Of course you knew I would have to comment on this one. First I need to mention that I admire Pete Kassan, the author of the piece. He has been an AI skeptic, though not as hardcore as I, for much longer than most and written some really excellent articles on the topic. Moreover, unlike myself, he comes out of the computer/software world so he has a lot more actual relevant knowledge and experience in the area. Because of this his words carry a lot more weight than mine ever will.

In a very broad sense I am sympathetic to his arguments against the simulation hypothesis. However, the fact that he only discusses the most common, strongest version of the hypothesis (what I call SH1), weakens his overall critique quite a bit. It reveals a lack of let’s call it imagination on his part and suggests his case against the hypothesis is not a well thought through, reasoned, rational one but rather a reactionary one. I cut him a lot of slack in this regard as I can see how in his mind he has intermingled his objections to AI with objections to the simulation hypothesis.

At first blush they appear very similar things and certainly there is much overlap but there are significant differences to be sure as well. Kassan here gets all wrapped up in worries about consciousness and the ability, or not, of a computer to simulate (create) consciousness. This is I believe, as I think as he does, impossible. However, if we are real persons living in a simulation that worry goes away completely and if we or the universe are only partially simulated those worries are also lessened or completely removed depending on the nature of the partial simulation.

There are no reasons I can think of to favor SH1 over any of the alternative simulation hypotheses (other than perhaps that Bostrom’s original reasoning making the case for it no longer makes sense. I do not pretend to fully understand the math, but again I do not see an obvious failure mode as the the logic can be extended to SH3 or any of the partial simulation variants, including SH2 [though the consciousness problem would still be a factor per Kassan], with only minor tweaks. I am working on a post describing how that might be done) therefore if one wants to make a case against it, it is imperative that their reasoning applies to all (at least most) possible variants. Perhaps there is a way to do that but, unless I missed something, in this post Kassan doesn’t even try.

All that said there are other problems with the simulation hypothesis that make it highly problematic or even untenable. The too good to be true problem that I gave discussed at length elsewhere being one of my biggest issues with it. Again, however it needs to be stressed that even that objection is less strong for less strong versions of the hypothesis such as the various partial simulation options that are available such as the ones I call SH2 and SH3 (discussed at length in link below)

Written by

Research scientist (Ph.D. micro/mol biology), Thought middle manager, Everyday junglist, Selecta (Ret.), Boulderer, Cat lover, Fish hater

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store