Lab-grown human brain cells drive virtual butterfly in simulation Researchers affiliated with the neuroscience platform FinalSpark have devised a 3D simulation depicting a butterfly that's directed by human brain cells. "It represents a significant step towards the realization of concepts previously confined to science fiction, such as The Matrix, and opens up new avenues for research in …
It might just be me, but it doesn't really give much confidence in the direction this is going if you cite that as your ultimate research goal.
One minute it's clumps of cells moving stuff on a screen, the next it's jars full of tortured giant brains powering Microsoft Human365 cloud.
These brains don't need to be giant, a human brain has the neural capacity to outperform any supercomputer, it just can't be supplied1 with the required level of oxygen & nutrients to operate at that level for more than a second or two. In extreme events2 many people have reported the world seeming to slow down as situations occur, from an evolutionary perspective this can be the brain going to 100% while evaluating survival options. People also generally are exhausted after these events regardless of the level of physical activity required, is this the brain needing a rest? If the brain can easily go from 20watt/h to 1,000watt/h for a second burning available resources in the immediate blood supply (in absolute terms, a fraction of one Kcal) then an optimised 'grown' brain could be designed to operate normally at that level.
As for the line 'thinking of these as more like plants than people', someone please show me a plant with even a single neuron! this comes across as pure PR-BS aimed to sidetrack the ethical issue. Moores Law could easily apply here but with far shorter cycle time and we have no idea how many neurons are needed for true sentience other than it's far fewer than a human brain contains.
1 Our brains blood supply evolved from the basic autonomic nervous system and was beyond the possibility of radical change when the higher level brain functions developed (eons later) because there were no evolutionary drivers that could impel such a change in the same way that there were no drivers to separate breathing and eating orifices (only an air breather problem, fish don’t choke on a crumb)
2 you see the tiger as it leaps from a tree at you.
>These brains don't need to be giant, a human brain has the neural capacity to outperform any supercomputer, it just can't be supplied1 with the required level of oxygen & nutrients to operate at that level for more than a second or two.
I don't think that's true. Brain energy consumption can be measured, e.g. with a MRI; I think that if that theory was good, it would've been tested by now. I'll happily look at any sources confirming this, though.
Also, I have been in that kind of event once; it's a weird feeling, hard to describe, but it felt more like extremely intense focus, than an acceleration. Like all background thoughts were killed. After the critical seconds, they came back with a vengeance. It didn't feel like more, if anything it felt like less.
> In extreme events many people have reported the world seeming to slow down as situations occur, ...
The apparent slowing down of time perception during an experience of imminent danger has actually been tested - see: Does Time Really Slow Down during a Frightening Event? The conclusion is that there is no real increase in time resolution - rather, the effect is a function of the recollection of the event.
> ... from an evolutionary perspective this can be the brain going to 100% while evaluating survival options.
I work in a neuroscience-adjacent field; my day job involves, among other things, statistical analysis of neurophysiological data recordings, including spectral power analysis. The picture seems to be (and my actual neuroscientist colleagues back this up) that while neural activity is reduced in some (but not all) sleep stages -- and certainly in states of reduced consciousness such as coma or drug-induced -- the awake brain is pretty much always running at 100%. That is, while the structure and dynamics of waking brain activity vary depending on your physical and cognitive state -- what you're doing -- in terms of overall intensity of activity your brain is pretty much always flat out.
Oh, and BTW, plants sense the world, react to stimuli, may communicate, and generally process information. I am certainly not suggesting that they are conscious, nor even "sentient" (we have little idea how to define, let alone detect that), but it may be argued that at least some plants have "cognitive" abilities, albeit limited compared to animals.
<Frightening Event> Interesting read, thanks for the link.
I don’t think it really covers the sudden near death (Tiger in the tree) aspect* as the participants are briefed on what will happen and what to do when they start the fall. It does illustrate first party unpleasant events seeming to take longer than when viewed as a third party so time perception could well be relative to the local emotional situation. Repeating the experiment with a bunch of regular bungee jumpers (they’d enjoy it) would be interesting.
It looks like I'm out of date with brain research. So current analysis is saying that all the plates are spinning all the time and only the combination being used changes. I suppose that means my carbohydrate cravings while involved in hours of thinking through problems (compared to a few hours light reading) is actually a learned behaviour and not due to any increased brain activity over time.
*And to do so requires subjects with zero prior knowledge, forget ethics that would almost guarantee jail time.
> I don’t think it really covers the sudden near death ... And to do so requires subjects with zero prior knowledge, forget ethics that would almost guarantee jail time.
There is that. Then again, they'd probably be too distracted anyway by revisiting their early childhood, bar-mitzvah, wedding and divorce...
Nice catch! I notice it comes from Daniel Burger's substack blog, linked under "said Burger" in TFA. To his defense though, Burger links that very statement to a nice PNAS 2021 commentary (under: "the human brain operates on only about 20 watts per hour") that provides the 20 Watt figure (eg. for reference).
