Neuralink Progress Update [video] | Hacker News Reader

Neuralink Progress Update [video]

youtube.com

281 pointslelf373 comments

A comment from techcrunch:

As the scientist who first published data on neuronal firing in the brain of freely-behaving primates (available at: Ludvig N. et al., 2001, Journal of Neuroscience Methods, vol. 900, pages 179-187) , I am curious to see this demonstration. I also know that in order to make any meaningful link between a human mind and a computer one must record the identified firing of at least 5 billion neurons from the "mind-generating" association cortex in behavioral and environmental contexts -- which is absolutely impossible with each and every currently available and envisioned electrophysiological technique loved or not, claimed or not, advertised or not by Musk.

--- Nandor Ludvig, MD, PhD

Source: https://tcrn.ch/3b3YGGC

Thank you for joining the discussion. What do you think about the counterargument that they don't need to create a full mind-body interface, that they could offer a lot of benefit just by doing simpler things like predicting limb movements?

It is not my comment, I just found it interesting. Sorry for the confusion.

He and his teams seem to have been quite successful with other businesses and ideas that were within the limits of known technology: electric cars and orbital rockets and LEO communications satellites. All of these things existed before (and all credit to the mountains of hard work of SpaceX and Tesla for polishing and improving them as much as they have).

Perhaps he’s not pushing the envelope enough if they haven’t attempted the actually impossible yet.

I’ll be cheering for them in any case; the impossible has a pesky habit of becoming possible in the decades that building some of this stuff takes. There is no shame in trying to do the impossible and failing; proving the current impossibility of hard things has value.

I'm not sure any of those things you mention are "successful businesses". Interesting technologies, sure. But they have yet to prove they can make a dime off any of them. So far they are nothing but capital burning enterprises.

(No,Tesla is not profitable)

Profitability aside, SpaceX has indeed produced reusable orbital rockets, and Tesla has indeed produced salable, useful electric passenger vehicles. My comment (and this thread) is about the feasibility of the core technology itself, which is in question.

The idea is that if all of the things you have tried have worked out and done what it says on the tin, perhaps you aren’t trying to do a sufficient quantity of impossible things. :)

There’s a difference between trying impossible things and pretending you’ve already solved them. Lying about your capabilities takes resources away from more honest entrepreneurs.

Elon's work with tesla + spacex has greater value than any other entrepreneur this decade. He is honest, just pushes timelines back

Not just decade, This century

Tesla showed that electric cars can work (even if sustainable energy generation is still a huge issue), despite the oil lobby trying to monopolize cars to use only petrol engines [1]

[1] https://archive.org/details/vimeo-210171457 / https://www.imdb.com/title/tt0489037/

More specifically they proved that electric cars can be "cool" and people would want to buy them regardless if they're electric. Before Tesla most electric cars were klunky, low performing, and generally meh.

You might be surprised to learn almost half of cars sold after the turn of the 20th century were electric. We knew electric cars could work for a long time.

How is Tesla not profitable?

Are you Tim Cook or something where anything under $10B is pretty much the same as 0?

Tesla has reported profit the last four quarters: https://fortune.com/2020/07/22/tesla-earnings-q2-2020-profit.... Are you saying their quarterly earnings reports are fraudulent?

I would say they could be fraudulent, certainly in the gray zone. Their profit in 1H 2020 was just $120M despite receiving $780M from regulatory credits, which is not sustainable and scalable. If governments stop giving Tesla money, they're deep in the red again.

The government gives Tesla money because Tesla is proven to actually pay that money back with interest.

They're not paying the money back with interest.

You're moving the goal posts. They're profitable, and the fact that a large part of their revenue is regulatory credits doesn't change that. There's a risk that they may not receive regulatory credits in the future, but we're not there yet.

If I moved the goal posts, what was their original position?

I didn't imply they were not profitable. You're not disputing anything I said.

I work on these brain machine interfaces - the 5 billion number is made up but his point that we need high bandwidth is reasonable. I think the right answer is much lower, depending on the desired application.

Yes — and putting down on ‘only one bit’ of information as output is the same as saying ‘no one will want a computer’

Sorry, I don’t mean to imply that I work at Neuralink, but I work on BMI devices in general.

Nandor Ludvig follow up comment:

> I did watch this presentation for an hour -- but it was so painful for me to experience this scene of incompetence and mockery of neuroscience getting worldwide attention simply because of Musk's money (while true scientists lose their jobs because of the lack of NIH or NSF grants for their quality research) that I add some sentences here and just leave. They did not show how their robot-controlled microelectrodes actually penetrate into the cortex and find cells -- because, as every single-cell recording expert knows, this is the difficulty: not just to move each microelectrode close enough to the targeted neuron but to make sure they can be kept there for long periods while not damaging the cell either. To do this, as claimed by Musk with 1,000 microelectrodes within an hour with "surgery without anesthesia", in the pulsing brain with no neurosurgeon present is not just impossible but even its proposal is an outright embarrassment for people with more education than the Twitter-audience encouraged to send their questions. Enough. Carl Sagan's prophetic 1996 book "The Demon-Haunted World: Science as a Candle in the Dark" predicted an America sinking in "superstition and darkness". This time has arrived. -- ---Nandor Ludvig, MD, PhD

I must say, it does feel like history repeating itself. It seems that anyone who is an expert in a field that Musk enters ALWAYS bets against Musk.

And then, in the fullness of time, they are often proven to be wrong because Musk finds another way they couldn't imagine - probably because he works as a multi domain expert and from a viewpoint of first priciples.

It will be interesting to see if that is how this one plays out.

It makes me sad that a technical comment from someone in the field thats widely echoed somehow compares to funding an EV and space travel startup. No one claimed Musk couldn't do that...

You don't think that Musk had any naysayers thinking SpaceX would fail?

Certainly there were: https://www.forbes.com/sites/beltway/2011/05/31/the-case-aga...

A quick googling probably will find early naysayers for any significant today's success.

Of course they had naysayers.

But launching rockets is not impossible. Recovering them is hard. But it's a bounded problem.

This is another one or two orders of magnitude harder.

But they were people loudly claiming that landing and recovering the rocket is impossible. And uneconomical. All the way past first demonstrations, and it was only until SpaceX started doing it as a matter of course that these voices finally stopped.

You had people saying that trains were impossible and that passengers would suffocate once they started moving (in the 1800s)

There are always going to be delusional people at every occasion in history.

Musk’s current wealth is higher than the total revenue of all the companies of which he was CEO, combined. Not profits, revenue. I don’t get why people are impressed that he sent astronauts into space when he’s been funded by equity markets and subsidies to the tune of billions of dollars. His feats are impressive in the same way building a skyscraper is impressive, but the trick is to find a way to do it sustainably and honestly. Musk has failed at both.

That being said, Tesla is currently bigger than the 4 biggest auto makers in the world combined. Remains to be seen if they can justify that valuation.

What does his wealth have anything to do with the companies revenues?

Nothing apparently.

Musk will drive his Tesla on Mars and you'll still be calling him a failed huckster because you don't like him.

He won't.

His Tesla is already in the general area of Mars, so he's half-way there :P.

No, it isn't.

It is, ∆v-wise.

Where did I say he failed? He’s obviously succeeded. Of course he’s a huckster but he’s good at it, or people are just easy to fool these days.

> His feats are impressive in the same way building a skyscraper is impressive

Um, but isn't he chief rocket engineer at SpaceX for example? And are there not numerous stories and anecdotes of people working at SpaceX who say they are amazed at how good his brain is, how fast he calculates ridiculously complex equations in real time. And, didn't SpaceX send people to ISS and create reusable rockets and design new machines to weld steel in a way that had never been done before and, well, so much more.

Isn't that a little bit more impressive than building a Skyscraper? But anyway, building a skyscraper is pretty cool why is that the bad thing to do here?

Musk the media legend is a much different person than Musk the man. He is not a rocket scientist, he didn’t drop out of a phd at Stanford, he didn’t found Tesla, and he almost killed PayPal. Link me to one piece of technical work he has ever done, not something his employees have done that he slapped his name on.

Are we applying the same measuring tape to Musk as to other great businessmen?

How many lines of code did Steve Jobs write? How many circuits boards did he design? Show me the work that wasn't done by his employees that he slapped his name on?

How many lines of code did Bill Gates write for Windows? For Excel? For Internet Explorer? He's clearly a technically incompetent fraudster stealing credit from his employees.

How many warehouses did Jeff Bezos design and build? What's his technical contribution to Kindle? How many packages did he personally deliver? Clearly he's technically incompetent fraudster stealing credit from his employees.

Musk conceived, funded and managed all his extremely successful companies. In case of Zip2 he was even the first (and only) coder. I really don't get this desire to bring him down by denying that his achievements are extremely impressive.

Musk didn't conceive Tesla, he joined as investor and is called founder after a payoff in a lawsuit to get the title.

Also unlike many other business leaders Musk has demonstrated his exceptional technical knowledge of the fields he works in as demonstrated by his Physics first principles approach.

This is moving the goal post. The reply that started this chain had the following to say:

> Um, but isn't he chief rocket engineer at SpaceX for example? And are there not numerous stories and anecdotes of people working at SpaceX who say they are amazed at how good his brain is, how fast he calculates ridiculously complex equations in real time.

There's a difference between being a good business person (Elon is exceptional at marketing and execution) and a genius scientist / engineer. Elon has tried to market himself as the latter, and has a following online (mostly male and interested in tech) that believes it.

There hasn't been a groundbreaking paper, a patent or some other innovation that Elon personally contributed to that I'm aware of.

> There hasn't been a groundbreaking paper, a patent or some other innovation that Elon personally contributed to that I'm aware of.

What about these?

https://www.google.com/search?tbm=pts&source=hp&ei=4iuVW4jkO...

This link is better, since the results in your link is doing a full text search on the document:

https://patents.google.com/?inventor=Elon+Musk

While great, having worked in these environments before, we almost always put our bosses on the patent if they contributed to the discussion in any way[1]. Giving the benefit of the doubt, would those contributions be considered substantial?

1 - https://patents.google.com/?inventor=Jeffrey+Bezos https://patents.google.com/?inventor=Steve+Jobs https://patents.google.com/?inventor=William+H.+Gates https://patents.google.com/?inventor=Mark+Zuckerberg

what the fuck does fucking being male have to do with this?

I don't know. It's the demographic on social media gravitating towards Elon.

If I had to guess: it's an inspirational story if you look, think and have a similar background to Elon and that demographic idolizes and sees Elon as a role model.

i just dont get why this deserves mentioning? espefially in a clearly negative way

> Are we applying the same measuring tape to Musk as to other great businessmen?

Why wouldn't we?

Gates was some kind of computer whiz kid in his early days, and who did write a lot of code. I noticed you picked certain later products, I guess he wasn't a involved in those? Jobs obviously wasn't the technical mind in his ventures. We can give these people credit for achievements as entrepreneurs without getting caught up in mis-attrtibuting credit for technical skill.

Steve Jobs never slapped his name on other people's work. He never claimed he was an engineer. The original Mac went to manufacturing with the names of the whole team engraved on the inside of the case [0]. For better and for worse, he viewed himself as a product visionary, and that's exactly what he was.

