[HN Gopher] A proof-of-concept neural brain implant providing sp...
___________________________________________________________________
A proof-of-concept neural brain implant providing speech
Author : LorenDB
Score : 106 points
Date : 2025-06-29 12:16 UTC (3 days ago)
(HTM) web link (arstechnica.com)
(TXT) w3m dump (arstechnica.com)
| yaris wrote:
| I often wonder how such teams build their devices - I assume it
| requires quite a few pieces of equipment that can't be bought at
| a nearby shop. Are such devices ordered from some manufacturer or
| are they built in-house somehow?
| numpad0 wrote:
| University researches in general? The boss always knows a small
| company that can make one, and staples a bidding notice to
| campus notice board for formality as the device is getting made
| and delivered. Isn't that how they do anywhere it might be?
| niemandhier wrote:
| There are specialized companies selling components, what you
| cannot get you manufacture yourself. Assembling the device for
| prototypes like this is usually done in house.
|
| Established labs often have a specialised section that
| cultivates all the little tricks how to do these things.
|
| Knowledge is transferred by hiring postdocs that have the
| skills you need or by sending a phd student over to be trained.
|
| As a scientist, if you do this for a while you end up with
| insane skills, but there is no place for them on the job
| market.
|
| Everybody else is living 15 years in the past from your
| perspective.
| pingou wrote:
| Telepathy is on its way. Next step they just skip the conversion
| of brain signals to words and just directly send the signals to
| another brain. But I think some conversion/translation would
| still be necessary.
| mettamage wrote:
| The year is 2100. The brain of the Eurasian president got
| hacked by the Antartic Federation. While humans have hard coded
| a moral code since birth, there are illegally born babies that
| do not undergo brain modification treatment. Moreover, the
| South Pacific Whale Society has no moral code. We should've
| never implanted this stuff into whales. The world will never be
| the same again.
| serf wrote:
| Ghost in the Shellfish.
|
| I'll see myself out.
| mettamage wrote:
| Ooooh! You're on to something. They are part of the evil
| villain tag team group that's secretly behind all this!
| They call themselves Eel-on Mollusk.
| briandw wrote:
| Like in the book "Accelerando (Singularity)", the lobsters
| that get a protected sentience status
| icoder wrote:
| So Long, and Thanks for All the Krill
| coldtea wrote:
| > _The year is 2100. The brain of the Eurasian president got
| hacked by the Antartic Federation._
|
| They'd stil have a president? They would probably already
| have a dictator that controls everybody through a mind-
| reading police state...
| alluro2 wrote:
| No, that's the Republic of Murica, where Trump The Third is
| a BDFL.
| coldtea wrote:
| Hardly better in Eurasia
| moffkalast wrote:
| The war with Eastasia and Oceania took a toll...
| gpm wrote:
| Lots of dictatorships use democratic language. See Putin,
| the President of the Russian Federation. Or the Democratic
| People's Republic of Korea.
| lambdaone wrote:
| Or they are a hive mind, like Alastair Reynolds'
| Conjoiners.
| thrance wrote:
| I think it would be deserved if whales started bombing our
| streets.
| Cthulhu_ wrote:
| I think (futurology / science fiction) that they will make some
| kind of brain link, but there won't be any translations
| happening in between, just raw brain signals from one to the
| other, like an extra sensory input; there won't be any encoding
| or data that can be translated to speech or images, but the
| connected brains will be able to learn to comprehend and send
| the signals to / from each other and learn to communicate that
| way.
| falcor84 wrote:
| That sort of connection would be very susceptible to psychic
| attacks - I'm thinking of the telepaths in Babylon 5, being
| trained for offensive capabilities, as well as just plain old
| spam advertising. So while "defaulting to trust" is often
| considered societally useful, I believe that it would be
| better for everyone if cross-brain messages are sent in a
| format that can be analyzed (and entirely blocked) by a
| filter on the receiving side.
| z3t4 wrote:
| We are so different, but I guess with a lot of training we
| could interpret each others thoughts. A first step would be
| to record your own thoughts and then replay them in order to
| see if you experience the same thing you did when the
| thoughts where recorded. It's possible that our brain is
| constantly re-configuring so that even your own recorded
| thoughts would make no sense.
| voidUpdate wrote:
| I think the main problem with that is that different people
| thing in different ways. I think in full sentences and 3D
| images, whereas other people might think without images at
| all. How do you translate that?
| yieldcrv wrote:
| If statements
|
| for how the brain chip chooses to function
| suspended_state wrote:
| It is very likely that this device works by perceiving and
| interpreting brain waves. Actually, from the article:
|
| > "We recorded neural activities from single neurons, which
| is the highest resolution of information we can get from
| our brain," Wairagkar says. The signal registered by the
| electrodes was then sent to an AI algorithm called a neural
| decoder that deciphered those signals and extracted speech
| features such as pitch or voicing.
| hearsathought wrote:
| I still fail to see how that's possible since it is assumed
| every brain "encodes" data uniquely. Communication between
| computers is possible because we have agreed upon standards.
| If every computer encoded characters differently, no
| communication would be possible. Without agreed upon ports or
| agreed upon mechanism to agree upon ports one computer could
| not communicate with another. So how can brain-to-brain
| communication work given that encoding/communication
| "standards" are impossible since each brain is different?
|
| For example, I see a tree and my brain generates a unique
| signal/encoding/storage representing the tree. Another person
| sees the tree and generates a unique signal/encoding/storage
| representing the tree. How would my brain communicate "tree"
| to his brain since both our "trees" are unique to our brains?
|
| My brain device reads my brain signal "1010101" for tree. My
| friend's device reads brain signal "1011101" for tree. How
| could we possibly map 1010101 to 1011101. Or is the
| assumption that human brains have identical signals/encoding
| for each thought.
