Why Kurzweil’s nanobot idea won’t work

Ray Kurzweil,[1] famous futurist, hired by Larry Page in 2012 „to bring natural language understanding to Google“,[2] predicts that humans will use nano-robots to connect their brains to the cloud. I think he overlooks a fundamental, show-stopping problem.

The core of the idea goes back to his 2006 book „The Singularity is Near“, where he predicted that „during the 2020s [it will become] possible to scan the brain from the inside using nanobots“.[3][4] In his 2012 book „How To Create a Mind“, he developed it into the full idea,[5] adding the cloud and „write“ capabilities to the originally „read only“ bots. In 2015, he said:

„In the 2030s, we are going to send nano-robots into the brain (via capillaries) that will provide full-immersion virtual reality from within the nervous system and will connect our neocortex to the cloud.[6]

He repeated the prediction only recently at the 7 Days of Genius festival in NYC.[7]

It’s a fascinating idea. And it’s true that brain-computer interfaces are being developed full steam ahead these days. But I think Kurzweil’s idea won’t work—even if it was successfully implemented.

Of course, the idea has a number of superficial problems, which fall into three categories:

1) Time schedule: „In the 2030s“ seems to be far from realistic, even for the original „read only“ idea. But that’s just a number. I’ll generously change the claim to „at some point in the future“.

2) Technical issues: The nanobots need some kind of long-term power supply. They also need a wireless connection to the cloud—each one of them! Which will account for considerable power consumption. And they will have a lifespan, which will come to an end, so they must be removable. And they will need to be injected into the brain since they will hardly be small enough to pass the blood-brain barrier. Which will make their removal a really demanding task, to say the least. And so on. But these are just technical problems which might be solvable—at some time in the future.

3) Ethical and legal issues: It will be very hard to impossible to get approval for testing the insertion of nanobots into the brains of living humans. But people find ways around legal issues, so I’ll assume that someone will do it.

So, my fundamental concern is not about the cloud or the bots. It’s about the brain being connected to the bots.

And it’s not about the „read“ part of that connection. If the technical problems are solved and someone’s actually brave enough to have a shot of bots injected into their brain, those could be used as single neuron brain scanners without problems.

My concern is about the „write“ part.

Adding millions of „read/write“ bots across the neocortex is the same as growing millions of new neurons and connecting them to the existing ones: A massive rewiring of the whole neocortex, which then has to rewire itself to make sense of the new signals coming in.

This has happened to all of us while we grew up. And it’s one of the reasons why we all have childhood amnesia—why we cannot recall specific events (who, what, when, where) before the age of 2–4 years.

Recent research has shown that this degrading of existing memories is linked to high neurogenesis, which leads to the replacement of synapses in existing memory circuits, effectively deleting the memories encoded by these circuits.[8][9]

So I think our brain’s adaption to the nanobots would result in a fundamental personality change. We would grow up a second time and mostly forget who we were before.

The two-year-old boy I once was doesn’t exist anymore. I have almost no idea how it really was to be him. After the nanobot integration, I as I know myself would no longer exist. I would hardly know who I was before.

For me, that doesn’t count as „working“.


PS: This post emerged from a Twitter thread started by Sara Nicolau.[10] Thanks for the inspiration!

  1. https://en.wikipedia.org/wiki/Ray_Kurzweil []
  2. http://www.wsj.com/articles/SB10001424127887324504704578412581386515510 []
  3. https://en.wikipedia.org/wiki/The_Singularity_Is_Near#The_brain []
  4. Kurzweil: „The Singularity is Near“ 2006, p. 163-167. []
  5. http://www.wsj.com/articles/ray-kurzweil-looks-into-the-future-1401490952?KEYWORDS=Ray+Kurzweil []
  6. http://singularityhub.com/2015/10/12/ray-kurzweils-wildest-prediction-nanobots-will-plug-our-brains-into-the-web-by-the-2030s/ []
  7. http://www.inc.com/graham-winfrey/ray-kurzweil-neil-degrasse-tyson-7-days-of-genius.html []
  8. Josselyn, Sheena A.; Frankland, Paul W. (2012-09-01). „Infantile amnesia: a neurogenic hypothesis“. Learning & Memory (Cold Spring Harbor, N.Y.) 19 (9): 423–433. doi:10.1101/lm.021311.110. ISSN 1549-5485. PMID 22904373. []
  9. Akers, Katherine G.; Martinez-Canabal, Alonso; Restivo, Leonardo; Yiu, Adelaide P.; De Cristofaro, Antonietta; Hsiang, Hwa-Lin Liz; Wheeler, Anne L.; Guskjolen, Axel; Niibori, Yosuke (2014-05-09). „Hippocampal neurogenesis regulates forgetting during adulthood and infancy“. Science 344 (6184): 598–602. doi:10.1126/science.1248903. ISSN 1095-9203. PMID 24812394. []
  10. https://twitter.com/SaraSnicolau/status/707831174677319680 []