What The F**k
213 segments
You know, I always thought it was going
to be the LLMs that make the end of the
world, but um I guess I was
significantly wrong. I just got done
reading this article and it's making me
feel all sorts of ways. Feelings I'm
just not used to feeling. Okay. Human
brain cells on a chip learn to play Doom
in a week. Now, you're probably thinking
this surely has to be some sort of
clickbait article. It's not like actual
human brain cells somehow wired up and
firing and actually playing a video
game, right? No, it is actual human
brain power being used to play Doom.
Okay, I have lots of things to say about
this. The first one obvious is that they
created a chip powered by human brain
cells. And what did they do? They made
the brain learn and play Doom over and
over again. What kind of weird purgatory
existence is this? You made something
that resembles some weird outline of
life and then forced it to play Doom for
all eternity.
You sickos. I know I'm like half joking
and all, but also it's like kind of
messed up, right? Like this is this is
just kind of weird. The version of Doom
they're playing doesn't look like real
Doom. This is some sort of mod andor
just like Teimu version of it. The fact
that you made a hellish experience for a
brain ship and then made them play Te-Mu
Doom, that is just not okay with me.
Okay, I'm not happy about this. But
first off, I figured we'd do a little
bit of background. So, in 2021, an
Australian company, Cortical Labs, used
its neuronpowered computer chips to play
Pong. And it looks like this. I honestly
I [laughter]
maybe I don't have enough brain cells to
understand what that even looks like,
what that's supposed to mean. If you
look at the game image, it's actually
doomed to play by itself. It's just
playing against a wall. Now, the chip
consisted of clumps of more than 800,000
brain cells on these micro electrode
arrays that can send and receive
electrical signals. And somehow that was
able to learn how to play Pong. Now, the
one that actually learned how to play
Doom used about a quarter as many
neurons as the Pong demonstration and
played Doom better than a randomly
firing player. Far below the performance
of the best human player. good, because
that means the best human player could
be matched by 200,000 brain cells.
However, it learned much faster than
traditional silicon-based machine
learning systems and should be able to
improve its performance with newer
learning algorithms. Who the hell forces
brain cells to learn how to play Doom?
Their entire reality is going around
hunting monsters with weapons and
picking up ammo and trying to find the
yellow, red, and blue key cards or
skulls where ultimately when you reach
the end of the first act, you have to
die. They actually created some sort of
Dante's Inferno brain torture machine
where you're just forced to be the Doom
guy and hunt demons. It's just actual
hellish experience. Think about it. That
poor brain just had to just process
nothing but violence all day. its only
existence was in violence and BFGs and
plasma rifles. But if you didn't have
that thought, your probably next thought
is why in the world would anybody
actually ever create this? What is the
purpose of having a brainbased computer
chip? Well, there's kind of like two of
them and this is where we get into like
the funny territory. So apparently the
realworld application that they're going
for is they want these controlling like
bio wetwware chips to control robotic
arms. They unironically want to create
the terminator.
>> The CPU is a neuron net processor a
learning computer
>> because apparently these brain cells
they just learn better than traditional
ML styled learning algorithms. The part
that's like the most terrifying of it
all. Besides for the fact that they
literally want to take brain chips and
shove them into robots and allow them to
walk around the world is that these
chips are actually controlled with
Python. Yes, that's right. The Cortical
Labs has developed an interface that
makes it easier to program these chips
using the popular programming language
Python. An independent developer Shawn
Cole then used Python to teach the chips
to play Doom, which he did in around a
week. You know what this means, dude? We
are going to be vibe coding robots. I
mean, we are we ourselves [laughter]
don't know.
Oh, we're going to be taking Python and
running just absolute critical
infrastructure software. Oh my.
[laughter]
Oh no. That means we're actually
entering into a world where pretty much
effectively everything that's on the
internet is potentially just made by a
robot and fake and then the real life is
just going to be filled with brain
chips. Like that's that's what's
happening. Somehow we're using brains to
make the robots walk around. It feels
deeply unsettling, you know, and I don't
know. And honestly, like just on on an
aside, this is really just kind of
feeling, you know. I I I must say that
this does not make me feel very good
because there's a lot more questions of
life as you add more and more brain
cells. What are we doing?
I don't I don't know if this is very
good. This just feels like a bad idea.
But I think my favorite part of the
entire thing isn't just the fact that
they're going to go freerange this on
some robot arms while vibe coding some
Python. It's that they don't even really
know how it works. Steve Ferour at the
University of Manchester, UK, agrees
that Doom is a significant level up from
playing Pong. But he says there's still
a lot we don't understand about how
these neurons are playing the game, such
as how the neurons know what is expected
of them, or how they can see the screen
with no eyes.
Oh my gosh. [laughter]
I mean, that in in of itself is just
like just that's an insane statement. We
have technology where we're just like,
well, actually, we don't know how they
learn or how they function or how they
see. We just know that when you give it
the data, it just works. And this
actually reminds me of one of my
favorite books of all time when I was
back in my master's day on intelligence,
which effectively one of the experiments
is that they took a ferret and they took
its eyeball eyeball cords and put it
where its ear holes should be and it
took its ear holes and put it where its
eyeballs should be. And guess what?
Couldn't even tell the difference. The
little ferret brain was able to take all
of that input and just be able to go,
"Oh yeah, I totally I totally get this.
Actually, nothing. That doesn't change
me at all." And now we're taking a maybe
what is equivalent to little ferret
brains and making them play Doom. Just a
true hellish experience. Your only
waking memories, your only knowledge of
life is seen through the old pixelated
Doom guy's perspective of just trying to
avenge his rabbit. Yeah, that's so
that's all I really wanted to yap about.
Um, dude, this has to be by far one of
the most interesting and also very
deeply uncomfortable stories I have ever
read. The name is Please don't vibe code
a machine and then put robot baby brain
intelligence on it and put it on a micro
electrode or array, whatever that means,
and then send it after me and my family.
Okay? Please, I'm begging you. Don't do
it. Okay? I Hey, I'm sorry for making
fun of vibe coding. I'm sorry for not
beating Doom enough. Okay? Hey, please,
Mr. Robot. Don't do it
again. Hey, do you want to learn how to
code? Do you want to become a better
back-end engineer? Well, you got to
check out boot.dev. Now, I personally
have made a couple courses from them. I
have live walkthroughs free available on
YouTube of the whole course, everything
on boot.dev, you can go through for
free. But if you want the gamified
experience, the tracking of your
learning and all that, then you got to
pay up the money. But hey, go check them
out. It's awesome. Many content creators
you know and you like make courses
there. boot.dev/prime dev/prime for 25%
off.
Ask follow-up questions or revisit key timestamps.
The video explores a scientific development where human brain cells grown on a chip learned to play the video game Doom in just one week. The narrator discusses how these 'bio-wetware' chips, developed by Cortical Labs, use hundreds of thousands of neurons to process information and learn significantly faster than traditional silicon-based machine learning. The technology involves interfacing the cells with Python and has potential applications in robotics, though it raises unsettling ethical questions about the nature of life and the mysterious way neurons can 'see' and interpret data without traditional sensory organs.
Videos recently processed by our community