Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

You know what would be really cool? Write a mini-me program, with enough money that hosting can be funded on interest, with a little left over for gifts.

Then, you die (painlessly, after a long and happy life, etc. etc.)

The program keeps running, tracking your descendants over time and gives them little random, appropriate gifts from the ghost of great^n grandpa or grandma.

I wrote about this in 2007 (https://tubelite.wordpress.com/2007/09/10/autonomous-softwar...). Perhaps Facebook is in the best position to do this, help people plan and create their ghosts.



Let's say that you save 10'000 $ in an account with 5% interest, then you'll have 500 $ a year to spend on presents. If every child has two children the amount children in a generation is 2^n. When they grow up we stop giving them presents for mathematical simplicity. Then it's possible to give every child in the 9th generation a $1 gift a year. Which is quite neat. If we have a generation length of between 25 and 30 years this will be between 225 and 270 years in the future.


In the long run, it is highly likely that you will either be an ancestor of every single human, or none of them. (http://www.bbc.co.uk/news/magazine-19331938) You don't need to give every single one of your descendants a gift. Give 2-5 gifts a year, randomized after filtering out undesirables using a naughty/nice algorithm.

Assuming total population remains roughly the same, if enough people did this, it would actually work. Most kids would get gifts from grand^n Santa.


Nice idea and an interesting article. If the goal is to be remembered for as long as possible giving more presents is obviously better. After 9 generations only a small part

But some might feel that a enormous list of children with ratings of their behavior would be questionable.


However, the site will never keep the API's the same. The mini-program may stop functioning a few years after the person's death, and there will be nobody to change it


True. The specific APIs will go out of date. That is why the intent has to be captured at a higher level, while the implementation of the intent can be swapped out and replaced as APIs change. Many aspects of human behaviour have basically remained the same in hundreds of years - giving a gift, helping out in a good cause, though the specific mechanism of searching and giving of gifts has changed.

Organizations, trusts and even countries have been running for hundreds of years, based on a limited set of rules written in a programming language called LEGALESE, updated from time to time. One way to look at the mini-me program is as a "software trustee", whose parameters are monitored once in a while by a programmer-lawyer human. This approach yields the same scale-out/low cost/automation benefits of the cloud, letting a single programmer-lawyer maintain hundreds of trusts, in much the same way as a single admin can manage hundreds of instances in the cloud. It reduces the cost of creating and administering a trust to the point where such options are available beyond the top 1%.

Think of virtualization, emulators of old video game machines... they haven't been manufactured in ages, and yet, we find a way to faithfully run them inside emulators. I think we'll find a way to translate the Amazon APIs while remaining faithful to the original human's intent.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: