Speaker 0:  today, we're looking at a massive update to how we define artificial intelligence. so forget chat to you for a second. scientists have successfully mapped the entire brain of a fruit fly and are actively turning it into software. if you ever wondered if the matrix could actually happen, maybe you should hang around and listen to this. yeah. this is the first time we've completely, i have a complete wiring diagram, connect to what do you call that? a connectome?

Speaker 1:  connectome. connectome? connectome. or connectome. yeah. yeah. yeah.

Speaker 0:  of a complex adult animal brain, the fruit fly. a startup called mamazing is using a map to build ai, bottom up, by mimicking biology rather than top down approach used by openai or google. it's headline news because it proves that we can digital digitize biological brain, paving the way for hyper efficient ai and eventually digital consciousness. what do you think about this, john?

Speaker 1:  yeah. so we did cover the fruit fly mapping, i think, but this is the next step.

Speaker 0:  we did cover the fruit fly mapping before, and we didn't know what they're like, how are they gonna take it now. they're creating an ai based on this mapping. so this is definitely the next step in all this.

Speaker 1:  yeah. this is great. yeah. so building on this fruit connect home, so the new venture mem memezing, i think.

Speaker 0:  amazing.

Speaker 1:  amazing. okay. aims to go beyond static maps and create software emulations of brains. so in technical terms, this involves converting a connectome into a computational model where neurons and synapses are represented as mathematical nodes and weighted connections. so these models can then be simulated and observed to see whether biological patterns give rise to intelligent behavior. so for example, researchers can simulate visual stimuli and evaluate whether the software fly model responds in ways comparable to a real fly, which is super cool and it's different from traditional ai approaches, which typically use statistical optimization on large datasets. and this is because this uses direct biological structure as the basis for computation. so, technically, emulating a brain connectome involves neuroscience, machine language, and high performance computing all at once. so the brain's wiring dictates how learning memory and perception emerge, but the exact mapping from structure to computation, that's what's not fully understood. so what amazing attempts is to operate within that structure using simulated environments to explore how synaptic strengths and networks dynamically translate into behavior. so what's cool about this is the project also intersects with effort in ai energy efficiency. since the biological brain operates with dramatically lower energy budgets than current ai data centers, a bio derived computing paradigm could lead to more efficient brain like artificial systems. and just on something else i think we did cover, a bio chip that, i'd be based in australia before. so, like, there's a lot of parallels here to all of this stuff. what do you think about this? is this i think this is kind of interesting, but i think we're far off from simulating the actual human brain.

Speaker 0:  yeah. definitely far off from that, and we'll have to see what kind of power we can get at this. but i this is really cool. so think of, to take what john said scientifically and put it into normal person's terms, think of a biological brain like a house. and until now, we've just been looking at the outside of the house. connectome is a full set of architectural blueprints, every wire, every switch, and every outlet inside the walls. so the brain emulation is taking those blueprints and building a digital replica of the house inside the computer. so, currently, ais like llms require massive power plants to run, but biological brains run on energy equivalent of a crumb of food. so copying biology could save, the ai energy crisis or could just change, you know, where energy comes from in the future. this is, definitely incredible.

Speaker 1:  mhmm.

Speaker 0:  you know, maybe you're, like, giving your laptop part of your sandwich in the future at lunchtime. it moves us from an ai that predicts text to an ai that perceives and navigates the world like a living creature. it's very interesting. so this, you know, could have impacts on our life, like smarter robotics, vacuum cleaners, drones that could navigate chaotic swarms as easily as a fly that tips around the kitchen without getting stuck. yeah. but imagine self driving that could handle this, and making those in moving, instinctive and automatic. ai chips in your phone that last for days, not hours because they mimic efficient biological circuits or even medical breakthroughs, testing drugs on digital brains instead of animals speeding up cures or for neurological diseases.

Speaker 1:  yeah. so the the and talking about ambitions and challenges here. well, the fruit fly, obviously, we talked about is magnitude, magnitude simpler than human brains. the thing is that connectome provides a manageable yet nontrivial test bed. so the early work focuses on simulating this nervous system and linking the model to simple robotic interfaces to test the behavior. so their long term ambition, obviously, is human brain emulation. so the the the challenges of that is scaling data collection, capturing neurochemical and temporal dynamics, understanding how abstract cognitive functions emerge from circuitry. and so there's a lot more there. but this effort, it does represent the most direct attempts yet to reverse engineer biological intelligence using software, which is really, really cool.

Speaker 0:  yeah. and this, raises some ethical and philosophical concerns. you know, brain emulation raised profound questions around, you know, as connectome startups progress towards sophisticated emulation, society must consider issues of digital consciousness, morale status, identity consent, and privacy, and what that means. but according to some of these timelines, even though i think with quantum computers, these are gonna be rapidly accelerated. right now with the fly brain as map, they say it's gonna take decades to get small mammals like mouse and such map and a hundred plus years to get a full human brain emulation.

Speaker 1:  yeah. wow. there's like, i love that they're moving forward with this. i'm curious to see how this would relate to the brain on a chip that we had discussed previously. so, yeah, this is really cool. i'm really excited for this.

Speaker 0:  well, i'm hoping that eventually, you know, all these companies, like, will borrow from each other. so the next version of, you know, our brain on the chip will use some of these things that amazing is publishing and talking about and vice versa to make it all better. so we'll have to wait and see. so keep an eye on amazing. they are the ones putting out these digital brains into robot bodies, so we'll see that. don't expect the mind uploading to be a consumer product in our lifetime. but this is a multigenerational scientific project that's interesting. so, you know, there's a shift from artificial intelligence to artificial life, building systems because they are wired to think, not just trained to predict.

Speaker 1:  yeah.

Speaker 0:  so the big question, if the tech existed tomorrow, would you upload your brain to the cloud or keep it analog? let us know what your comments are, and we'll keep watching this story closely.