Neuralink is 99% hype


A couple of years ago, Elon Musk bought a company called Neuralink, which is trying to build a bigger, better brain-machine interface. The hype was incredible. Here’s a small sample (note: the author confuses a concept called a “neural lace” from Ian Banks’ science fiction novels with Musk’s Neuralink over and over again, which tells you how unreal every thing is):

As an immediate application, Neural Lace could potentially help patients suffering from brain injuries and certain illnesses. However, the utimate goal and mission of Neuralink are to successfully merge the human brain with machine, fusing human intelligence with Artificial Intelligence. As a result, this is expected to bring humanity up to a higher level of cognitive reasoning.

At some point, Neural Lace is going to enable humans to upload and download information directly from a computer. Just in a similar way how Neo from the Matrix does in order to learn new skills and acquire new information.

In order to insert Neural Lace, a tiny needle containing the rolled up mesh is placed inside the skull. The mesh is then injected and unveiled upon injection, encompassing the brain.

The Neural Lace integrates itself with the human brain. It creates a perfect symbiosis between human and machine. This technology could be the catalyst for the technological Singularity.

Nope. None of that is true. It serves Musk’s interest to have these absurd claims floated about. I wrote about this nonsense at the time. I pointed out the reality then: “Elon Musk has bought a company, and is cunningly trying to inflate its value by drowning the curious in glurge, techno-mysticism, and making shit up, which, because he has this mystique among young male engineers, will probably succeed in making him more money and fame.” That’s what it’s really about, not science, not cool biomedical engineering, not even impractical wish-fulfillment. It’s a game for Elon Musk to pump up his ego and pretend to be cutting-edge in yet another thing, while skimming off lots of money.

Also, he’s got this bizarre paranoia about artificial intelligence, which he doesn’t understand either, and he thinks this is a way to combat the existential risk of the robots taking over.

Anyway, I bring this up again not because Musk has done anything useful with his hype machine, but because a pretty good video came out explaining the details.

She also makes the very good point that this is not gadgetry to benefit the masses, but to make the rich richer and widen the divide between the haves and have-nots, if it worked. Which it doesn’t.

Comments

  1. Dunc says

    Yeah, and it’s not like there aren’t some potential issues with the technology, even in The Culture…

    It was a little bundle of what looked like thin, glisteningly blue threads, lying in a shallow bowl; a net, like something you’d put on the end of a stick and go fishing for little fish in a stream. She tried to pick it up; it was impossibly slinky and the material slipped through her fingers like oil; the holes in the net were just too small to put a finger-tip through. Eventually she had to tip the bowl up and pour the blue mesh into her palm. It was very light. Something about it stirred a vague memory in her, but she couldn’t recall what it was. She asked the ship what it was, via her neural lace.

    ~ That is a neural lace, it informed her. ~ A more exquisite and economical method of torturing creatures such as yourself has yet to be invented.

    [Excesssion, Iain M. Banks]

  2. slithey tove (twas brillig (stevem)) says

    Yes, I once thought it would be really neat to have a watch built into my brain so I would instantly know the exact time without worrying about damaging the watch on my wrist. This grew into having built-in access to the internet so I could tap into the so-called collective consciousness of millions of people to answer questions that pop up sporadically. Yes I knew this was all fantasy.

    The point she made about how this will actually fractionalize us into Haves_v_HaveNots, being less able to confront the menace hypothesized by Musk, [fractionalizing is the divide and conquer strategy the GOP is waging against the Democrats to ensure their fascist oligarch remains in office] This makes me think Elon is in a superhero mindset were a global bad guy (AI) is threatening the world, and a single superhero rescues the world by defeating the bad guy with his overwhelming super technology, regardless how expensive it is. EG Tony Stark, the billionaire, had enough money and industry to develop his Iron Man suit, and be the hero against Theron, (with a little help from a team).

    Is being constantly connected to the Internet really a good thing when everyone has a smartphone that can get there easily?. Isn’t disconnecting a thing now, where people seek vacation spots with NO Wifi, and NO cell phone coverage, in order to disconnect from the overwhelming firehose of, information and disinformation, that can be difficult to distinguish which is which?

    BMI is only good for certain prosthetics such as replacement hands, ears, or vision. The Internet is not a valid cognitive prosthetic, it is only an assistant, not a replacement or enhancement.

  3. steveht says

    Many many great points here. My biggest GAH moment was Musk saying they were pushing hard to find human subjects, when as near as I can tell, the “published” work (doi: https://doi.org/10.1101/703801) includes data from just 2 rats, and has not yet been peer reviewed.

    I know that’s how Musk does tech development, but there’s no excuse for that approach in invasive biomedical applications.

  4. mailliw says

    I think most of what Neuralink is doing comes very firmly under what neuroscientist Gina Rippon in her book The Gendered Brain describes as neurotrash.

    We simply don’t know enough about the brain yet to come anywhere near to what Musk is promising in the foreseeable future.

