3 Comments

Very interesting! You have some really great ideas. I want these ideas to become popular. I'm not sure about the improvements to the Bussard ramjet, but I find the Caplan thruster interesting. Although the last time I read a third of the paper was a few days ago. Haha.

Expand full comment

On 'Population Growth Targeting': you could think that encouraging a lower interest rate might increase population through impacts on housing affordability (assuming supply also rises), more employment, and reducing opportunity costs of money spent on children (which might make people delay parenting until they are older and have accumulated more wealth).

On 'Deniable communication': If you are using a set of one-time pads for encryption, then in a simple case you can just use the xor of your alternative message with the ciphertext as a second padset. For example: you would send a plaintext message M1 that is encrypted to ciphertext as C = Pad1 ^ M1. You then have your second plaintext message M2 and have C = Pad2 ^ M2 -> Pad2 = C ^ M2. You could then exchange either Pad1 or Pad2 and it would seem to match two different messages. This gets more complicated if your pads need to be generated from some function (even if you set the keys) since you would need to basically invert the generating function to find the right secret key to get Pad2 which would probably make it vulnerable to a chosen plaintext attack. You would also potentially be leaking information between the messages, but I'm not sure.

Expand full comment
author

On population growth targeting: interesting I hadn't thought of the pronatal effects of inflation rates. Presumably the increased population growth due to low inflation would create deflationary pressure as kids grow up and join the workforce?

For the deniable communication: yeah that's a good way to avoid sending multiple messages to hide your original message! I think to maintain the deniability, the person you are conversing with will have to respond to both messages. But responding to the decoy messages can probably be done with language models.

Expand full comment