[0] https://www.cultofmac.com/122408/signed-by-steve-jobs-co-sig...

>> Steve Jobs never slapped his name on other people's work

I'm not sure he told Atari he couldn't program:

>Jobs convinced Wozniak to work on the game during his day job at Hewlett-Packard, when he was meant to be designing calculators. At night the two would collaborate on building it at Atari: Wozniak as engineer, Jobs as breadboarder and tester.

>Allegedly, Jobs told Wozniak that he could have half of a $700 bounty if they were able to get the chip count under 50 (typical games of the day tended to require around 100 chips). After four sleepless days that gave both of them a case of mono (an artificial time limit, it turns out: Jobs had a plane to catch, Atari wasn't in that much of a rush), the brilliantly gifted Wozniak delivered a working board with just 46 chips.

>Jobs made good on his promise and gave Wozniak his promised $350. What he didn't tell him -- and what Wozniak didn't find out until several years later -- was that Jobs also pocketed a bonus somewhere in the neighborhood of $5,000. Though it's often reported that this caused a rift in their friendship, Wozniak seems to have no hard feelings.

https://www.gamasutra.com/view/news/127537/Steve_Jobs_Atari_...

Jobs could and did program. He couldn't do what Woz could, though. Nobody could. And nobody could create or sell a vision like Jobs. Apple was a great division of labor.

> What he didn't tell him -- and what Wozniak didn't find out until several years later -- was that Jobs also pocketed a bonus somewhere in the neighborhood of $5,000.

That part is crappy for sure.

https://en.wikipedia.org/wiki/X.com

https://en.wikipedia.org/wiki/Zip2

> how fast he calculates ridiculously complex equations in real time

The sad part is that this is not sarcasm.

I vaguely recall from the first Neuralink presentation that the cameras used on the robots are fast enough to take the pulsing of the blood into consideration and compensate for it.

Personally I think even though it's an amazing technical accomplishment, the robotic electrode installation might be the simplest part of this whole thing. Once you have the electrodes installed, you might be able to pick out certain patterns, and maybe even restore motor function to people who have lost it.

But the part about being able to replay memories seems incredibly far-fetched. We have no idea how to really interface with the brain. We can only really talk to a small fraction of the neurons, because you can only fit so many electrodes and threads compared to the billions of neurons. It might end up being incredibly challenging to connect a new interface, like if you suddenly grew a tail and had to learn how to control it.

Consider that advances in decoding how memory and other aspects of the brain work isn't going to be possible until you can read from neurons in real time with high enough temporal and spatial resolution.

Neuralink may be the enabling factor in massive advances in neuroscience, not actually a consumer product, but a microscope.

That's not necessarily useful. Trying to understand the brain starting from the neuron level is arbitrary and certain to be an extremely difficult problem, because of the incredible scale of the brain.

However, computation doesn't happen solely at the inter-neuron communication level, but also inside each neuron itself.

Finally, trying all of this in mammals and even humans, instead of trying to first do it on something like a nematode or even simpler organism is such an obvious way of putting the cart before the horse that I can't take it seriously.

There is 0 chance that this approach will have any kind of success in understanding memory and thought. I am extremely hopeful they will do some extremely good work in helping give people with various spinal chord injuries some amount of control over their limbs, including the ability to sense, which would be a tremendous life improvement and would make the whole thing worth it. But it's much, much easier to reach half the speed of light with chemical rockets than it is to understand software by studying individual transistors in a running processor.

"There is 0 chance that this approach will have any kind of success in understanding memory and thought"

You don't need to understand it, you just need to be able to derive a correlation. Look, we've already been able to do this in a crude fashion with both brainwaves and PEM and fMRI.

What you're saying is that you'll never understand biology without studying DNA, and molecular biology, but you can "understand" a complex system at many different levels. Or locating say, the language center of the brain at a gross anatomy level is useless. All of these coarse grained high level scientific discoveries have advanced the state of the art.

Neuralink isn't try to decode the bits in your neurons. They're trying to construct a broad association between spike patterns and thought processes. We've already had success doing this, they're just upping the resolution. And to say that this can't grant your more than 0 insight into how memory works is as silly as saying that brain injuries to things like the hippocampus didn't grant any kind of help in understanding memory formation.

Put another way, 1) https://www.youtube.com/watch?v=1_yaQTR3KHI 2) https://www.youtube.com/watch?v=gWoPI-VoFV0 3) https://www.youtube.com/watch?v=fs2GDSYYCoA

I think the main difference here is that Musk's other ventures are all about technologies that already existed, but were prohibitively expensive or difficult to sell to a mass audience. He and his companies have done great things to promote these technologies, and to refine the tech to make it cheaper and more sustainable, not to mention fighting entrenched players who have a financial interest in seeing him fail.

But figuring out how the brain works and how to interpret its signals using a computer is in a completely different league. Maybe he'll prevail nonetheless -- he's assembled some really smart people in the field to tackle the problem -- but I think it's reasonable to remain skeptical at this point.

Yes exactly! He prefaced his presentation by saying the purpose of it is to attract the more talent. That’s the right move but by definition it will be very promotional in order to convey a vision for engineering things that are physically possible but face significant engineering hurdles.

To get this technology to work as envisioned will require hundreds of technical breakthroughs across many fields.

> by definition it will be very promotional in order to convey a vision

I think that this is where Musk's endeavors are different and why Tesla stock prices are constant source of disbelief.

He is able to successfully sell grand vision - and moreover - support it by showing results. Sure there is good old Toyota, it makes good reliable cars but it's just a car, there is no vision - I think that some people may want to buy Tesla or other Musk's company product just to indirectly support other projects.

> true scientists lose their jobs because of the lack of NIH or NSF grants for their quality research

Advertising and marketing is important. Doing things in the open does matter, live streaming double booster landing, pointing out that small steps here and there lead to big grand dream of landing on other planet does matter.

I am skeptical but hopeful as the technology can make a big personal difference for me. However i think one must be happy that someone is tackling the problem in the way Elon is doing.

Agreed! We're not going to get anywhere if we don't work on the problem. But I think people are expecting Neuralink to put an AI in their brain in 10 or 20 years, which isn't realistic.

And sure, as another commenter posted, a monkey could control a prosthetic back in 2008, but even now controlling a prosthetic with your brain is fatiguing to the point that you can't do it for very long. Time and research will improve that, but I think people are expecting that on a pretty short timeline too.

In a loosely sense, this is also a very old technology but refined. There are some fotages of monkeys able to brain control a prothetic arm back in 2008 on Youtube. [1] So, what they're currently doing is just an update on steroids from old science. It will remain a mistery to me if they'll be able to crack some basic science about consciousness as Hodak pointed out. The side effect while developing this tech will produce improvement of countless lives. That is highly doable, no doubts about that.

[1]: https://youtu.be/sm2d0w87wQE?t=49

Not only can you have animals control stuff, but you can control the animals with just thoughts too: https://www.cnet.com/news/scientists-connect-a-human-brain-a...

Intercontinental control via human-brain to rat-brain happened in 2016: https://journals.plos.org/plosone/article?id=10.1371/journal...

Heck, you can buy a controller kit for roaches right now, marketed for the kiddos: https://backyardbrains.com/products/roboroach

If you'd like, you can also hook your brain up to the cockroach : https://journals.plos.org/plosone/article?id=10.1371/journal...

This kind of control is nothing compared to what Neuralink has set out to do. Also note that all of these existing brain control devices are incredibly fatiguing. You get minutes of use before you just can't do it anymore. That's light years away from embedding an AI in someone's brain. Their goal isn't steroids on old science, it's all the steroids in the entire world, plus all the cocaine in the entire world, plus all the caffeine in the entire world... and then some.

Also not sure how 12 years old is considered "old technology", when we're comparing this to SpaceX, that has been improving on tech originally developed in the 1960s.

Retro-propulsive landing of an orbital class booster has not been ever demonstrated. Falcon 9 First Stage is the only booster that ever existed able to do it. The full-flow staged combustion rocket engine is also a new technology that will fly for the first time in Super Heavy Starship. Rocket science is literally a synonym with "difficult".

Here you have how SpaceX looked like in 2002: https://i.redd.it/00ikcxo2z0f01.jpg

My guess is that in 20 years people will be also saying how he took an existing technology and sold it to a mass audience.

DC-X.

> Here you have how SpaceX looked like in 2002: https://i.redd.it/00ikcxo2z0f01.jpg

I've seen that dancing Elon picture everywhere but now I finally know the source. Thanks for sharing!

You should compare apples to apples right? SpaceX is a government contractor. The question is, in 18 years as a contractor, have they done more impressive work than other contractors. It took them 18 years to send people to space. Is that impressive for a military contractor? Just for context, project Mercury took 3 years to send someone to space.

The sad thing about SpaceX is that because NASA was tasked to build a commercial space industry (funding SpaceX and others) they slowed the pace of development of the technology by limiting the funds. I'm sure if NASA made SpaceX and it's other contractors cooperate, the funds would have been spent more efficiently too.

> project Mercury took 3 years to send someone to space.

On Nazi rockets. It wasn't from scratch.

Ha, and the other literal synonym for “difficult”?

Brain surgery!

Mitchell & Webb made a pretty clever comedy sketch about those two professions.

https://www.youtube.com/watch?v=THNPmhBl-8I

I think you underestimate the complexity of the human brain, and overestimate how much we understand it.

There is no even remotely existing technology that does what Musk wants to do. The stuff that does exist is in comparison a stone-age tool.

While SpaceX's achievements are considerable, they are not in the league of the achievements necessary to do what Neuralink has set out to do.

Again, I acknowledge that I could certainly be wrong about this in the long term, and Musk has put together an impressive team, but I think skepticism is warranted, and comparisons to Tesla or SpaceX are lacking in scale.

As somebody said, if a old and renowned scientist says something is possible, he's most likely right; if he says something is impossible, he has a very serious chance of being wrong.

Clarke’s First Law: “ When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.”

"If an expert says something can be done he is probably correct, but if he says it is impossible then consider getting another opinion." — Richard Hamming

If it is possible, it's possible in the sense that it's possible to approach lightspeed with a chemical rocket. The predominant method in modern neuroscience for recording from a large number of cells is fluorescent microscopy with genetically encoded fluorescent sensors. Whole brain imaging is now routine (at least for drosophila, and albeit probably not at cellular resolution), afaik ephys tops out at 1000s of cells.

Gotta keep in mind that there are already several extant well-funded and highly motivated organizations devoted to studying the brain... and they're not doing whole-brain ephys.. they're doing advanced fluorescent microscopy

On February 25, a director at the CDC held a White House press briefing in which she said that the CDC anticipates that lockdowns will be necessary to contain COVID-19 [0]. In response, President Trump held a press conference on Feb 26 in which he said that the number of US COVID-19 cases is going to be 0 within a couple of days [1].

What I'm saying is that you can ignore and disparage the experts all you like and sometimes get away with it. But then sometimes reality sets in.

[0] https://www.statnews.com/2020/02/25/cdc-expects-community-sp...

[1] https://www.whitehouse.gov/briefings-statements/remarks-pres...

I mean, there's a difference between "impossible" and "impossible today".

and when an md/phd says something is "impossible today" and a salesman who thinks biology is a "software problem" says it's possible for the right price, I know who I'll bet on

People who say it cannot be done should not interrupt those who are doing it.

Even if it doesn't work for stimulation, this is _HUGE_ for capture. Nobody has captured anywhere near this much data non-destructively before.

It still isn't clear those will end up non-destructive in the long-term. biological processes surrounding physical brain health are far from fully understood, and it is possible there could be long-term maladaptive outcomes.

That's why they're testing on pigs first. But even if this is not viable in humans, this will be HUGE for neuroscience and our understanding of the brain. Because right now our understanding of the brain is limited to a large extent by the inability to capture information from it in realtime, and inability to selectively stimulate it. That neuroscientist quoted upthread should be in awe of these new capabilities. It's like he's working with stick and twine and someone showed him a rocket engine. Instead they just sound like sour grapes - Neuralink is doing something an academic has no ability to do and will never have the ability to do because he's in academia and he can't build surgical robots or microscopic tungsten carbide wires.

I have a hunch this will take longer but it will lead to interesting science and applications nonetheless. Also i believe the existence of a publicly discussed company like this increases the speed of research and nowadays people are much more responsive to accepting that not everything will work perfectly right out of the gate, the journey of Tesla is actually a great example.

So, Theranos for brain science?

Well, that's if you ignore Hyperloop, The Boring Company, Tesla auto-pilot - many failed experiments as well, or over-promising and under-delivering.

> Tesla auto-pilot

Huh? It's being improved constantly. Nothing like it in the car industry.

It is not, and will not become in the next 20 years, a self-driving mechanism, as it has been designed to be.

IDK about the 20 years claim, but he's definitely consistently said that full autonomy is just around the corner and that when you buy a new Tesla the hardware is already fit for full autonomy and just needs a software update with the algorithms to enable it... Both are bold claims and at least the first claim is consistently being proven wrong. I guess it creates more enthusiasm than "we are working on it", and saying that the hardware of a new Tesla is fit for autonomy means that fans buy a new one and not a second hand Tesla.

I think that given enough dedication and investment, it's certainly possible that we'll have fully autonomous cars one day, and it's likely that Tesla will be one of the first companies with such a car, but I'm just annoyed by their constant "it will be around in 6-12 months" that they repeat year after year.

Right now Tesla may literally be the last company likely to get there:

https://www.cnet.com/roadshow/news/self-driving-study-naviga...

And those are not 'bold claims' as much as 'outright lies'. There is no known path right now to full autonomy, so to claim your hardware is ready for it is nothing but a fantasy.

That list may not be the most reliable. It pretty much seems like they put Tesla last because of the bold estimates Musk gives for full self-driving. The part where they criticize the software seems disregard-able because it's impossible to find a standardized measure of comparison when all the other competitors don't have a product one can buy and test, and the environments they run in are far less diverse than the landscape encompassed by Tesla owners.

What about the production efforts from Nissan and GM which have launched level 3 with higher safety and no fanfare?

What about the other 10 companies who have launched similar systems at level 2 with much better safety records and again no fanfare?

What about them? They only did it because Tesla is forcing them to do it. You think they would have pushed EVs and things like autopilot to the consumer at this pace for any other reason?

They haven’t really earned the right to be talked about.

Tesla is a pioneer of mass market EVs, but is not in any way a pioneer on self-driving cars.

Okay, which major car company is putting in auto pilot features in their cars today (as in, if I go out and buy this car I’ll have it today).

Open to being educated here, and I’m hoping the answer is ‘a couple believe it or not’.

Every Cadillac currently + the new GM EVs come with Supercruise.

Ever heard of the Nissan Leaf?

The Boring Company and Tesla auto-pilot are still alive though

The Boring Company is boring boring regular tunnels, nothing like Musk's fanciful visions that civil engineers have scoffed at.

Tesla auto-pilot is far behind even the state of the art in self driving (Waymo), and even that is far behind actual self driving. So sure, some good has come out of the project, no doubt about it, but the naysayers have also been right.

The tunnels themselves were never the innovation, it was the TBMs. Have they not made progress there? I imagine not much since I haven't seen them on here in a while.

What are you basing your assessment of Tesla auto pilot on? Genuinely curious, if you could point me in the direction of some resources on the topic I would appreciate it.

He has repeatedly said "full self-driving" is coming soon with the hardware that they've already sold. That is utter nonsense. Here is a very recent example but he's been saying it for years:

https://www.bbc.co.uk/news/technology-53349313

I've seen this described numerous times. Here is a quote from a research company called Navigant[0]:

"Tesla continues to make high-profile promises, including having one million robotaxi-capable vehicles on the road by the end of 2020," Navigant's report reads. "However, the performance of its systems remains inconsistent and its products do not match its proposed mobility business model."

[0]https://www.cnet.com/roadshow/news/self-driving-study-naviga...

Hyperloop was an idea, not a company. Musk only released a white paper containing his thoughts on the topic. You can't compare a successful (and arguably revolutionary) company to a glorified blogpost!

The Boring Company is still going on, with about as much speed as it was intended to.

Tesla auto-pilot is still work-in-progress. Behind the promised schedule, but so were propulsive landing of rockets and Falcon Heavy, and yet here we are.

> Behind the promised schedule

The promise has been full self driving robo taxi fleet by the end of the year for years now. Is there an actual "schedule" of features that are missing to reach that goal or is he just repeating the exact same promise without any sign of progress to back it up? His competition still doesn't have a general purpose solution and they at least aren't cheapening out on the sensor data.

Musk said that they feel confident to achieve the implant of the link in under an hour and without general anesthesia. Is their any difference between anesthesia and general anesthesia when working on the brain?

What do you mean by the qualifier 'when working on the brain'?

Typically avoiding GA is a laudable goal -- it's expensive (even in countries with public health cover), there's a lot more risk, and it would usually involve an overnight stay in a hospital (which might just come under the cost and added risk categories).

In the case of this process, it would also involve a human (anaesthesiologist), so on top of risk and cost, you then have resource constraint issues compared to avoiding GA and letting a robot do application of LA and the operation.

Open brain surgery often involves keeping the patient conscious as a good way of verifying nothing wrong and/or something right is happening during the operation -- this is probably less of an issue with this kind of repetitive, relatively low-risk, procedure.

> Open brain surgery often involves keeping the patient conscious

Many times they aren't fully conscious but still heavily drugged, just in a state of "not asleep" enough for basic tasks/communication needed to perform the surgery.

Also he doesn't explain why most universities lose funding. Unlike private companies (especially managed by Elon) most research teams are either incredibly inefficient or have to deal with an incredibly inefficient bureaucracy machine. Companies like Neuralink or SpaceX bring brilliant, efficient teams together with an efficient "machine" (the company), which is a winning combination.

>getting worldwide attention simply because of Musk's money

It's not though. A bit the money but more his track record of building cool high tech stuff that works - rockets, cars etc.

> To do this, as claimed by Musk with 1,000 microelectrodes within an hour with "surgery without anesthesia", in the pulsing brain with no neurosurgeon present is not just impossible...

The not-so-secret secret sauce that will make it possible is AI.

I'm certain it will be possible, in the not too distant future, to build AI models that can successfully control microelectrodes to penetrate into the cortex and find cells with superhuman precision.

Consider that we now have "dumb" AI machines like GPT-3 that can model the statistical structure of human language. And note that there's nothing new or interesting about GPT-3's design except its size. GPT-3 is just a sequence of really large transformer layers.

I can imagine a robot with the scale of a GPT-4 or GPT-5 successfully controlling those microelectrodes.

I must say, it does feel like history repeating itself. It seems that anyone who is an expert in a field that Musk enters ALWAYS bets against Musk.

And that's exactly what Musk said to himself when he told people to take their hands off their Tesla's steering wheel. [0]

[0] https://www.wired.com/story/elon-musk-tesla-autopilot-60-min...

What an idiot. This has all the hallmarks of being wrong. It’s highly emotional. Lacks specific details. Best of all, it reveals a total lack of understanding of what is going on, which might have something to do with the fact that he freely admits he doesnt even bother to watch things all the way through because it’s... embarrassing? Pegging Elon as some fool with lots of money, the biggest indicator that this guy is just another academic idiot.

I just checked this guy out. He made yet another comment on tc about how musk is just another Theranos (obviously wrong, go watch the next booster landing if you disagree) and how his own device was much better, but couldn’t reach production because of evil venture capital culture, which musk is a part of. He really emphasized that his device was much better. Looking at the patents he filed, it was a sub-dermal drug delivery device with some electrodes to sense. So nothing to do with neuralink, not better than it in any way except maybe one could argue that the much simpler and practical device would be of more benefit overall in the short term. This guy is a liar on record as I have just shown. Academia is filled with these kinds of people. Loud mouths who are good at squawking the loudest to get grant money and brown nose all the necessary people to move upwards. Typical academia in 2020...

I really want to know the inflammatory response over the long run — this feels too good to be true.

> I also know that in order to make any meaningful link between a human mind and a computer one must record the identified firing of at least 5 billion neurons from the "mind-generating" association cortex in behavioral and environmental contexts

Wrong. I'm no brain scientist, but this fungus [1] shows the way.

If the scope of the device is sufficiently constrained it wouldn't need to interact with 5 billion neurons.

> They did not show how their robot-controlled microelectrodes actually penetrate into the cortex and find cells -- because, as every single-cell recording expert knows, this is the difficulty (below)

If the difficulty lies in actually embedding the implants, then a biological solution such as this fungus should be used, with the work being done to genetically engineer it to make it programmable through a device.

Just mail me my royalties.

[1] https://www.theatlantic.com/science/archive/2017/11/how-the-...

> If the difficulty lies in actually embedding the implants, then a biological solution such as this fungus should be used

That article says that the fungus doesn't go anywhere near the brain.

"Hughes’s team found that fungal cells infiltrate the ant’s entire body, including its head, but they leave its brain untouched."

It's good that they are trying to do such stuff, but a lot of their goals are indeed impossible. This is like taking a running processor, connecting to a couple of its legs and trying to figure out what an operating system is currently doing.

Impossible is a word that has been often used with SpaceX and Tesla. I would be careful using that word to describe anything Elon is working on.

I like Musk, he is an awesome manager and inspirer, and clearly technically smart. But the actual breakthroughs come from the actual science by the actual scientists. And the actual scientists know how little do we know about the world (and brain in particular).

This is a great example of the walled garden. There is a serious issue where experts refuse to accept that outsiders can have impact on their field. Let’s say neuralink does fail, it’s still very likely they push the field forward in some meaningful ways. Most likely they will innovate on hardware pushing the curve for what is possible outwards. But sure, let’s diminish all work that isn’t done by academics as impossible and merely false advertising. Track records don’t matter if you don’t have the degree.

I'd be happy to be corrected by an actual neuroscientist, but I don't think some folks realize just how far away the Neuralink team is from being able to do any of the things they are claiming. Inserting electrodes in the brain is the easy part. The reason no one else has yet bothered with this level of miniaturization is because it won't be of much benefit until you can figure out a way to decode the signals reliably. This is extremely daunting, and it's something that has occupied lots of brilliant people full time for decades.

