Hacker News Reader

Comment by zellyn | Hacker News Reader

There are also real-world solutions. I spent some time brainstorming this a while back.

1) The safest solution, albeit somewhat expensive: launch a Voyager-like probe. It constantly generates private and public key pairs, broadcasting the public keys immediately and the private ones on a schedule. That way, you can pick a known timeframe for decryption. Launching more than one probe, for redundancy, is probably a good idea: you can encrypt with multiple keys.

2) Cheaper but less reliable solutions involve balloon-dropping transmitters that wake up and divulge decryption keys across a huge geographic area where it would be infeasible to find them. A variation is sinking modules very deep in the ocean that wake up and float to the surface.

View on HN

What about launching a probe that echoes any signals it gets? If it were sufficiently far away, one could send it the private key and destroy the local copy. Nobody's going to leave from Earth (or wherever) and catch up to your private key!

Since you're launching the probe well in advance of sending out the key, a sufficiently well-funded and a paranoid party could send a probe that just trails yours, intercepting your signal and beaming it back to the party in question (encrypted with their private key, of course.)

They already have the info, the problem is preventing the spreading (well, their problem on this scenario) so is just necessary to block the signal back

And even if they did, barring FTL, they couldn't really get it down here faster than the probe itself.

If the probe is destroyed so is the the possibility of decryption. Same for the second option.

In fact the problem is not related to time, but to the processing time and it's hard for two reasons, the math has to be so incredible/good/error free that going the physics way is cheaper and by cheaper I mean the only way (there is nothing more expensive than that thing it does not exist). Now Assumptions related to physics, which part do you assume as "hardcoded", (gwern decided for the cache speed; I found the problem on specialized hardware for it, AMD builds cpu on spec (back to physics))speed of light, and the amount of energy it takes to perform a job, resistance of circuits near absolute zero and therefore the amount of energy required to perform the job given that you can suppose on non paralelizable tasks therefore scalar cpu and therefore given a minimum scheme of circuitry the amount of time on a perfect conditions. This works unless reality(trademark) goes into play; -how much time will it take to humanity to hit the perfect conditions? -which conditions or values do you consider appropriate to measure the time (for releasing the information given the conditions are met?)

What if you just hide (like for example by burring them in the sand on the sahara) 100 devices from which 10 are the original ones and the other ones are backups of the original and you need ten of them to decrypt? (scale times 10^x for paranoia level)

That's why you launch a few of them. Also, the odd's of the probe getting destroyed e.g by an impact are really low. They get lower once you leave earth orbit and really low when you hit interstellar space. And you can probably get to less impact stuff quicker if you go perpendicular to the plain of the ecliptic rather than out.

Your bigger issue is component failure.

So launch a enough to deal with component failure, wait long enough for them to be out of harms way, and then you have a time lock.

Sure it's expensive, but time locks with a known time to unlock or even a decent estimate are not really possible algorithmically.

Excellent, but the point is to make the most secure/cheap solution, he already has a mountain a he could build a room with gigantic 100mt of steel walls with a piece of paper with the pass on it recorded by 100 webcams on 100 diff connections, or send a 10000 probes. Math for building an idea to hack is so good because ideas and intelligence to build them are so expensive the other parts of the problem look more feasible to hack/fix/try

Also it's cheaper than your options, also your options are not that "real world" factible without some serious money.

third tough, what if we take something people would do for something they desire a lot... money, what if we replicate the bitcoin schema and make use of all the hashes as required to decrypt the file? you do already have times for some real world calculations for that.

still playing the "hiding" card. put random players on online games with the answer and make them release on certain time.

That would be a cool use of a CubeSat (as long as it never becomes worthwhile for a wealthy entity to go snatch it out of orbit...)

Oops: s/balloon/parachute/

Reply on news.ycombinator.com