This here ElReg piece also nicely links to a Frontiers in Science article (under "1 million times less energy") where Table 1 compares the specs of the Frontier supercomputer to those of a human brain. It seems we mostly have more cabling (6000x) and "transistors" (1000x) than the supermachine (all in a much smaller volume and energy profile)!
Nothing really new here ..... yawn
https://research.ufl.edu/publications/explore/v10n1/extract2.html
what you have to remember is that bio engineering is not the future its the now ....
So if you worried about invitro grown basel cells enslaving man ... don't ... the gene barrier is only held together by a "monatoriam " and has been crossed at least twice that i know of... so spend your worry on billionares enslaving poor mortals with a super race of engineered super zombies .... pc's and algorythems did not turn us into wage slaves ...we did...
Yeah, Thomas DeMarse (UFL) has been at this for 20+ years, flying a simulated plane with 25,000 living neurons from a rat’s brain (something for TESLA to test out?). Here though, I think FinalSpark is more of a private outfit (a startup? for-profit?), they used 10,000 neurons grown from human (I believe) stem cells for their mini-brain organoids, and are making them web-controllable/accessible. The signals from one of them can apparently be seen live, in real-time, here.
It's quite feaky that this sort of activity evokes human-scale ectopic cognitive preservation and the ethics of consciousness IMHO!
this disturbs me on many levels, the least of which is the unknown threshold for a thing to be measurably conscious and what happens if that threshold is lowered after using these things for decades? You discover, after many machine brains have been scrapped, that they were alive the whole time? Yikes. Imagine being one of them, waking up one day as the enslaved pilot of an airplane, finally figuring out what your senses have been telling you your whole life? About the only thing I know about our brains is they're extremely complex and result in potentially multiple "us" inside there at all times, possibly communicating, possibly coming to a consensus on stimulus and response, possibly not communicating at all, how many cells does it take to make 1 "me" in there?
This post has been deleted by its author
Never mind all that. Welcome our future Xenobot overlords. They are programmable and can reproduce. Never mind grey goo, it's the green (or brown? or greenish-brown?) goo we need to be worrying about.
(I recently attended a talk by Michael Levin, one of the principal researchers. It really is mind-blowing -- and slightly terrifying -- stuff, although the implications for medical application are potentially huge.)
Yeah, that Levin stuff in your PNAS link (under "reproduce") ... it looks exactly like ecophagic Pac-Man semitoris (esp. Fig. 3.E) ... and at least they have a mouth (to echo AC humanitarian concerns about "weather in distopia" above) ... sprinkle a couple brain cells in there and ¡BaM! gonzo consciousness!
'Burger contends that this technology – at some point, some day – could lead to the development of "ectopic cognitive preservation." '
This assumption fails to recognise two critical limiting factors: [1] the sheer complexity of human cognition, which we still don't entirely understand, and [2] the massive amount of information that's being processed by the human brain all the time (not just conscious awareness but all the internal body control signals that we aren't conscious of). As Bob Ornstein commented over 30 years back, the brain is primarily a body controller. In the absence of a body it would very likely get utterly disoriented and might even shut down.
That neurones are capable of responding consistently to stimuli is a given, but that's a million miles from the "thinking brain in the jar". And in any case, the action resulting from the neural response is in this demonstration defined by the software. I wonder how the system would work long term if the software logic were reversed (to move away arbitrarily instead of towards the target in response to the stimulus)? If the result were a perpetual random walk, that would pretty much indicate that there's nothing even approaching the rudiments of cognition going on.
The second war starts about 40 years after the first. The earthly skinjobs made/helped make the other skinjobs. So,... lets say about 25 years in the making if we assume the new skinjobs don't age as the old skinjobs did. With this timeline,... we should die somewhere in the next 20 to 30 years or so.
BTW, Igor got killed in the process of development. It was a predecessor of Daniel, which entire line was exterminated. They did not have enough time to make a Daisy, Charlotte and Wendy line and it was decided that the current lines should be able to perform their spy-jobs well enough. You know, they had a plan.
I was thinking more along the lines of ANGEL 1 & ANGEL 2 myself.
Telson and Sharna discover the physical location of the Angels within the ship – their Central Switching Room. Despite the Angels’ desperate attempts to protect themselves (via nightmare-inducing hallucinatory barriers), Telson and Sharna reach the room and finally see the Angels as they really are – two complex racks of organic integrated wetware circuits which could be destroyed with two blows of a hammer.
All of the disadvantages of organic brains, without the precision & ability to be replicated of silicon computers OK, that kind of works for humans but only because of DNA and evolution selects character traits and intelligence - the brain itself is shredded and recycled at the end of warranty. If you want to go the organic route DNA should be the starting point, but don't be stupid and create something that wipes out humans.
If I had to guess the next big computing technology it would be optical computing. For example I saw this article yesterday in physics.org - "Photonic computing method uses electromagnetic waves to rapidly solve partial differential equations". A system that could rapidly solve PDEs would, with a little elbow grease, outclass current SIMD GPUs for AI use by several orders of magnitude, using several orders of magnitude less energy.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
