Predictions: Retrospective

Prediction blogs seem to be in vogue. I suppose the idea is if you can demonstrate a track record of correct predictions then maybe you have some real insight? Or got lucky a few times? In any case, making bad predictions would seem to indicate poor insight. So I’m going to do two prediction blogs. This one is retrospective. I’m mining my past public statements for predictions and grading them on accuracy. The next one is prospective, talking about stuff that’s coming up or things that people have asked me.

Starship architecture
In January 2017, I published a book on transport architecture to Mars. I did this in an attempt to anticipate the SpaceX Starship architecture before it was officially unveiled at the IAU conference that year. What I got right included a lifting body reusable upper stage that is refueled in orbit and capable of flying single stage return from Mars to Earth. Even in 2021 this part of the architecture is a bit controversial, but the book was written to explain and motivate it, starting from first principles. What I got wrong was mostly scale and execution. Starship is much much bigger (as it must be, to be useful) and does a tail flip maneuver to land, ensuring that its load paths are simplified and structural efficiency is improved.

In March 2015 I published this paper on Hyperloop routing. In it I recognize that the biggest obstacle to transsonic surface transportation is not levitation, propulsion, life support, vacuum, or any other exotic concern, but rather routing of the very low curvature right of way over the bumpy Earth. For generic use cases, nearly all the route has to be tunneled, which with current technology is extremely expensive. This is why, I believe, Boring Company came to exist. This fact is still not widely recognized.

The price of speed | Trains and Boats and Planes
There is a reason for the gap on the efficiency frontier at about 100 mph. Too slow for efficient flight, and too fast for typically bumpy terrain.

Grid battery dominance
In a number of blogs as early as 2018, I predicted that the PV+battery grid power modality would be able to control the market well enough to force spinning generation operators, mostly gas peaker plants, to lose money. I later discovered that Tesla had already built the battery “autobidder” software which can do just this. In 2021 the implications are still not widely appreciated, but they enable grid battery operators to systematically transfer revenue from the competition, if they want to. I also wrote that 2020 would see nameplate solar power at <$10/MWh, which proved to be a little optimistic. This plant managed $13.50/MWh. I also predicted that nuclear power would never overcome its decades of stagnation.

Carbon capture
I wrote in June 2020 that carbon capture and sequestration technology was impractical as the cost of capturing coal-produced carbon would always exceed the revenue derived from burning it. While I continue to believe that “clean coal” is total bullshit, I’ve more recently decided that electrical reduction of direct captured CO2 will be cost favorable to fracking by about 2026.

I had the opportunity to tour SpaceX in 2011 and soon after dumped my extremely meager life savings into Tesla stock. Later I sold a bit of it to fund my Green Card application. I lost track but I think overall that investment has done 120x or so over the last 10 years. I think Tesla still has a long way to go. It is hard to overstate how much better their entire tech stack is than any of the traditional automakers. Every few weeks we hear that Volvo has pledged to stop making ICE cars by 2035, or some other vague and boring goal. No-one will buy ICE cars well before then, the only real question is whether they make cars at all, by 2035.

In November 2019, I published a short note on Starlink. For the most part the calculations and predictions I made turned out very well.

Life on Mars
In December 2018, I wrote on Quora about reasons to suspect that there is life on Mars. I predict that there was life on Mars in the distant past, and that there is still life on Mars, in areas with geothermal heat. I predict that we will find evidence for past life in the next few years of Perseverance’s mission, ideally by the end of 2021. I predict that when we sequence the DNA of current life, some time before 2035, we will find strong evidence that Earth’s last universal common ancestor was a Mars organism. This view is actually not uncommon among interested scientists, though of course in one’s official capacity one must be somewhat sober about the current paucity of evidence.

7 thoughts on “Predictions: Retrospective

    1. Thank you and thank you! The interesting part is in devising the merit function to be optimized around. Short term revenue? Maximum lifetime revenue less depreciation? Maximum revenue less competitor’s revenue? Optimal medium term market position?

      Liked by 1 person

  1. “I’ve more recently decided that electrical reduction of direct captured CO2 will be cost favorable to fracking by about 2026.”

    Wow. I’m glad to hear that.

    Will this be useful to run intermittently, as a means of storage to offset the extreme variability of wind and solar, or will the capital cost be high enough relative to energy cost that it will need to run continuously to be economically viable?

    “I predict that there was life on Mars in the distant past”

    I don’t remember how much I’ve said about this somewhere on here, but my guess is that Mars never had life. When I imagine how life could have begun, it seems hard. Life consists of a metabolism that can support the genetics, and a genetics that can support the metabolism. It’s a severe chicken-and-egg problem. Not insurmountable, but when I try to imagine how it could be surmounted I have to appeal to conditions that probably never were in place on Mars.


    1. Some of the molecules found by Curiosity seem to only occur as a result of biological metabolism. This is VERY VERY hard to discern, and the team of scientists trying to replicate the spectra obtained on Mars have been working very hard for years. Hence, some significant error bars, so nobody is standing on the rooftops yelling “Eureka!!” just yet.


  2. High speed rail (the Japanese & European versions) seems to be missing in your chart on speed vs. power. I’m thinking railroads in general should be the most efficient form of transportation, but it might mess up your chart a little bit. Higher ground velocities require more expense in road/track preparation, might need a 3rd dimension on your chart 🙂


  3. Also there is a Soviet/Russian/Belarus technology that is mostly unknown in the West. It’s called SkyWay (string transport) and it’s an innovative take on railroads. There is also a completely redone technology stack and there is a test park in Belarus where the tech is demonstrated. There is now a pilot project under construction in UAE.

    It appears that SkyWay is 5-10 times more effective than railroads, meaning it’s a contender to dominate transportation globally, replacing railroads and planes.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s