What new technology, for what institutions?

post by bhauth · 2023-05-14T17:33:28.889Z · LW · GW · 6 comments

This is a question post.

Contents

6 comments

Thinking about designs for new technology is something of a hobby of mine. For simple examples, you can see this battery design [LW · GW] or some of my blog posts. Even if you can't come up with something useful (and you probably can't) having a problem in mind is a good way to guide your learning about a field.

Whether or not my designs are good isn't the point of this post. It's a general question about what LW members think that people outside major institutions should be working on, and who would deserve access to things if they were actually interested. (I'm just a generalist, as half-assed as Eratosthenes, but there are presumably people in some private communities as smart as me but specialized.)

It shouldn't be surprising that big companies and big university labs aren't that interested in pursuing outside ideas that they can't take credit for. Getting new ideas implemented requires specific conditions, which is why inventions don't usually get implemented until they're obvious enough that a bunch of people think of them. But before you start trying to find a way to get some institution to use your invention, you should probably figure out if you should be trying to give/sell it to that particular institution. What's the point of new technology if it just gives institutions more room to be corrupt, right? My view is that modern technology and weak competition have relaxed constraints on some institutions, and resulted in them being unusually bad by historical standards. For example:

• I think American universities have become corrupt beyond the point where long-term recovery is possible. If you try to actually go somewhere, you'll find the wind is against you.
• A lot of professors are terrible people who game publication metrics, publish fake data, abuse grad students, etc.
• A lot of other American institutions are based around university credentials.
• A lot of corporate executives are terrible people who will, say, knowingly sell toxic compounds and say they're safe.
• Management at large corporations is full of moral mazes [? · GW].
• Current politicians are...current politicians.

This is a question I don't often see public debate about. Some people think of technological developments as natural events like rainstorms. Other people think of the progress of science as research in a Civilization game, where money and PhD-hours are used to fill up the requisite number of beakers with Science Fluid. But at least in my opinion, the limiting factor is usually "smart people trying to work on the problem", and smart people have some ability to decide what to work on.

Most of the time, when people suggest a topic to work on, one of the following applies:

• There's already a perfectly good solution that just wasn't marketed very well.
• The proposal is fundamentally misguided, and impractical no matter how good you are at design.
• I know somebody who made a good design, but large investments would be needed and investors don't have enough ability to evaluate technology. For example, here's Bill Gates showing off how well he understands physics. Oh, I could tell you some stories about dumb investors, if I could tell you them.

Yes, forum-posters tend to suggest stupid stuff like "you should work on a nuclear reactor that uses thorium" or "try to find some single chemical that fixes aging". Most people don't understand the shape of the current technological frontier, let alone the shape of what's plausibly feasible, but let's be real: neither do the high-level officials and executives determining what programs to direct research money into. A vote to spend billions of dollars on something silly instead of hydrogen fuel research isn't really any worse, and at least it's democratic. And the people here are smart, right?

Anyway, I'm curious what technologies the people here think should get some more design work done, and what institutions they (you) think should get those designs if they were interested. Designing stuff is just a hobby of mine, but it's somehow more fun if it seems like somebody would deserve to have it. If you can't think of any technologies, here are some hypothetical examples:

• You designed a new electromechanical actuator that could make robotic arms substantially cheaper, something like planetary roller screws or cycloidal drives. Who deserves to have it?
• You designed a very-low-cost long-range UAV. Say, something small enough to take off from civilian roads in Japan (impractical to destroy with missile strikes) but big enough to reach China from there with a sensor platform and a 500lb guided bomb. Who deserves to have it? Should the USA have it to deter China from taking Taiwan, or would giving the US government the ability to drone strike people really cheaply be bad?
• You designed a better neural network ASIC. Who deserves to have it?
• You designed a new AI architecture, maybe something about self-improving multi-agent systems. Who deserves to have it?
• You found a way that a bird flu could be modified to be highly infectious in humans and then make something fun like superantigens after a delay. Who deserves to know about that approach?

Answers

6 comments

Comments sorted by top scores.