A neuromorphic supercomputer called DeepSouth will be capable of 228 trillion synaptic operations per second, which is on par with the estimated number of operations in the human brain
A better title would be “Supercomputer that could conceivably simulate entire human brain, based on a rough estimate of what it would take to do that if we had any idea how to do that, will switch on in 2024”.
For real. I’m reading the title all wondering how the fuck they mapped all the neuron connections and… nope, the real innovative part of the story is clickbait
That’s only counting connections. The brain learns by making new connections, through complex location and timing dependent inputs from other neurons. It’s way more complex than the number of connections, and if neuroscientists are still studying the building blocks we don’t have much hope of recreating it.
This also ignores that the brain is not wholly an electrical system. The are all kinds of chemical receptors within the brain that alter all kinds of neurological function. Kid of the reason why drugs are a thing. On small scales we have a pretty good idea how these work, at least for the receptors that we're aware of. On larger scales it's mostly guessing at this point. The brain has a knack of doing more than the sum of all parts on a pretty regular basis.
Not to mention the scale and nature of the “dataset” that our brains were trained on. Millions of years of instinct encoded in DNA, plus a few years gathering data from dozens of senses 24/7 (including chemical receptors, like you said) and in turn manipulating our bodies, interacting with the environment, and observing the results. We’ve been doing all of this since embryo.
We can’t just feed a model raw image and text data and expect it’s intelligence to be comparable to ours. However you quantify intelligence/consciousness whatever, the text/image model’s thought processes will be alien to ours, which makes sense because their “environment” is nothing like ours - just text and image input and output.
A computer doesn’t need wet neurotransmitters. It can simulate node activation of all types pretty easily. It just needs to be trained on the proper models. The problem that we will face is choosing when we will believe it to be sentient versus following a complex series of patterns. At a certain point, we’ll have to remember that we are basically machines running in code as well. There will come a time when we will start to feel a moral obligation to grant AI citizenship.
Four grad students out there hand-entering NXML rows while squinting at AI enhanced SEM images should be able to get all 228T done by… next quarter, right?
This is setting aside that bus capacity is the bottleneck vs. compute power and they have yet to demonstrate bus performance of a full 228T connections/second with implicit timing which, to my knowledge, has never been demonstrated in a system a tiny fraction of this size. Though that’s not to say it’s impossible, but while this machine is incredibly powerful the comparison to human brains is predictably inaccurate…
Not too long ago it would take a room like that to mimic a fraction of the power in my watch. Heck, I’ve got more power on my wrist than it took to get to the moon.
It’s a BangleJS. So, not super powerful, but I can program it myself and it has gps, gyro, Bluetooth, and two weeks of battery (assuming I’m not using that stuff constantly.)
Yup. One of those cards that plays a tune when you open it is more powerful than the Apollo computer. Apollo was only working on 4kb of RAM and 74kb of ROM.
If I had been in charge of figuring out how to make that work everyone would definitely be dead.
Those cards are more powerful than the watches I'm thinking of when people say "Casio nothing-watch", I think is what I was getting at. I'm thinking simple digital watches from the 80s/90s/2000s, kind of thing. As far as I know they have no real programmable logic, and anything that might be considered RAM is under 1KB.
But yeah those cards that actually play recorded samples are probably more powerful general computers than onboard Apollo, that's a good comparison.
Still, those computers and stuff like the Saturn instrument unit were freaking marvels, considering what all they could actually do with so little.
I’ve been hearing that quote about watches being more powerful since I was a child in the 80s. And I think it refers to processing speed - or even moreso, calculations per energy. Sure the watches may not have as much RAM/ROM, but they work fast enough to fully update every second, and do so using little enough power to last years on a coin cell.
To be fair our brain took millions of years of evolution, while this simulated brain took only a few years to be developed, maybe in the future this can all fit in a phone perhaps. Enough for this simulated brain to watch memes of beans from this era.
Reminds me back in like the late 90s / early 2000s when we used those rainbow text color fader generator tools for chats and messaging apps. I remember there was a pretty popular one back then but I forgot the actual name.
Step 1: Write your copy. Maybe give it an edit pass for sake of spelling and grammar.
Step 2: Do a websearch for unicode text converter. This cool font was found here.
Step 3: Find a format you like. Copy to clipboard.
Step 4: Paste. This is commonly used for bold, italics, underline and strikeout. There are also modifiers to make text i̽n̽cr̽̽̽e̽̽̽a̽̽̽̽̽s̽̽̽̽̽̽̽i̽̽̽̽̽̽̽n̽̽̽̽̽̽̽g̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽l̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽y̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽ ̽̽̽̽̽̽̽c̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽r̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽a̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽z̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽y̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽̽ by stacking mods. Older systems or websites with incomplete unicode fonts will replace unknown charaxters and mods with boxes, so there is only limited backward compatibility.
Add comment