I actually really hope that is the case. I have a feeling of existential dread every time I imagine all the effort I am putting into learning and acquiring new skills all going to waste when those tasks become absolutely second nature to everyone, irrespective of effort or time investment.

Although, I am all for fixing the existing problems that it obviously could solve; I fear my mother will succumb to Alzheimer's, and something like this is dearly needed. It just seems like we won't know when to pull the brakes.

> I actually really hope that is the case. I have a feeling of existential dread every time I imagine all the effort I am putting into learning and acquiring new skills all going to waste when those tasks become absolutely second nature to everyone, irrespective of effort or time investment.

I'm not a neuroscientist, but with such a limited understanding of the brain, beyond just the anatomy and the nature of neurotransmitters, let alone consciousness, I think you can allay such fears for the time being.

I think this could have significant benefits for neural diseases like Alzheimers, Parkinsons, ALS and possibly anxiety and addiction/mania based maladies but the idea of simply being able to 'download' the ability to learn a new skill in real time seems entirely preposterous.

I say that because some of the skills that I have been able to turn into careers have actually required I learn something where my brain no longer interferes with the process so that it is done effectively.

The notion of muscle memory is real, meaning you cease thinking about it and it just occurs out without any effort due to continual practice and repetition: my question is how can your body recover from an unexpected/unpredictable error if you haven't already practiced that same scenario over and over, and you were only given the default normal operation of x skill? Most learning is adaptable and gained from observed behavior, be it our own or others, to mistakes.

Experience is a funny thing, and our brains are flawed in recalling many of our most cherished or dreaded events and experiences accurately; I honestly think this could be a medical boon for the aforementioned diseases, but the Matrix-like learning will probably be a quixotic pursuit, in my opinion. One that sounds awesome on paper but would be horrible in practice.

I agree with most of what you say; however.. if you consider some novel ability like speaking a foreign language, I would think that your aptitude in such a skill is a product of many factors, albeit memory and recall being paramount. If what Elon says about this device is true, and we can use it to mentally interface with the web, then I don't see what would stop us from speaking new languages "on-demand". This is a very naive and layman perspective but I believe I've heard Elon claiming ideas like this around the potential for this device. Not sure what the delta is between practical applications like disease treatment and this somewhat sci-fi scenario, but I hope it is large.

I don't feel existential dread, but I certainly do a feel a level of demotivation.

That being said, neuralink not being able to do this at the present moment is of no help to people who need to learn right now.

This is exactly how I feel about this. I hope this will help the people who are in need but I really don't think we'll know when to stop. I don't like the idea of this implant to be necessary to compete in the future. Granted it is still far in the future, but it is still kinda frightening knowing just like AI this could go out of our hands anytime.

Unfortunately, there is 0 chance that that this technology will be able to help with stuff like Alzheimer's (a disease that we don't even know the biology of - plaques are looking increasingly unlikely as the real mechanism) or any kind of higher brain function within our grandchildren's lifetimes. We don't even understand how computation happens in single-celled organisms, trying to jump to interpreting memory in humans is just so far off that it's like 10th century surgeons dreaming of re-attaching nerves.

However, there is probably hope for helping people who have lost control of their limbs through spinal chord injuries and similar effects within our lifetimes - this is a much more feasible problem that is already being worked on. Maybe even some fun gadgets for "thought"-based gameplay (simple movements).

Perhaps having more bandwidth would improve signal recognition and decoding?

Well, in the update they showed themselves predicting the position of a pig's limbs as it walked on a treadmill.

If your bar is reading thoughts & dreams, you'll probably be right. But that's probably not what is needed to be useful in helping people with variety of ailments.

Not to take anything away from the Neuralink team, but predicting joint movements in a steady gait is one thing. Predicting join movements in starting, stopping, changing speed, and accurately manipulating objects is quite another.

Presumably they start with easy walking and build capabilities incrementally.

The model to push impossible tech has worked... cars and rockets.

That's kind of the thing though...those took 26 bil in subsidies to generate 20 bil in revenue to duplicate existing tech (spacex def gets credit for reusable, obvi, but Tesla man, yuck)

I’m not sure about the numbers, was thinking about the model to progress technology rapidly.

The reusable rocket tech seems worth it alone.

Not to take anything away from the Neuralink team, but predicting joint movements and angles is about the edge case of what can be achieved with semi-invasive and non-invasive bci.

Recording from the nerves predicting joint movements in a steady gait is definitely already achievable.

It can be predicted with the same precision/throughput, better than left/right:1/0 once per second?

Can you share some link?

Is your hypothesis that there is nothing they could do that would both be not too far from what they have now, plus valuable?

I believe that Neuralink has potential to do great things. However, like many other Musk ventures, they vastly exaggerate their own accomplishments while downplaying remaining challenges. Most brain-machine interface advances will probably be made elsewhere.

I think that's just how progress gets made quickly. As Larry Page would say, "have a healthy disrespect for the impossible."

For what they are trying to solve for initially, we’re talking about people that cannot even move a finger.

Not to burst your bubble but I don't think they're not predicting X,Y,Z limb positions in that demo. The traces on the left don't line up with the video, what would they actually represent?

It looks a lot more like the result of adding physical movement markers to a pig with an implant, and then mining the recorded neural data for signatures which correlate with the marker data. I'd be willing to wager that's exactly what is shown in that video.

I don’t see the point of this comment. Of course they are relatively far away. But they are also closer than anybody else.

No, not by a long shot. See, for example:

https://www.sciencedaily.com/releases/2019/06/190619142542.h...

Do they have a robot surgeon that installs it though?

It's non-invasive. No need. (Maybe you were joking.)

It’s an interesting article, but I’m not sure why you would imply this orthogonal approach is further along than Neuralink.

The novel research being done in the paper you linked isn’t even around the sensor suite, it’s around training AI to extract more precise signal from a noisy sensor — in this case an EEG.

The premise of the article is that EEG is being used because direct neural monitoring is too invasive and difficult.

Neuralink is essentially attempting to disprove the premise of that paper.

I’m going to guess the signal processing Neuralink will be doing is much more sophisticated than whatever that paper was doing on top of an EEG feed, just because the data rate from Neuralink is so much higher.

The demo in which they were able to read limb position was pretty impressive and shows that they should be relatively close to being able to help people that have lost limb control.

The brain is also quite adaptable and learns via feedback loops so getting the electrodes 'close enough' might be sufficient along with some physical therapy.

This has been there for a while, it is pretty simple to decipher impulses from the cortex. They are just publicizing years of research from other folks i.e. professor Jose M. Carmena from Berkeley and so on. We need more folks like Elon to publicize science...It is actually great work the team is doing.

It seems like the key advance here is the miniaturisation. Imagine one of these puppies paired up with an advanced prosthetic arm.

The key advance is the increase in order of magnitude and precision of placement of filaments embedded in the brain.

One of my favorite facts about neural interfaces... all electronics use negative charges for communications. But neurons use positive charge in the form of calcium flow. This makes it really hard to bridge that difference.

I wonder why is it so, in the end it's just electrons flowing from negative to positive end to create current, why do they behave differently inside neurons?

It’s because the flow of electrons driving current is mostly used only in one part of neurons: mitochondria. It’s used to generate ATP, which is the actual currency of energy in cells. This in turn powers systems that send and receive ions that are generally positively charged (why this is, I’m not sure) for neural communication.

So, while reading electrical signals is possible, returning them is tough because you’re so far removed from the language of the neurons when using electrons.

I totally agree, I am totally unimpressed, Elon knows how to create hype and get smart people to do something. Smart people sometimes can't do it, it take a generation of research to access brain raw information. Codos to him to collect such smart minds but we are away 10-20 years from meaningful read and write to the brain.

> it take a generation of research to access brain raw information.

I don't see why this has to be the case. I'd more expect that the pace of research will accelerate as devices like Neuralink come onto the market and allow much higher resolution and more precise data to be collected across many more individuals.

10-20 years is reasonable for more advanced capabilities but they've figured out prediction of pigs' limb movements in ~1 year, so we could have neurally-controlled human prostheses very soon after trials begin.

i can imagine how seing the recent progress made by ML technologies in processing huge amount of information in real time to extract cognitive information, one could think the same kind of technics could be applied to other fields, provided you manage to reach the same amount of raw information.

Combining the touch detection with other devices would already be useful on its own. A phone and neuralink could work together and figure out when I'm holding it and unlock automatically and then lock it when I stop holding it. The connection between them would have to be trusted and secured of course - it would need to verify the unlock signal actually came from my device.

Getting something like x,y,z coordinates of your limbs relative to the device would be a good primitive for many features (same usefulness as acceleration and orientation sensors in phones). If that was fast and accurate enough, that could be used to control robotics.

Pairing with other devices and issuing gesture commands to them would be another possibility. Sign language might be translate-able directly to bytes.

We're talking about building Pong at this point. There will be lots of creative solutions found within the constraints.

> A phone and neuralink could work together and figure out when I'm holding it and unlock automatically and then lock it when I stop holding it.

I mean, maybe, but I feel like you’re missing the potential of this device!

We're not :/ we just know our field

Disclaimer: I truly dislike this technology for my own personal reasons.

That aside, I think the problem you bring up is pretty much "solved" as the only thing that is missing are training sets. Pretty much like when Tesla just needed a bunch of driving footage to start improving its self-driving AI, and they got it.

The first Neuralinks (I believe) will be intended to gather as much data as possible in order to tackle this concrete issue. After that is going to be a quite vanilla "big data + ML" type of job.

It would be cool to see, if different people have different neuronal dynamics for the same sort of "activity". Who knows what it will be found. Regarding "the pig demo", we don't know for sure if they can do that to any pig wearing a neuralink (i.e. like installing software) or if the data/model was produced and meant to work on that particular pig only.

Agreed, ML will be able to extract signals from this data, no question. The electrodes really are the limiting factor I think, because you're going to need millions placed all over the brain and deep inside too for a truly useful implant.

I don’t think that’s entirely true because brains are quite adaptable. In similar sorts of studies, humans managed to train themselves to use the electrodes to direct things. It may never be ‘perfect’, but I think you’d be surprised how much could be achieved using such a method.

>because you're going to need millions

For sure, increasing # of electrodes + bandwidth must be on the roadmap.

You may not need to go as deep and broad as you believe. Most processing happens on the cortex anyway, the interior (white matter) is pretty much interconnection. Also, the action of one neuron could trigger thousands of others in quite remote regions, so you could probably get a decent picture of what's happening without having to cover it all.

I was honestly quite impressed by "the pig demo", I wouldn't have guessed that would be possible with the current state of their tech.

> That aside, I think the problem you bring up is pretty much "solved" as the only thing that is missing are training sets. Pretty much like when Tesla just needed a bunch of driving footage to start improving its self-driving AI, and they got it.

That is absolutely not the case, and we are absurdly far from anything like it. We haven't even been able to decode very much useful information from the "brain" of a nematode yet, and there we don't care in the slightest about its health.

