[ExI] A small step towards brain emulation

[Anders Sandberg]

On 21/11/2012 07:17, Giulio Prisco wrote:
> Anders "the probabilities get serious for decent uploads around 2050,
> and merely working ones in 2037"
>
> This sounds like sweet music to me, especially because there is a
> little possibility that I am still alive in 2050, and a more solid
> possibility that I am still alive in 2037.
>
> But my gut feeling is that it will take longer, perhaps much longer.
> Perhaps the first experimental uploads around the end of the century.
> I hope I am wrong, and I would love to hear solid arguments in support
> of Anders' timeline.

I really should publish my full analysis as a tech report, especially 
since Stuart Armstrong praised it at the Singularity Summit. It should 
of course be taken with a mountain of salt. Basically it is a Monte 
Carlo model of different scenarios of Moore's law, the complexity of the 
brain, progress in neuroscience and scanning tech.

It is worth remembering that I get a probability distribution with my 
full model, not a single date: the previous post on this list was a back 
of the envelope sketch based on single numbers. The full distribution 
looks a bit like a lognormal distribution starting to build height in 
the 2030s and having a peak mid-century, with a very very long tail 
stretching into the future.

I'll promise I will work on a report when I get some time after the big 
AGI conference in December.

But, yeah, we should get cryonics contracts and work on other ways of 
life extension. Uploading is not going to be easy.

-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University