you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 13 Sep 2026
22 points (100.0% liked)
TechTakes
2698 readers
133 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 3 years ago
MODERATORS
I haven't seen any mention of this in sneer space, so kind of a request for context/comments. I recently learned about Cortical Labs from a news story that came out about a month ago. Basically (and apologies if I've got this wrong), the story is that they've built a "biological data centre" where the computing units use lab-grown human brain neurones. They've been in business (actual business) since 2019. Some shallow searching on reddit and lesswrong shows a flitter of interest from those guys here and there.
I've been keeping extensive notes on these bozos but I don't quite have a story yet. Not even the US VCs are going big into "biocomputing". I'm trying to work out what the fuck Cortical Labs think they're on about.
If I had the time and interest I’d look at their literature but unfortunately I lack both. Filing this under: probably a grift
my first guess is some sort of weird sci fi eugenics
I think its more just getting money from people who think something sounds cool despite being not actually useful. Long storied history of this sort of thing.
The new Maas biochips
If you wanna see something that's actually potentially useful and interesting, just look up the term "Mortal computation thesis"
Hey, why not both
from what i understand, it's also impossible to copy this thing and every single instance will turn out slightly different. also it will continue to learn on whatever data it is fed so you might end up with something useful, that will degrade, is on a timer and has no backups
hmm. Sounds like an average chapter of Franken Fran
Looks like they have 'neuron culture and interface as a service' going on, for all the good its for.
Its a real thing. You grow some neurons and stick some electrodes into it, and you can measure the activity of neurons or groups of neurons through some electrodes and induce some to fire through other electrodes. Now, if you have some external process that you define that takes the output from the electrodes and does some computation on it and returns signals to the input electrodes, the way that neurons work and change their synaptic weights over time ensures that the network will learn to produce outputs that make the inputs more predictable. Hence teaching a few thousand neurons to play 'doom' - you have a program that takes the screen and turns it into a set of input strengths, take some outputs and turn them into controls, and it feeds back, and you eventually have something that learns to do predictable actions.
Their innovation appears to be having standardized software you can use to define the input and output electrodes with respect to whatever other software you are running connected to the stuff. Who is willing to pay for this is anyone's guess.
Maybe all the AI hype has lowered my standards too much, but that actually sounds pretty neat. No idea if it's actually useful and the biological component gives me a kind of instinctive discomfort that I can neither justify nor ignore, but it does feel like there's an actual interesting development there, at least compared to throwing bigger matrices at the natural language problem and waiting for God to pop out at the end.
Ah yes, a torment nexus.