Self-driving is actually an interesting example (though it is likely many orders of magnitude simpler than interfacing with the higher functions of the neural system of, say, an insect). We are still much farther away from real self-driving with the skills of an average human driver (such as not hurtling into static obstacles on a road) than anyone was thinking a few years ago. In fact, the CEO of Waymo believes that we won't achieve full, all-condition self-driving in our lifetimes (though he does believe that we will achieve useful self-driving in common conditions).

Regarding the brain, we don't even know yet where the computation takes place yet - to what extent does it happen at the cellular level, and to what extent and the neural network level? Further, the electrical signals are unlikely to be the only important part - there are many chemical substances that impact our judgement and reasoning, so I don't see why we would assume that even capturing all of the electrical activity of the entire neural network would be enough to decode thought patterns.

>That is absolutely not the case, and we are absurdly far from anything like it. We haven't even been able to decode very much useful information from the "brain" of a nematode yet

Huh? What?! Have you actually watched the update?

They decode the position and movement of the limbs of a pig as it is walking, in real-time, with uncanny accuracy.

Yes, motor signals. This is not necessarily entirely new, we have been able to take simple readings of motor signals for years, this is 'just' pushing the frontier (not to say that it isn't a great accomplishment!).

Thought processes though are a completely different beast, and we have no realistic hope of interpreting them at this time. We wouldn't even know what to look for. We're missing much more fundamental research in how thought works in even simple animals before we could have the first hope of finding something in a brain. Even thought processes such as 'that direction is more promising for food'.

Oh, I see your point. Yeah, we are not even sure how consciousness emerges. But, for sure, some sort of crude i/o with someone's thoughts could be established, and not far away.

I'm not even talking about consciousness, just basic computation in neural networks. How does the very simplest animal decide whether to move left or right in the presence of a stimulus? Can we influence that thought, not at the sensory/motor level, but at the computational level?

These are unanswered problems in worms. They are unanswered even in single-celled organisms, or at least in colonies of single-celled organisms, such as slime mold.

So I don't believe that any kind of IO will be achievable in the next hundred years, beyond possibly interpreting motor functions, and perhaps providing simple sensory input.

It's hard to build a model of a system without observing it - perhaps a reason we don't know how to decode neural signals yet is that we didn't have a way to read them. I'm sure that this vastly improved sensor will lead to better models of the brain and neural activity.

We have plenty of ways to observe neural activity, and have had them for years. This chip fab represents an incremental step. In fact if I had to choose for my lab this new chip or some other tech currently available, there are about 5-6 other things I'd prioritize over this Neuralink device.

Eg. what alternative?

1. Fiber photometry rig

2. A 1-/2-/multi-photon microscope with head fixation stage for GRIN lens imaging

3. New electrophysiology rig for single cell, paired cell, multi cell, voltage clamp, patch clamp, etc recording

4. An automated micromanipulator for molecular uncaging and optogenetics experiments

5. a rig for superres single molecule tracking experiments, PALM/STED/STORM/uPaint

6. A set of headmount Miniscopes (elon could actually help to vastly improve these)

Thanks.

Which of these could you use in vivo?

We’ve actually had ways of observing neural signals for quite some time. In order to identify the regions of the brain requiring surgery for epilepsy, patients get electrodes implanted and spend several weeks in hospital. Various researchers have recruited volunteers from this cohort for neural signal studies. A team recently managed to decode speech from neural signals from such a study!

https://www.ucsf.edu/news/2019/04/414296/synthetic-speech-ge...

Cool! This seems very similar to what Neuralink hopes to build out, aside from some technical details

That looks like a key point behind the existence of the Neuralink company.

Even among neuroscientists there will be disagreements about what is possible when.

Seems to me this is a circular problem, until somebody comes along and does the miniaturization and figures out all the software and systems you need, no serious progress will be made.

How about first trying to understand how the handful of neural cells in a jelly fish of worm can guide the animal to food and allow it to avoid danger, before trying to imagine we could help human beings recover their memory?

If you are so much better at solving these problem and your approach is so much superior go compete against them. If somebody is willing to dump 100M into making break-threw medical advances I'm not gone question their approach unless I am really, really sure I actually know better.

Well, unfortunately, unlike Elon Musk I didn't have rich parents, connected friends, or the luck to be part of a really great startup idea, so I don't really have 100M to spend on my own fantasies.

But the fact that someone is willing to put money into a fantasy will not make it a reality. As I said elsewhere, I am truly hopeful that much good will come out of Neuralink in the area of helping people with motor problems and prosthetics. I absolutely commend this goal and am happy to see more people invest in this area.

But that doesn't mean money will magically supplant the decades or more of research we still need to do to until we can meaningfully even think about what thought means at a physical level.

And just because Musk has delusions of grandeur about the way he will change the world, enough so as to put his money where his fantasies lie, is no reason to start believing in the same things.

I read we can decode signals reliably, but the hardware required is the size of a large desktop, so not something anyone could feasibly lug around. With that said, I think the use of GPUs is relatively new in the field, and I don't believe anyone is using ASICs, so there may be some significant gains possible.

everyone thinks it's pretty far away, i'm just glad someones taking a serious go at this

The people taking a serious go at trying to understand how the brain works are the ones trying to simulate a nematode brain, the ones working on understanding animal behaviors, the ones studying how infants acquire knowledge etc. These are all prerequisites for any kind of useful understanding that would allow us to build a working model of the brain and start imagining how to interface it more directly with other systems.

Not doing the fundamental research and trying to pick up useful signals by arbitrarily choosing to focus on inter-neuron communication in a few areas of the exterior of the brain is unlikely to even advance our understanding one iota. We're learning much more about thought (not neuro-motor transmission, mind you!) from the people training slime molds to navigate mazes than we are learning from this.

I'm glad though that there are more people working on perhaps helping people who have lost control of their bodies to regain it. That is an absolutely worthy goal that the team from Neuralink do have a hope of achieving in our lifetimes, and where I absolutely believe they can push the state of the art.

There are a half-dozen groups (companies & in academia) that are building neutral stim & recording platforms with thousands of channels & high bandwidth comms so that we can at least capture, analyse, & transmit signals to start learning & closing the feedback loop. Think of it like designing the first real integrated circuits so that we could eventually get the 8086.

This needn't be directly in the brain. E.g. there are a lot of groups looking at the parasympathetic nervous system (ie vagus nerve) for a host of organ control applications.

Been almost 5 years since I was working in the field... But there seems to have been a lot of progress already. Big limiting factor for iteration velocity is identifying viable test subjects -- elective participants aren't really an option (few volunteers, ethically questionable, & regulation); you have to find people with diseases or conditions that can justify such extreme interventions.

I’ve had parts of my body inactive due to peripheral nerve damage after an injury (peripheral nerves mostly always heal, so function came back), and I can tell you I would have easily opted into any experimental surgery that could restore function.

You don’t have to go far to find candidates, walk into any ER today and you’ll find trauma victims with irreversible nerve damage from a common car/motorcycle accident.

I agree with your assessment: it's easier to find subjects with peripheral nerve issues. It gets progressively harder to find subjects as you move up the chain, as the risks compound quickly: peripheral > central > brain.

As you can imagine, neural link is targeting some of the most invasive -- hence animal subjects.

There is a famous quote that we should all remember today " we overestimate what can be accomplished in the short term and underestimate what can be accomplished in the long term." - Arthur C Clarke

As an actual neuroscientist (PhD student, in a past life, now a googler), I would disagree with this part: > Inserting electrodes in the brain is the easy part. The reason no one else has yet bothered with this level of miniaturization is because it won't be of much benefit until you can figure out a way to decode the signals reliably.

I performed and analyzed multielectrode recordings not entirely different from the ones used by Neuralink (Neuronexus is the brand if you want to look it up). These were pretty much the state of the art as far as electrode density in vivo goes. And we would absolutely have loved to have higher density. We even went so far as trying to simulate higher density by moving the electrode array around with a microdrive and repeating the stimulus.

I am quite sure the constraint is not on the demand for higher density arrays from scientists, it is more on the manufacturing side, the economics for further miniaturization are probably not there, in the absence of the Elon Musk cash spigot.

I fully agree that many of the claims Neuralink are making are quite a ways off though.

From some of the Q & A, it seems material design is going to be one of the biggest bottlenecks. Making the electrodes thin enough to give enough granularity, but also durable enough to provide a strong and reliable signal for many years.

Sounds like an engineering problem the can be solved with the right mix of experts, motivation driven by mission, and limitless funding by the likes of Musk to grease the wheels of innovation in this domain and problem space.

I asked my neuroscientist friend about this a few months ago and his response was that inserting electrodes and interacting with the brain is the easy part: preventing the body from rejecting them or forming scar tissue over them after a few weeks is the hard part that nobody has been able to solve yet.

Reminds me of cyberbrain sclerosis from Ghost In The Shell.

Which is precisely what this has solved!

Ah, well that is exciting then (I did watch the video and know they said the pig had the unit for two months, but I wasn't sure if they had actually solved it or not, although two months is certainly a great sign).

I'm surprised nobody talked about the potential implications of effectively embedding, now that the device is standalone, a lithium ion battery into your skull.

I'm not a battery hater whatsoever but in the case of a thermal runoff if the battery starts burning... there is no way to get it out.

I wondered about this, but is it not already the case with other existing devices like pacemakers etc?

Yes, and that's why there are such high standards for medical devices.

In the presentation Elon specifically mentions their involvement with the FDA and highlighted their focus on safety (drawing a parallel with Tesla five star ratings)

Seems like a nightmare now that I think about it, like you get a ransomware popup beamed into your visual cortex saying:

"We are in your brain, wire the contents of your bank account to the following bitcoin adres: <adress>. If you do not comply, we will blow up your neuralink, if you try to alert anyone, we will blow up your neuralink. We are watching, you have 2 hours, good luck."

"01:59:59"

"01:59:58"

...

"it was rumored that hackers for big media companies had figured out a way to get through the defenses that were built into such systems, and run junk advertisements in your peripheral vision (or even spang in the ... middle all the time - even when your eyes were closed. Bud knew a guy like that who's somehow gotten infected with a meme that ran advertisements for roach motels, in Hindi, superimposed on the bottom right-hand corner of his visual field, twenty-four hours a day, until the guy whacked himself."

-Neal Stephenson The Diamond Age

This imaginary problem can be solved by having a hardware off switch.

Imagine your brain got so used to the neuralink interactions and benefits that turning it off causes an insurmontable pain (like drugs)

Sure if someone has two improbable situations combined, they are doomed. But they are more likely to be killed by a lightning.

/If we sense that you think about turning off your neuralink, we will blow it up/

If we can read your mind, we don't need to ask you to do anything, we will just read all your passwords and do everything ourselves.

Reminds me of https://mobile.twitter.com/tylerthecreator/status/2856708222...

To be clear, I don't think this is a solution to either.

The entire reason why people suffer from anti-social behavior delivered through computers is that computers are a real part of our world at this point. Simply disengaging may not be an effective way of dealing with it.

It's still a funny tweet though.

Lithium ion is not one thing, one type of battery, but describes a very wide range of batteries. If you are willing to pay more and use more advanced materials for the difference parts of the battery, you can make it way secure.

We don't do that yet for cars as for the amount of batteries you need, it would make the car to expensive. However for medical devices, such batteries are already in us and with all the research into batteries today, this will improve even more.

