Well, jeepers. My initial judgment, based on Lux’s summary alone, was that the project seemed dangerous and maybe a little evil; on closer inspection, the project seems dangerous and maybe a little evil!
On the first point, this is a risk even when one brings life into the world the old fashioned way. On the second (and no one should underestimate how seriously Keyholez takes the charge of uncoolness), I think I’ve made it clear how I feel about souls. As explained by Keyholez2009:
This has to do with being a materialist — i.e. not a religion-crazed imbecile. Where do you, o wise and honorable Reverend Greif, think love happens? The soul, you little muppet? […]
These people call themselves intellectuals, but they have yet to adjust their worldview to align with the basic scientific discoveries of the last several centuries, so they end up accidentally spouting nonsense, invoking some ghostly person-essence that does all the loving and the smooching, totally independent from nasty, Scrooge-like “brain chemistry,” which no one should trust their daughter with. [Emphasis added]
Here’s the hypothetical I’d like to run through: We fabricate this soul in this box and run the torturous program once. So, we do a little bit of evil. Then we do it again and again and again. More evil? Now, what if we put the soul in the box and just run a simulation that is totally awesome (or design a brain that enjoys what another considers torture, but we’ll leave that alone for the moment). Negative evil (“goodness”?) points? What if we run more happy simulations than we do sad ones and still come out with more insight into the operation of the brain? Net positive on the good/evil thing and science points! Everybody wins!
The point is that ethics gets pretty murky if you can just start generating more utils by running more supercomputers no matter what you do with that capability. Hey, what if we can simulate trillions of happy existences for the same resources it would take to provide for a million “real” people over the course of their lives? If the soul in the box is equally ethically salient (as Keyholez2010 is suggesting) we’d be remiss for wasting our money on inefficient version 1.0 embodied souls.
But I don’t think any of this is relevant for the same reason I’ve always dismissed speculation about singularity-style technological advancement scenarios that focus overly narrowly on AI/whole-brain simulation: We’ll get halfway there first, and that will be wacky enough.
I wish I’d been using tags so I could call up all the examples of brain-computer interfaces I’ve posted about and try to demonstrate that there’s progress being made in the field (as opposed to directionless, largely non-sequential localized advances). Instead I’ll just finally getting around to posting about the new X PRIZE: $10 million for some sort of big progress on brain-computer interface systems (they’ve neither fully defined the criteria for success nor finished raising the money, but it’s the thought that counts).
The last X PRIZE went to Virgin Galactic for progress in commercial space flight, and that’s not exactly commonplace. I realize that just because a lot of nerds are trying to buy incremental progress on this front isn’t going to usher in a new era all that soon.
But that line of work and the sort of things the Bluebrain folks are working one will converge before either reaches any sort revolutionary potential on its own. We’ll have ambiguous hybrid cases before we have to ask about whether what’s in the box is a soul created ex-nihilio or not.
