Wow. Your channel keeps exposing me to these incredible ideas. Your Dr Michael Levin interview led me down about 20 hours of mind blowing lectures… this interview is equally as amazing. In fact if you could arrange for the two of them to chat to each other I’ll grab the popcorn!
Excellent interview. A well explained and fascinating brief look into a complex protein. Best of luck to your guest with his research and his contribution to science. 👍
What's more, ML is based on the idea that neurons that fire together connect together. But apparently this would be a second affinity, after connecting them together we decide how much they should be connected. A discreet step followed by a continuous affinity process. Cool insight.
I may be way off-base. . . . but here goes. I've been following Mike Levins work and it occurred to me that he's looking for a mechanism whereby as cells multiply, they need to know what the next cell is going to be, and when they should stop dividing. Thus if cells are to 'make a foot' they must pass on information on when to stop at the boundary. Something mechanism is needed to pass that information on. His group has identified ion channels as part of that path. I do wonder if Talin may be involved in the cells passing that information on? I have some knowledge of AI and it seems unlikely that a Talin molecule would be used to hold weights.
This is indeed quite similar to a neural network that have affinity between neurons, but then they were based in how they supposed the brain worked. So it's one value per synapse? That's 1 trillion? GPT 3 has 175 billion parameters, so not that far out.
Hi Abdelilah, good question, treatments that reduce ECM stiffness are being used successfully in cancer therapies (i.e. LOX inhibitors which reduce collagen crosslinking) so perhaps similar strategies might one day be possible for other types of ECM.
@@bengoult1 thx for your kind answer. Could certain natural interventions like enzymes or monitoring of A1C and insulin through relatively safe geroprotective molecules like metformin have also some benefit in delaying stiffness of ECM?
we say ECM a lot of times in this video - if we didn't define it, we means "ExtraCellular Matrix" !
Thanks Eleanor! This was a really fun discussion. Keep up the excellent work!
Thanks Ben and thank you again for coming on the show!
Wow. Your channel keeps exposing me to these incredible ideas. Your Dr Michael Levin interview led me down about 20 hours of mind blowing lectures… this interview is equally as amazing. In fact if you could arrange for the two of them to chat to each other I’ll grab the popcorn!
Excellent interview. A well explained and fascinating brief look into a complex protein. Best of luck to your guest with his research and his contribution to science. 👍
What's more, ML is based on the idea that neurons that fire together connect together. But apparently this would be a second affinity, after connecting them together we decide how much they should be connected. A discreet step followed by a continuous affinity process. Cool insight.
Fascinating, thanks for producing. Would it be helpful for a volunteer to proofread the transcripts of your shows?
Wow, coincidence... This week i studied about ECM in my coursework and you uploaded this 😅
I may be way off-base. . . . but here goes. I've been following Mike Levins work and it occurred to me that he's looking for a mechanism whereby as cells multiply, they need to know what the next cell is going to be, and when they should stop dividing. Thus if cells are to 'make a foot' they must pass on information on when to stop at the boundary. Something mechanism is needed to pass that information on. His group has identified ion channels as part of that path. I do wonder if Talin may be involved in the cells passing that information on? I have some knowledge of AI and it seems unlikely that a Talin molecule would be used to hold weights.
I have been 🤔 thinking about this alot!?!!!
40:00 mm Perhaps these work like digital circuits in telecommunications where the information is compressed encoded and uncompressed and decoded.
This is indeed quite similar to a neural network that have affinity between neurons, but then they were based in how they supposed the brain worked. So it's one value per synapse? That's 1 trillion? GPT 3 has 175 billion parameters, so not that far out.
So yet another reason to worry about glycation of proteins?
How do we extract the data from the brian and upload it to a computer!? That should be the real question we need that!!!😏
Thx for this interesting video. What interventions could delay stiffness of ECM ?
Hi Abdelilah, good question, treatments that reduce ECM stiffness are being used successfully in cancer therapies (i.e. LOX inhibitors which reduce collagen crosslinking) so perhaps similar strategies might one day be possible for other types of ECM.
@@bengoult1 thx for your kind answer. Could certain natural interventions like enzymes or monitoring of A1C and insulin through relatively safe geroprotective molecules like
metformin have also some benefit in delaying stiffness of ECM?
If this is correct, it's huge, like, on the order of Watson and Crick.
＋
Elon Musk is waiting with his Neural Implant !!