I'm more concerned about my head being drilled open, metal put into my head and having a BLE connection. The battery is the least of my worries.

The only major technical update from its initial announcement a year ago is that they're no longer hiding the transmitter behind the participant's ear. (and pigs!) Other than that, the whole thing seemed like a recruitment attempt, but that's fine. I fully support mission, and I'd love to work for them if given opportunity. To apply: https://jobs.lever.co/neuralink

Since they started the presentation announcing it was a recruitment event, I think you are right.

They put it in a pig brain and can use neural signals to predict the pigs behavior - that's major progress

Seriously! How much faster would people expect them to be able to run with this?

Something like this is the work of a lifetime. Elon is 49, and isn’t running out of funding any time soon. It’s an incredibly costly and risky endeavor which could improve a lot of lives.

I think the idea of putting it into an otherwise healthy individual is way, way off and kind of a distraction? But in the meantime if this can somehow help treat e.g. Alzheimer’s it would be an absolute godsend for millions of people.

Not behavior, motor signals. Major difference, but still, a nice achievement.

The efforts made by Neuralink’s team are impressive and, IMO, the idea is very serious, with lots of practical short term consequences as well. Someone said that getting electrodes into the brain is the easiest part, but I beg to differ on this, for the simple reason that data is more accessible, abundant and clean than ever. Accessibility of data has obvious implications on the pace of progress in using said data.

Now, on the demo and its tech: some other commenters are concerned about the safety of the lithium battery, but what about the Bluetooth? The stacks I used in my embedded projects were simply horrible and I’m heavily biased against BT even on principle :D

What about the electrodes being, in the end, simple conductors and, so, antennas? I’d hate to be in a NMR machine with a NL in my head (thunderstorms too? An attacker with a jammer?!).

I was expecting a write demo too, but maybe it’s not something you want to be streamed online for the general public.

Last thought: the pig limbs prediction task. I think that the prediction was greatly aided by the chosen setting of the experiment (tapis roulant), but still impressive. I’m sure that today we saw a very important piece of human history unfold before out very own eyes

> some other commenters are concerned about the safety of the lithium battery, but what about the Bluetooth?

Being Bluetooth, you probably have to retry a few times to get it to connect and will fail frequently from WiFi interference. Also, the BT software stack is a mess.

For example I have problems connecting AirPods to MacBooks (recent ones), same company on both sides, still not working reliably. BT is a shit technology if there ever was one.

They talked about security and explicitly mentioned that the Bluetooth radio is hardware isolated in some way. They mentioned using encryption without detail but I expect that the radios are treated as untrusted. Also they are using third party penetration testing.

> Accessibility of data has obvious implications on the pace of progress in using said data.

That is only true if the data is likely to be the right data. For motor control, electrical signals seem to be pretty much it. But we have no clear proof that the same is true for higher thought processes - it could well be that chemical signals and processing inside each neuron itself are major components of that, so the data would be massively limited.

It would also be interesting to see how well such an approach would do when tried on a microprocessor - could we tell that a microprocessor is displaying "hello world" by sampling electrical signals from its transistors and trying to apply some kind of machine learning on that data?

You might like this article titled "Could a Neuroscientist Understand a Microprocessor?": https://journals.plos.org/ploscompbiol/article?id=10.1371/jo...

Neuralink is a very interesting effort, but with brain-computer interfaces there is an elephant in the room that seems to be brushed off or overlooked all the time: the security of computers that get connected to your brain. You don't want something so vital to be vulnerable to buffer overflows and speculative execution attacks. I wonder if this is ever going to be addressed seriously.

They did. Very limited attack surface.

Limited attack surface is not acceptable for a device inside your brain.

And nobody ever said that its the only thing they do.

Okay, but it is Neuralink’s responsibility to say everything they do. If the only thing OP can find is that they guarantee a “limited attack surface”, that’s frankly not acceptable for a device in your brain.

> it is Neuralink’s responsibility to say everything they do

When they bring a product to market, that is.

The presentation was in general about Neuralink. You had 1 question about security and the guy gave an 30s answer saying it is very important, they are doing it from ground up and try to bake in security from the beginning. He mentioned one example, of seperating BLE stack on a hardware level. Seems like a reasonable answer for a Q&A of this format.

Neuralink has actually no responsibility to tell you their whole security architecture. They are presenting this for recruiting and showing the public some of their progress.

The only reason why software is unsecure is because we decided to tradeoff security for development speed. Using secure languages, formal software verification,etc you could create secure systems. Brain software would definitely have to be put in much higher standard.

Making it secure will take more time, and companies will not willingly waste time because it puts them behind competitors. Therefore security is a solution that can only be enforced via government regulation, and of course not all governments are going to regulate it, which will naturally lead to a race to deregulate under pain of being left behind.

So we might be able to make it secure hypothetically, but my bets are, we will choose not to. Case in point would be self-driving cars, which are hilariously underregulated.

> Case in point would be self-driving cars, which are hilariously underregulated.

Why is it bad?

Self-driving cars kill less people than human-driver cars per mile.

If we heavily regulate self-driving cars, we will slow down self-driving cars development. Cars will probably be safer, but the number of cars will be smaller, thus number of deaths will be higher.

No, we don't need self-driving cars regularions at this moment.

> Why is it bad?

I didn't say it was bad, I was just observing and predicting.

> Self-driving cars kill less people than human-driver cars per mile.

The types of self driving cars you are referring to are the kind being sold to consumers now, which are more like "assisted driving" than self driving. There are very few of them (i.e. they are not widely tested), they are owned by rich people (selection bias), and they ship with a giant warning not to let the car actually drive itself without supervision. So you're really testing the crash rate of wealthy educated people with driving assistance against the rest of humanity without assistance, it is no wonder the deaths are lower. The type I am referring to are the ones that actually drive themselves, the shipping trucks and whatnot.

> If we heavily regulate self-driving cars, we will slow down self-driving cars development. Cars will probably be safer, but the number of cars will be smaller, thus number of deaths will be higher.

This simply restates the point I made originally: we will be able to make them safe, but we will choose not to for whatever reason.

>formal software verification

There is a big misconception that this provides some sort of guarantee that the software will not behave in unexpected ways. It is just another tool and as such it has its own limits.

Who is to say that attacks might come from outside? Nations might mandate a backdoor from the Neuralink developers directly. Malevolent ones will surely want a literal kill switch feature, or copy the technology with their own. Imagine the horrifying power of a nation state that can kill any citizen at any time in any place with a keystroke. With that everpresent threat, you might not even need to be served propaganda to fall in line with the whims of the government.

<Aramaki> What do you mean I bricked my device?

If human brains merge with computers, I hope we are running free software in our heads. The idea of code you cannot observe or control, in your head, seems like an attack on humanity. I'd like to see this put into actual law from the start, possibly at the level of a UN Human Rights declaration. I'm not interested in living in a world with an app store for your brain.

As an aside, the technical capability for true mind-reading doesn't seem too far fetched anymore. I don't think society is currently structured to accept the consequences of that. People are not what they appear to be on the outside.

If you don't want proprietary software in your head, don't install it in your head. Why do you need the UN to stop you from doing that to yourself?

It's naive to think this will be optional for long once it's viable.

Oh you have X,Y,Z mental illness? Treatment from or inclusion in society is contingent upon getting this chip installed.

Want this job? Even if it doesn't get to the level of "all employees need to be running OurCorp Headware (tm)", being required to be running Microsoft Headwindows in order to get a job with medical benefits seems ominously plausible.

Want to use the neuraIM to talk with family and friends? I know you'd prefer a FLOSS nIRC client, but everyone you know is already on WhatsBrain.

Let alone forcibly installing such a chip as "for the persons own good". Or more likely "to stop terrorists" due to needing an excuse.

That's almost definitely going to happen. If the tech were available today we'd probably be hearing disturbing leaks of forced chip installation in people heads in various countries already. :(

We can’t even get people to wear masks in the middle of a fucking pandemic. Do you really think this is a likely and universalizable scenario?

That's a minority of people and they are chastised for it. The vast majority of people DO wear a mask. So yes, this is REALLY a likely an realizable scenario.

seeing as people can get fired for not wearing a mask, this can also be a stipulation for employment. install or starve, citizen.

What if your employer asks for it?

Tell that to the children whose parents will opt for the neural implant after birth.

Plenty of countries strut about on the world stage while maintaining complete disregard for UN declarations. I don’t think that’s going to help.

There are rare moments in big developments when individuals can significantly affect the outcomes - before the distributed consensus of the market takes over and forces us down some ugly path. In this case, as long as Musk is spearheading this development, perhaps we should get Nick Bostrom to write a pop-philosophy book specifically about merging the mind and the machine? Musk is pretty receptive to Bostrom's ideas.

And yet this technology may give some people their humanity.

> The idea of code you cannot observe or control, in your head, seems like an attack on humanity

That sounds like a description of regular brain to me.

Yeah but at least no one else has access to it.

And this assumption is based on what exactly?

applied psychology

Don't tell marketing

Reminds of a scene out of Westworld where they are talking about how the "hosts" are more human than humans since they can modify their "core directives" or the code running inside them. Cuts to the heart of the "free will" argument since humans cannot change "our core wiring."

Sure we can, that’s how we learn new skills and knowledge.

You can’t modify the process by which you learn those new skills though, can you?

That would be my thinking as well. Sure you can train yourself to learn new things "better" but there are some people who seem to be real polymaths at a level which the rest of us cannot seem to attain even with years of practice.

That's literally what https://en.wikipedia.org/wiki/Memory_sport is about.

I actually like the idea, that the brain could start to hack itself. It changes its own premise. Feels loopy too!

we can hardly observe running code on computers, even when we have the source. Programmers have trouble understanding what their own code is doing!

Can you observe and control what gets into your brain right now?

To a degree. And at least I don't have a physical corporate device implanted in my brain.

You might find the Multireal trilogy by Davis Louis Edelman interesting: https://www.amazon.com/Infoquake-Jump-225-Trilogy-v/dp/15910...

Not the best writing in the world but several of these ideas are touched upon before it kinda goes off the deep end. A fun read at least.

If the apps are going to be controlled on the smartphone as Elon mentioned, forget free software for now. Apple and Google are ready to take their 30% cut

Plenty of people here on HN will jump at the opportunity to have Apple control what they are allowed to think/view/experience. They already do this and passionately argue they love it.

“We have been subordinate to our limitations until now. The time has come to cast aside these bonds and to elevate our consciousness to a higher plane. It is time to become a part of all things.”

-Puppet Master

Sometimes I think that we are going to become the AI that Elon Musk is trying to prepare us to fight against to. This technology is literally opening pandora's box. Both the amazing and scariest thing in the world will come out. Pretty crazy times we are living. Cannot highlight this enough, we're summoning the demon hacking our brains.

If AI becomes vastly more intelligent than humans as he predicts, a neural implant is nothing more than a teddy bear to make us feel better. If we augment ourselves with AI, our expanded selves will see the brain as vestigial and slowing us down, eventually excising it, then we become full machines.

