I really enjoyed Brian Merchant’s Torching the Google Car, and didn’t enjoy Eric Newcomer’s opposing article in favor of self-driving cars. I do, though, like a lot of Eric Newcomer’s writing and have been a subscriber to his Substack in the past. And I also don’t universally like the content of Brian’s Blood in the Machine newsletter. So, in other words - let’s have a friendly conversation!

I think the gist of my feeling is this:

Self-driving cars have been over-promised and they’ve failed so far. I documented this in a visualization before: basically every car company promised to have a self-driving car by now. Everyone missed the mark, massively. They vastly underestimated the challenge of building such a thing, and all the macho “machine learning for self driving cars” engineers collected fat paychecks but didn’t create the thing they were paid to create.

No, the “self-driving cars” in San Francisco don’t count: they have remote drivers, who have to take over often, and they are only approved to operate in a small, highly-mapped area. They also aren’t production vehicles and haven’t met Level 5 qualifications. They are the mechanical turk: you can fool yourself into thinking there’s no guy operating the chessboard from within the suspiciously-large box below it, but the guy is there.

And now, having failed to produce a self-driving car using LIDAR, machine learning, cameras, high-resolution maps, these companies want permission to use our roads and pedestrians as training data. With all due respect, they aren’t very good at this. Why do they deserve to put people in danger?

Our streets are thoroughly political and we are so fucking tired of people dying because car companies and drivers get their way, every time. The things that safe-streets advocates have been talking about for years haven’t changed: the speed and size of cars needs to decrease. This is obvious. There is no discussion necessary. Cities that lower speed limits have fewer deaths. It has been working for Hoboken, which has reduced the city speed limit to 20mph and achieved zero pedestrian deaths in years. Bigger cars kill more. We know which streets are dangerous and how we’d fix them, but our leaders aren’t fixing them and people continue to die.

I don’t know. It’s tiring, there is only so long that you can shout and point at the things that are so obviously wrong and so clearly fixable, and be met with silence and backsliding. Cars are getting more dangerous for pedestrians and the National Transportation Safety Board is allowing it. Regulators from the NTSB are concerned, they say, but not concerned enough to do anything, because that responsibility falls in the hands of the NHTSA, which also doesn’t do anything. Thoughts and prayers to anyone not tall enough to be seen in front of your monster pickup truck.

And, yeah, I don’t know - maybe the two aren’t in conflict. But I must say, in conversations I’ve had, they are. Both car-fans and electeds argue that soon we don’t need safer streets or better public transit because self-driving cars will perfect the art of driving. And a suspicious number of people who talk about the urgency of pedestrian deaths as a reason to support self-driving cars are silent about the more basic reforms that could happen today. This includes Newcomer, who I hope is more vocal about the basics in the future.

Anyway, the thing with self-driving car companies is that they don’t need your tweets of approval, your supportive articles, your pro bono trolling in the replyguy trollcaves. They have lobbyists and money, and longstanding relationships with regulators. Money wins.

There’s no upside for safe streets advocates, nothing to sell. We’ve been saying the same thing for years. The only upside is that people would die less if better policies are implemented. Is that enough?