Hi David,
It's also not a good idea to jump all over people who are just trying to get their heads round something and get it wrong (this I know because I do that too often as well). If you're trying to help someone understand something, almost always the wrong way to begin is by attacking their current understanding. It just makes them defensive. This does not apply to people who claim to know plenty about a subject. So, for example, when I claim to KNOW that prediction MUST precede recognition, I'm actually demanding attacks on that idea. I want to be certain in a scientific sense, and the best way is to invite smart people to try and prove me wrong. The "worst case" scenario is I am proven wrong in my original idea, but we now know why. That will always lead to something better. Now, when asked again and again about these fears coming from highly regarded authorities in other fields, Jeff understandably treats the question in a certain way. Instead of discussing "how" we can do this (Jeff's lifelong quest), this is a completely different question about what'll happen if we do it and we mess up the deployment. It is an important question, but it's not what Jeff is interested in. It's hard enough to get anywhere near there, and then it's even harder to make a smart computer dangerous. It's much easier to give the dangerous powers to bad people than it is to build a smart computer and make it bad and then give it the power to be dangerous. So it's in a certain sense a really stupid question. People who are scared of the possibility of dangerous AI should spend their time figuring out how to keep power away from bad intelligences, regardless of their substrate. Clearly, as history shows, we're much worse at this than we are at developing potentially destructive technologies. AI's are hardly likely to make that less true. (I wrote this before the video was linked to, so I'll send this and then watch that - on my phone which can't do both). Cheers Fergal -- Fergal Byrne, Brenter IT Author, Real Machine Intelligence with Clortex and NuPIC https://leanpub.com/realsmartmachines Speaking on Clortex and HTM/CLA at euroClojure Krakow, June 2014: http://euroclojure.com/2014/ and at LambdaJam Chicago, July 2014: http://www.lambdajam.com http://inbits.com - Better Living through Thoughtful Technology http://ie.linkedin.com/in/fergbyrne/ - https://github.com/fergalbyrne e:[email protected] t:+353 83 4214179 Join the quest for Machine Intelligence at http://numenta.org Formerly of Adnet [email protected] http://www.adnet.ie On Mon, May 25, 2015 at 5:04 PM, cogmission (David Ray) <[email protected]> wrote: > This was someone's response to Jeff's interview (see here: > https://www.facebook.com/fareedzakaria/posts/10152703985901330) > Please read and comment if you feel the need... > Cheers, > David > -- > *With kind regards,* > David Ray > Java Solutions Architect > *Cortical.io <http://cortical.io/>* > Sponsor of: HTM.java <https://github.com/numenta/htm.java> > [email protected] > http://cortical.io
