The Future of the Brain - Miguel Nicolelis, MD, PhD | The FitMind Podcast

Miguel Nicolelis, MD, PhD is a Brazilian neuroscientist and physician famous for his pioneering work in the field of brain-machine interface (BMI) technology.
Dr. Nicolelis’s extensive research with brain-machine interfaces blurs the lines between real life and science fiction.
In the episode, we discuss neuroprosthetics, monkeys controlling robots, brain-to-brain communication, and the key implications of these wild, futuristic technologies for us all.
For the full list of FitMind Podcast episodes, visit www.fitmind.com/podcast.
ABOUT FITMIND
FitMind provides mental fitness training based on the latest neuroscience. We created a mobile app that teaches 50+ science-based methods adapted from various traditions (i.e., Loving-kindness, Stoicism, Yoga, Buddhism, Sufism, Taoism, Indo-Tibetan Bön, etc.). Check it out here: bit.ly/fitmindtraining
Website: www.fitmind.com
Podcast: ​pod.link/fitmind
Instagram: bit.ly/IGfitmind
SHOW NOTES
0:00 | Intro to Dr. Miguel Nicolelis
2:56 | Monkeys Controlling Robots
9:04 | Mirror Neurons & Brain-Machine Interface (BMI)
11:53 | Advances & Applications of BMIs
16:26 | BMIs for Parkinson’s Disease
20:54 | Upgrading the Human Mind?
22:02 | Why Machines Will Never Think Like a Human
28:30 | Brains are Changed by Digital Systems
31:18 | Limitations of Computers
35:50 | How Nature Shaped the Brain
39:19 | We Only Experience a Sliver of Reality
46:33 | Viruses of the Mind
62:25 | Rapid Fire Questions

[@coder5382]

I've been a self-taught guy since I was 9 years old. I started my journey learning English on my own. I'm Brazilian, and learning English in public school wasn't that good, so I decided to take it on my terms. I'm now a coder, besides my studies in electronics (both analog and digital electronics). I've also developed a love for learning languages. I now speak three languages: Portuguese, English, and Japanese. That has also led me to have a big fascination with the brain and its inner workings. Miguel, you've always been a big inspiration to me. Long before Elon started his NeuraLink project, capturing electrical activities in the brain, you were way ahead in your studies and experiments doing that. The brain is really a fascinating organ, or machine, I should say. Our brains will go backward if we start to depend too much on technology for everything and don't challenge our brains too much. That could create a huge problem for the evolution of the brain. Instead of learning a second or third language, for example, one uses machine translation. I'm a strong believer that a brain will only evolve if it is challenged. "Use it or lose it." Thank you for uploading, FitMind.

[@AnaCristina-mp5fe]

Nicolelis’ amazing!! 💛

[@JOKENCANAL]

Simplesmente fantástico❤

[@AnitaMindFit]

Ameiiiiii. The Future is ours...

[@ecstazyrm]

Excellent Dr Nicolelis! You are the boss 😊

[@axeldietrich1244]

Fascinating man

[@bluerose2682]

Great interview!!!!

[@nerdacademico5042]

Perfetto ❤

[@landofreikiaroma]

Wonderful interview 🎉

[@AnitaMindFit]

Darwin is everything

[@AnitaMindFit]

Monkeys and natural behaviours

[@taupring]

Actually, while I do like most of the interview and insight, a number of responses around the future Machine Intelligence (IMHO) are just wrong. First, I would agree that the simplistic deep neural networks that we have now will not achieve true AGI. Just starting to create architectures that mimic brain responses is leading us to some remarkable results (GPT3, DALLE, etc), but the research that will get us much, much closer is simulating (with extreme precision) the true complex nature of the brain (including the analog responses and multiple morphologies of neurons). Creating multiple neural networks with this level of realism, connecting them, and then refining with results from actual brain patch clamps will provide a level of simulation and response that will simply be uncanny and almost magical. To say that this won't happen in 15 to 20 years time is naive (very naive in the case of a researcher that can understand the level of progress that is being made).

[@opaleyeakintunde4827]

hhhhmmmm

[@taupring]

What is missing from the leaked Q* algorithm (and ChatGPT/transformer based models) is agency. Without agency you won't have true AGI. As I mentioned, yes these are getting very sophisticated and will be able to do almost magical things, but sentient beings (and true AGI) don't just respond to questions when asked. They initiate new thought without being prodded/asked. What I believe will be needed to get agency into these systems is (at a low level) a neural circuit that contains a very tiny amount of randomness/jitter (very tiny). This slight amount of jitter provokes random thought creation internal to the NONN (Network Of Neural Networks) that make up the AGI. Most of these internal thoughts get rejected after being passed by an internal filter, but a few make it through (and form the basis of agency for the AGI). Again, it would be much more sophisticated than described, but this is an example of mechanisms that would need to also be in place before we can have true AGI. Of which, transformer based models are just a part of what will be needed and by themselves will not provide true AGI.@@opaleyeakintunde4827