In a way, transitioning like this, rather than only creating non-human AI in parallel, would hopefully transfer some of our humanity into the machines. In any case we would look back at this transition as a natural metamorphosis. I wouldn't be surprised if this metamorphosis is a natural emergent phenomenon among civilizations across the universe.

My point is that Human + AI = AI. There's no AI around the corner as far as we know, but you can argue that the portion of machine intelligence is growing at a crazy rate. So, we're becoming slowly an AI as you correctly pointed out, maybe it's just plain evolution. I wonder if no AI comes out after all (by itself), we'll be the One created through ourselves.

The AI is already here. It's called technocapitalism. We're not becoming an AI, we're being controlled and modified by one, so as to better fit its teleology.

Is anybody else sick about literally everything people don't like being <xy>capitalism?

I'm more sick of <xy>capitalism being a thing that people like unconditionally.

Posting on HN is the height of capitalism.

I disagree, Humans already have a neural interface with AI via the visual cortex and the auditory nerve. You are forming a neural link with an AI when you have a conversation with your Google Home.

I imagine a electrode neural link would with much the same way, but faster and less disruptive. Humans would off load tasks like: information discovery, complex mathematical calculations, and making arrangements for holidays to their AI companion.

Big difference is that devices that work with these interfaces (i.e. screens and speakers) are loosely coupled and maintain context with the rest of the environment. You can remove yourself from their influence in an instant.

I disagree with the premise that the fact that I can see my living room in my peripheral vision reduces the influence of my media consumption.

The kwown argument that <<you're already a cyborg>> seems to fit nicely with the current status of humanity. The problem I see lies when you decide to integrate these machine learning models (not "sentient" AIs) as part of neural substrate at a convinient broader bandwidth. We are already seen the disasters of biases, filter bubbles and many more to come. There is a huge gap between a cyborg (or Augmented human) and an AGI. The latter doesn't exist yet but we are slowly building it through the links of our mind. That is both the real summoning the demon and quantum leap of faith in evolution.

I see the topics of biased information and neural input bandwidth as unrelated as the current rate of bandwidth is sufficient. As you mentioned, humans have repeatedly demonstrated the ability to whip themselves up into a frenzy throughout all of recorded history. So it’s a bit strange to suggest we shouldn’t adopt a technology on the basis that it might encourage us to do what we are already doing.

Also filter bubbles in media have generally been created by quite basic algorithms that only meet the definition of AI when you really stretch it. You might argue that a more advanced AI might create more toxic bubbles, but it seems to me that our society has just about perfected toxic bubble creation and any further advancements in the field would suffer severely from diminishing returns.

Taking very Elon's words: “This was futile. I tried for years. Nobody listened.”. If going backwards is just not an option, we must be more conspicuous of the dangers of new high tech deployments. I think everyone here share the same feeling: thrill but chilling at the same time. It's Inevitable by the way, on the Kevin Kelly's sense of the word: this is our technium taking shape and we better learn how to deal with it fast, like it or not.

Human society is part AI already. Humans to human society is as cells to a human being.

I think another possibility is that connection of our brains to a primitive AI will boot the AI with some human intelligence. Also, an AI based on current silicon seems very fragile. It could make sense for AIs to build synthetic biological bodies.

Don't worry, we and our children and grandchildren will be long dead before the first general AI and/or brain-computer interface (at the level of thought, not motor control) ever happen.

I feel people are vastly under-estimating just how complex general intelligence is. Even simplistic intelligence like what you find in a worm is far beyond us, not to mention insects. Thinking that we could decode human thought out of electrical signals in the brain at this stage of our understanding is like the ancient Greeks thinking they could build a rocket to the moon.

This could be the contemporary equivalent of the lie detector test -- unlikely to be proven wrong for a long time, yet influential despite of questionable results.

"This is starting to sound like a black mirror episode... Well, I guess they're just very good at predicting" - Elon Musk

It must be fun to work in a team of those assorted specialists

Not really: https://outline.com/FnfbkD

"The company is now down to just two of its eight original founding scientists."

My thoughts on @neuralink

The Good: Explore by direct observation/interaction the brain, cure countless disseases along the way, and eventually augment humans.

The Bad: Privacy, Hacking, identity theft, memory implants, infinite dystopias.

The Ugly: The rich/poor divide will be cognitive.

There were thousands of similar predictions when mobile phones were created.

Literally, when any new technology created, people started predicting dooms.

Yes, business as usual but what is actually unpredictable are second or even third order effects from the deployment of technologies. Taking social networks as example, USA's president sounds like doom to me.

Before social networks TV influenced president elections, radio before that, and newspapers before radio. Each time media changes, it is doom.

It is curious to see that the rate of change is the sole thing that accelerates between medias across time. Let's imagine a 1000000x speed scenario among virtually all human endevours. Maybe not doom but that's all folks.

By the way, I'm optimistic about the future.

This is a little bit of a false equivalency imo. You can walk away from a TV. This is literally inside your brain. Imagine if a bad actor got the reigns, or if a nation had a back door installed for their version for direct access to their citizens. Literal mind control that you can't escape from at all. No one is at the TV for 24 hours a day. Imagine if a kill switch were possible, ushering the era of death by keystroke hundreds of miles away. You can't kill someone with a TV, or a radio, or a newspaper.

> if a nation had a back door installed

So far no nation managed to install backdoors on all phones and laptops. I have no reason to believe it will be different with brain implants.

> Literal mind control

It’s very far to that.

> that you can't escape from at all

You can always turn it off. With some hardware button.

mobile phones were telephones with a battery, not something directly wired to your brain and thought processes.

For every new technology there’s an explanation why it changes everything.

And phones are just telegraph with sound and telegraph is just a mail over wires, and mail is just a messenger without a person.

Also, it is very very very far to wiring something “directly” to the “thought process”. Society will have time to adapt.

“ The rich/poor divide will be cognitive.”

It has already largely been cognitive for decades because of standardized testing and assortative mating.

Try thousands of years due to differential evolutionary pressure, population bottlenecks in Europe and East Asia and interbreeding with Neanderthals and Denisovians.

This makes me feel uneasy, and weird, but... I think I like it.

It's probably all those movies written about the subject (ie Matrix) that add to the uneasiness.

I like how he focuses on safety aspects and that it's reversible etc.

I have to wonder if the electrodes and implants are really necessary? Mary Lou Jepsen is working on reading and writing activity to individual neurons using infrared holography -- a wearable device.

If someone else is interested, I found this recent video on the topic:https://youtu.be/CqsglRFjEKI

Nobody asked the single hard question, so I'll put it here:

Those types of electrodes are known to cause permanent brain tissue scaring after 1-3 years, similar to asbestos scaring.

How do they intend to solve this problem?

If I'm quadriplegic, I'd prefer my brain tissues scarred after 1-3 years and then new implant installed in a different spot rather than being a vegetable during that time.

(And when we have these implants helping paralized people for 10 years, we will think of something. We don't need to solve all problems at once.)

Actually every comment section out there is riddled with this question and it is directly addressed in the video. Perhaps you should watch it...

I think there are a lot of questions one might ask but I’m going to say that it’s too early to ask most of them. I’m sure that once this makes it past prototype stage and human trials are on the horizon they will need to have good questions to a lot of these questions and I have no doubt they are thinking about basic and not so basic scenarios.

This is a good PR/recruiting stunt but nothing more. (Also Musk could use a little bit of training when it comes to the way he is speaking and is timing his speech)

Musk's presentation style has been this way since forever — after watching a couple of his presentations, you kind of get used to it. He's not much of a showman.

This absolutely should not be allowed to be done. We are treading on dangerous territory, except everyone is excitedly plodding forward without a second thought.

I agree. This work is incredibly cool scientifically but this seems too dangerous to pursue further

- Said everyone who saw the first computers, planes, automobiles, trains, bikes, phones, telegram, and likely the wheel.

I'm sure everybody who saw the first nuclear bomb said this too, and they were right. I am become death

One could argue nuclear bombs and MAD are the reason there hasn't been a third world war. Someone asked an engineer who designed ICBMs : How do you feel that things you worked on may never actually be used? He said, then I have succeeded. The purpose has always been to act as an deterrent. If my work was used, then somewhere something has failed. Off topic to neuralink, but found it interesting.

Could you elaborate on why this is dangerous and shouldn't be done?

This was already being done way before Musk, for medical reasons. You can't expect people to put society in front of their own health.

Unimpressed, but great work..I can see potential but this is like making a plane for landing on moon we are one decade early on this technology ...Electrode impedance is a big deal that I didn't see addressed here...It hasn't been addressed for the past decade...There are so many impressive work I see like the work by Berkeley professor Jose M. Carmena that could address this issue...next Decade can be exciting.

Elon was pretty clear since beginning that it is nowhere near complete and the demo is to attract talent

Could you elaborate on which aspects of the electrode impedance?

Not only are you going to have to charge your phone, watch, headphones, and glasses, but you're gonna have to remember to charge your brain now too

Well, un-dismissible pop up screens that can even wake you from sleep might help... until it’s used for more than battery charge notifications

Although presumably, given the device is plugged directly into your brain, they can make it remind you of this in friendly way.

> but you're gonna have to remember to charge your brain now too

Well, try not sleeping for a few days :)

touché

"Can I borrow your charger my brain is dead" lol

The wireless charger could be installed in the bed itself, or in the headrest of your car seat.

Can't wait for the daily rush on the cranial induction seats on commuter trains.

I haven't watched the whole thing, did they address the problem of writing? All I saw was detecting firing neurons from a small region of the brain (which is cool and useful in itself) but saw nothing about actually transmitting to the brain signals that the brain can interpret correctly. Which seems by far the hardest part, when the device stops being a pure observer of a tiny region of the brain.

I believe that they did show the "threads" firing, and the neurons lighting up in response.

I imagine that initially they will be focusing on blunter interventions into the brain, like detecting impending epileptic attacks, and send some sort of signal to try and interrupt it. I think such things already exist in a cruder form.

It's going to be a long while, and a lot of iterations before you can expect to just "know" the contents of your latest text message.

What is technological progress without the courage to try. To challenge what is impossible by all modern reason.

The big difference between him and top notch medical experts seems to be this : piles of cash and cutting edge tech across all disciplines thrown at a problem. It's not common that a research proposal gets granted hundreds of millions of funding.

But here is what I've learned from Tesla/SpaceX: Cutting-edge tech changes the initial conditions of a problem. How things are measured, analyzed, manufactured, simulated etc.

The only things that are fundamentally constraining are the laws of nature. As long as you don't violate them, things can be done, potentially (in principle).

That why Musk has been like: OK, I see a feasible trajectory, so let's throw money and talent at the problem and see how it goes. Which by the way is how progress works. Pushing things.

Experts (especially scientists) have acquired their expertise within certain initial conditions. Initial conditions incorporate assumptions that might be outdated after having thrown money and tech at the issue.

That's why Musk doesn't give a damn about conventional wisdom but in my opinion only cares about the fundamental physics when it comes to basic feasibility.

What seems to be different compared to a lot of others seems to be that he is a scientific mind by training and a hardcore engineer by profession. Plus basic economics literacy and tons of money.

You usually find only one or those traits in people.

Tldr : his mantra is 'question your constraints'... Let's see how well that goes with this project

Once again a project that seems to have gone from the drawing table to redesign the prototype. And now well under way !