    The supposed threat of machines acquiring superhuman intelligence is profoundly unlikely in the lifetime of anyone alive today. What is notable about current “neural networks” and “machine learning” is how woefully ineffective they are in comparison with human beings. Rodney Brooks is well worth reading for a realistic and well informed view of the current state of play in AI research: https://rodneybrooks.com/the-seven-deadly-sins-of-predicting-the-future-of-ai/.

    I used to work for a company that among other things was starting to develop software to assist doctors. They were very keen to avoid this software being classified as a medical device, because it would have to undergo far more stringent tests before being allowed into doctors’ surgeries. I think attempting to perform brain surgery on someone to implant a BMI falls very firmly under the classification of a medical procedure and should obviously be very strictly regulated.

  5. wzrd1 says

    I remember thousands of vaporware projects like this evaporating in the dot-com bubble collapse of 2000 – 2004. Looks like some haven’t learned from that lesson.

  6. says

    People behind projects like this really seem to have a really naïve understand of biology. They seem to think the electronics are the hard part.

  7. garysturgess says

    Human level AI isn’t scary. There are 7 billion human level intelligences already. I believe the scary bit (to the extent that any of it is scary, which is doubtful) is the potential for exponentially increasing that human level intelligence to super human intelligence and beyond.

    So, fair enough – let’s take Musk’s fears as valid. How does constant access to the Internet slightly faster than having to type on my smart phone help us reach super human levels of intelligence?

    Even if you take them at their word, this solution offers nothing but snake oil.

  8. John Morales says

    garysturgess:

    Human level AI isn’t scary.

    Ah, that hitherto unattainable goal (you mean AGI, but anyway; note AI is already superhuman in its particular limited domains).
    There is no such thing, nor is it in any immediate horizon.

    But sure, not scary to you.

    Thing is, for at least some people, those who think utterly unlike them are in fact scary — e.g. sociopaths — and I don’t think the case that human intelligence is the upper limit is a credible contention.

    Back whenever the movie Star Wars came out, I remember snickering when mere humans sat in the swivel chairs and clumsily (luckily) scored a hit.

    Even if you take them at their word, this solution offers nothing but snake oil.

    How so? Current AIs ain’t conscious, but still superhuman at particular tasks.

    (That potential ain’t snake oil)

  9. garysturgess says

    John Morales@8:

    Ah, that hitherto unattainable goal (you mean AGI, but anyway; note AI is already superhuman in its particular limited domains).

    Right, but that’s not what Musk says, “Human level AI” is his words – remember, I’m “taking them at their word” here. While not by any means an AI specialist, I do have a BSc in computer science, and I’m well aware of the CS use of the term. I submit that Elon Musk, if he has ever heard actual computer scientists use the term, has subsequently ignored/forgotten what they said.

    How so? Current AIs ain’t conscious, but still superhuman at particular tasks.

    (That potential ain’t snake oil)

    Because being able to access the internet slightly faster will not help us compete against superhuman AGIs.

    Now, that said, I am not suggesting such a capability would be entirely useless. The obvious accessiblity benefits are worth some research in and of itself – but as the video points out, such things are not exactly new and Neural Link is not a particularly noteworthy example of such.

  10. John Morales says

    Hm. One more thing. Human intelligence has some probability distribution, but pretty obviously the median is not that close to the tail.

    (What would be scary about a world of Einsteins? But then, he was benevolent)

  11. garysturgess says

    John Morales@10:

    What would be scary about a world of Einsteins? But then, he was benevolent.

    I again reiterate: slightly faster internet access is not going to help us against a world of malevolent Einsteins/Hawkings/Newtons either.

  12. John Morales says

    garysturgess,

    “Human level AI” is his words – remember, I’m “taking them at their word” here.

    Yeah, but he refers to artificial autonomous sapient entities thereby, which is not a thing.

    I submit that Elon Musk, if he has ever heard actual computer scientists use the term, has subsequently ignored/forgotten what they said.

    He’s a pretty smart cookie, so I doubt that.
    Perhaps he may be referring to automated systems that are as capable as a typical person at particular inferential tasks; if that were so, not that hard to set a billion “watchers” to watch a billion people on an ongoing basis, and infer their degree of loyalty to the system thereby. Or perhaps he’s just guruing publicity.

    (Anyway, Roko’s Basilisk is scary, if you buy into it. But it’s just plain silly, ain’t it?)

    Because being able to access the internet slightly faster will not help us compete against superhuman AGIs.

    Gotta think like a visionary (I’m not one); it’s not about “being able to access the internet slightly faster”, it’s about expanding our sensorium. It’s about not needing an interface to access the net.

  13. John Morales says

    Also, re the video, I had to fast forward/back till I went past the promo (a goodly bit of its content), then had to cope with the multiple stitched-together snippets.

    Slightly irritating, I prefer speakers who can just say what they want to say without stitching together a myriad takes (e.g. Lindybeige).

    (But still, that’s about the presentation, not the content, which is fine)