What I'm the most curious about is how they plan to handle exceptions and failure modes to keep it unharmful.

>> What Dorothy illustrates is that you can put in the Neuralink, remove it, and be healthy, happy and indistinguishable from a normal pig.

If I understand correctly, this is a device that can turn a person into a pig?

Nah, that's the web browser, invented a while ago.

I'm so excited by this and I'm glad that someone like Elon is finally concentrating on agri-tech.

This is great news for pig farmers. Imagine being able to control the amount of exercise they take, to bring them inside automatically and send them back out on a schedule. If they get rustled or lost somehow they could be ordered to walk back to the farm.

I'm sure this will revolutionize farming.

Can't wait to have neural data associated to text and images for multi-modal training of neural nets. Up to present we have had the basic modalities: video, image, audio and text, now we can have a new modality which is more direct than even speech, remains to see how much of the embedding space it can cover.

I would love to be able to control vim entirely through Neuralink. Using vim already feels like I’m coding at the speed of thought, this would literally be the speed of thought. The era of the 100x developer is coming.

Is there some type of calibration that would be needed after implant to allow for variance? Or does the robot surgeon just have perfect accuracy in where the threads are implanted?

I don’t think it matters so much to be super precise on the positioning. The spike detection is pretty robust and then the real heavy lifting is in analyzing the patterns of spikes across the set of 1,000+ locations. And that is automatically calibrated as part of the process of getting it to work at all (that is, getting the predictions to match the reality).

they mentioned you can "record backup and replay memories" and potentially transfer them into a new body

and they mentioned it is gonna be in the future (some completely undefined amount of time lately, so technically it is not a lie)

>transfer

More like, copy to the new body and kill the old one.

https://en.wikipedia.org/wiki/Teletransportation_paradox

Neuralink v1.0 has all the hardware for self-driving.

How do they decide where to place the electrodes?

And how do they keep them fixed in the right position?

- They focus on avoiding blood vessels. It is up to the software to learn the mapping, after implantation. - They talked about flexibility, slack of electrodes, and pig head-butting. :)

which should I learn now assembly or c for neuralink programming? or I don't need to learn anything after surgery I can install all directly to my brain?

Was there any mention of glial scarring anywhere?

How much would this technology cost?

They said goal is to make it as affordable as LASIK

where can I find the source code for neuralink in GitHub?

Neuroscientist here. I have personally implanted many animals of different species with similar apparatus as the ones described by neuralink, and published numerous papers in the field, including theoretical papers on neural activity coding and modeling. Here are my two cents about this. Sorry if it’s long, I m hoping that this will help provide perspective on the topic at hand.

As a scientist, the first thing that is striking is that this is another example of science by press conference. It’s probably why there is a lot of eye rolling in the field. No data has been released (including in the single author elon musk paper published a couple of years ago), and it is therefore impossible to judge the results independently, let alone have a scientific discussion about it.

First off, it should be noted that not much here is completely new. Chronic implantation of multielectrode arrays, with wireless recordings have been around for roughly 15 years. These arrays are able to record and stimulate. This is usually done in mice, in which space is even more limited than the pig. The number of contacts (i.e recording points) of 1000 is also on par with the current state of the art (see, e.g the neuropixel implants). The fact that neuralink managed to reproduce this seemingly from scratch is an impressive feat, but in no way is it novel. Dozens of labs around the world do this every day.

The issues with electrode recordings are to get close enough to each cell without damanging anything. If the contact is too far, the recording is polluted by other neurons, and one has to rely on blind signal separation to make out each cell. As a result in typical experiments, each electrode is manually positioned In the close vicinity of the neuron (think micrometers). This is a challenge with large arrays because each contact is not independant which reduces the yield. This likely happens here.

Another important problem arises with chronic implantantion. When you insert something in the brain, make it as thin as you wish the brain will not like it. There is going to be pushing and shoving of cells, and when the brain reset in position (within minutes) it will have moved. This requires fine tuning of each electrode’s position to keep it close to the neurons. More importantly, it will create a scar. Cells will die around it, the organism will try and isolate the foreign object. This takes weeks. This is the reason why, roughly speaking, no single cell has been recorded from for more than weeks (the record i believe is around 3 months): after a while the electrode is simply too isolated from the actual neurons to pick up a signal. There are endeavors around this, With other electrode materials, etc but neuralink does not do that. (Another really cool way to do chronic implant is with calcium imaging with small implanted confocal microscopes)

In addition, this is likely bad for the organism, and possibly not something you would want happen in your brain. BTW I believe that (in addition to size constraints) this is why they are doing studies on pigs now (although their neural activity is not well documented as opposed to mice/rats or primates): it is commonly used as a model for human injuries.

Finally, do be aware that some humans are already implanted chronically in the brain with electrodes: - For parkinsons, in a specific, non cortical area. These electrodes can record and stimulate to alleviate crises - For hearing, it’s not stricto sensu in the brain (in the cochlear cavities) but it stimulates neurons - On the surface of the brain, with ecog electrodes mostly used during surgery but sometimes chronically for epilepsy diagnosis

In addition to these technical issues (which are serious enough, by the way), the claims made here are just sort of outlandish. Even in very well controlled laboratory conditions, there is very little evidence that it is remotely possible to “record” memories (let alone replay them). For the most part it is because we have _strictly no idea_ how _anything is encoded_ in the brain (there is even raging debate as to whether it “encodes” anything). (Sorry to my neuroscientist friends and colleagues, but we are not close to anything the layperson would qualify as “understanding what the brain does”, by a long stretch.) In addition, some of the less outlandish claims (e.g. controlling robotic limbs, etc) made by neuralink are things on which neuroscientists have been working on for decades. It’s not impossible for sure (as many of us argue in their grant proposals), but it is hard. And the plan presented by neuralink (more electrodes) is in no way innovative to an insider in the field. In fact, I doubt that it would get NIH funding.

TLDR there are many well documented issues with this endeavor that many people have been working on for decades. It’s frustrating that the general public be exposed to the uninteresting sugar coated sci fi version of an otherwise complex, nuanced and thus interesting field of science.

Thanks for this. I know absolutely nothing about neuroscience, but I was very skeptical when he claimed that memories could be saved and replayed. The idea of encoding and the actual mechanism of our memories is really intruiging though.

It’s routine for labs to insert threads into brains? It’s routine for non-rigid, fine threads to be inserted via robot? I’m just being honest, your comment is not very convincing. And it skims over important details like the fact that all electrode solutions before were not sub-dermal.

I did not really say that any of this is routine. Nor did I try to be particularly convincing.

But implantation of ultra flexible and thin electrodes is not new (and it is promising, sure), several groups have been working on « mesh electrodes ». As of now it has not really proved safe enough for chronic use in humans.

The subdermal aspect is also certainly a nice advance, but relatively useless in a lab setting (and impossible in rodents).

The robotization is not really impressive, all electrophysiological apparati are robotized, for obvious reasons: nobody can reproducibly place electrodes with micrometer precision by hand.

But hey, let’s wait for some papers to be out and we’ll judge then.

Ok so I looked it up and it appears that this concept of very thin and light electrodes is basically taking off right now with multiple implementations. In this context, Neuralink is basically competing against other products for a share of the neuro-implant market, which will be big at some point. Why doesn’t your comment mention any of this? Or the relative merits of each competitor? It’s not clear that neuralink was the first or the last to do any of this. Why should their effort to find an electrode material that prevents loss of signal be minimized? Shouldn’t they be encouraged to find a solution like the other people trying to find a solution in academia?

And I know you won’t answer this because nobody ever answers questions that aren’t inflammatory or insulting... but how could information not be encoded in the brain? When you dream, your brain could not possibly be generating the raw sensory signals... the only way to explain lucid dreams is heavy encoding of information and a really powerful guessing system to fill in gaps where signal decoding was weak or didn’t happen.

Sorry if I did not make that clear but I fully agree that their endeavor towards better electrodes should be encouraged, and maybe even saluted when they provide evidence of improvement over the state of the art.

The question of information in the brain, including whether it is encoded is a really hard one. Of course we can easily find some areas of the brain whose response is _correlated_ with aspects of the outside world. But what does it mean? Who (or what) is decoding it, why? Many different metaphora and analogies can be (and are) readily applied to the brain. As of now, it is safe to say that a comprehensive theory of neural processing has not really been able to answer all of these questions. It’s just all very complex.

With a ton of respect for your work and field, the tone I am picking up here is that Neuralink getting too much attention, unfairly, compared to other good and important work.

What would be the downside to another organisation in the field? Putting fairness aside, if it means even a chance someone with a disability can have an improved quality of life within the next 20 years, is this not a good thing?

(throwaway.)

I've been deeply interested in this field for a long time and BCIs specifically. (not to brag, just to explain:) I have a graduate degree in a related neuro field and strong & fast coding abilities.

Working @ Neuralink interests me a lot.

But I've also read the stories about Musk, how he treats employees, the working conditions, etc. [0]

I'm not afraid of hard work but can folks here shed light whether it might be worth it (or not) to go work there?

Are Tesla, SpaceX, etc. engineers generally... happy/satisfied?

[0] https://news.ycombinator.com/item?id=18679715

You're assuming that Elon Musk personally run each company he found as if it's his full time job.

There's only one Musk.

Specifically Neurolink is likely only about 5% of the time. Its about half each for Tesla and SpaceX according to him, more or less depending on what happening.

> Are Tesla, SpaceX, etc. engineers generally... happy/satisfied?

I don't work at either but I have watched plenty of SpaceX product, err, rocket launches, and they seem to frequently live-stream from their Hawthorn CA headquarters with lots of gleeful & cheering employees. My impression is that at least for SpaceX, it's probably both highly demanding and highly rewarding. In another life I'd love to work at a company like that!

There are plenty of jobs with poor working conditions where you don't get to build rockets, cars or BCIs. Not that it makes it ok, but it appears plenty of people are consciously making that tradeoff.

I enjoyed this joke from the stream:

> If you bombard earth with photons long enough, it will emit a Tesla

I think that goes back to a Karpathy quote [1], don't know where he got it from

[1] https://twitter.com/karpathy/status/961307010246492160

They mentioned it was a joke they had heard (presumably this tweet).

Yes it’s one of my favorite thoughts, which is why I pinned it :)

It's a good one! You helping out with the ML at Neuralink at all? Maybe someday Autopilot could learn something from recordings of people's brains while driving :)

Whats the name of the employee sitting next to Elon who said it?

Max Hodak (https://twitter.com/max_hodak)

Oh, didn't realize he was a founder.

Also "we're hydrogen evolved".

Oof. I wish YouTube had more than 2x playback rate. Elon is hard to listen to at less.

  document.getElementsByTagName('video')[0].playbackRate = 3

thanks!

The fact Elon talks about the universe stared as leptons and not hydrogen as many pop sci shows do is a good example of his whole game plan.

Understand things at their lowest level and build.

Disrupt at the lowest level you can.

Why pick leptons? Why not quarks? It's arbitrary.

Elon did say quarks and leptons.

Did you say this because you really hate Elon or are really into the Universes first few second :)

(Elon got it right)

Reply on news.ycombinator.com