| goopypoop wrote:
| I already learned to interpret touch, taste, smision etc.
| when I was just a baby. How hard can a new one be?
| lambdaone wrote:
| There would probably be a Universal Common Embedding used as
| an intermediate representation between people's individual
| private neural representations. Likely the distant descendant
| of our open-source neural models.
|
| And machines would of course also use the Universal Common
| Embedding to communicate, as man and machine meld into a
| seamless distributed whole.
|
| It all seems a little bit too inevitable for my liking at
| this point.
| msgodel wrote:
| You could skip a lot of stuff by training a transformer on some
| kind of neural embeddings. You probably get effective FTL
| communication and limited immortality that way.
| falcor84 wrote:
| How did you arrive at FTL? At best we'll have comms at radio
| transmission speed, right?
| msgodel wrote:
| It's faster than light because your model is already there
| so no information has to be exchanged.
| throwaway34564 wrote:
| very cool concept - kind of like an offline inference
| model of your conscious. You could have periodic (over
| the wire) sync updates to your real self
| falcor84 wrote:
| But isn't it the same sort of FTL and "immortality" that
| a creator gets from us listening to a downloaded podcast
| that they prerecorded?
|
| To quote Woody Allen:
|
| > "I don't want to achieve immortality through my work; I
| want to achieve immortality through not dying. I don't
| want to live on in the hearts of my countrymen; I want to
| live on in my apartment."
| msgodel wrote:
| Seems like a technicality tbh.
| goopypoop wrote:
| I look forward to learning that your favourite fruit is
| AAAARRGGHHHH
| dotancohen wrote:
| Maybe he was dictating?
| coldtea wrote:
| Next step: techno-slavery
| midtake wrote:
| Like the trisolarans
| spinlock_ wrote:
| Don't forget the injection of ads for the basic subscription
| plan, Black Mirror S7/E1 vibes.
| Lapsa wrote:
| It's already here. And likely you are deliberately trying to
| deny that. Here's a nice little rabbit hole for you:
| https://pastebin.com/raw/8TQyPKUF
| goopypoop wrote:
| the scary thing isn't the sci-fi in this stuff, it's that the
| tech already exists to do the polygraph scam again
| seydor wrote:
| Not without ads being injected. But it's a small price to pay
| for such amazing capabilities
| germinalphrase wrote:
| Telepathy, or maybe memory and experience sharing media
| machine, or maybe humanity id unification device, or maybe
| flesh robot actualizer, or maybe a looping torture horror show
| mask. The possibilities are endless!
| dylan604 wrote:
| would you need an antenna for your brain implant to tx/rx those
| signals? what shape would that antenna need to be? would it
| need to be a rigid antenna that would have to stick up out of
| one's head? would it be a curly pigtail style or straight semi-
| rigid yet flexible so it could bend when passing through doors
| like the CB antennas of yesteryear? Could it be very flexible
| like a cable so you could run it down your back and under your
| clothing?
|
| would your antenna be susceptible to crosstalk, and would that
| interference come across as new voices in your head? in fact, i
| wonder what the signalling protocol would be so that the
| message is only be received by the intended recipient,
| something like wifi? to that, would someone be able to tune
| into the spectrum around them and see the metadata of these
| telepathic signals to see who was talking to whom, when, and
| for how long. obviously, i'm assuming the actual signals will
| be encrypted so that these conversations will be private. will
| the NSA be able to pick up these signals from their satellites
| and be able to listen in? or will the gov't force backdoors
| into these communications?
|
| i know this might sound conspiratorial, but all of these are
| valid concerns being dealt with now, and only sound
| conspiratorial if you've have your head in the sand. these are
| also questions that startups tend to ignore. look at the IoT
| market that never considered any potential security issues, and
| now we have massive bot farms. i know i wouldn't want to be
| used as a bot because my neural implant manufacture never
| considered what a hacker might do once they gained access
| andrybak wrote:
| > In this second test, the word error rate was 43.75 percent,
| meaning participants identified a bit more than half of the
| recorded words correctly.
|
| > [...]
|
| > "We're not at the point where it could be used in open-ended
| conversations. I think of this as a proof of concept," [Sergey
| Stavisky, a neuroscientist at UC Davis and a senior author of the
| study] says.
|
| The ability to produce sound without a use of a dictionary sounds
| awesome. It is an interesting result, a proof of concept as the
| author of the study says, but the title is editorialized at best
| and effectively clickbait at worst, because most readers will
| assume that "near instantaneous speech" means "clear intelligible
| speech and ability to communicate".
| dang wrote:
| Ok, I've taken "near instantaneous" out of the title and put
| "proof of concept" in there, which is a phrase used by one of
| the researchers in the article.
| aitchnyu wrote:
| Cant wait for a man-choker that executes "I'm having costlier
| rice, check if glucose spike is lower than usual rice". Yes, both
| devices are outside my body.
| stephenlf wrote:
| I don't understand what this means. Did you use a translate
| app?
| TheCapeGreek wrote:
| I think it's perfectly legible?
|
| man-choker would just be a choker (you know, the accessory
| usually for women) with some tech on it, in this case to
| accept a command to check with another bio-device if the
| glucose spike of more expensive rice is better or worse than
| cheap rice.
| shawabawa3 wrote:
| it's legible but a very confusing sentence
|
| man-choker is not a word, and choker is a niche garment,
| why not "necklace" or just "wearable"?
|
| Then it "executes" a question?
|
| and then there's a reference to "both devices" - what
| devices?
|
| You need to put together a lot of context clues and
| assumptions to get to: They are probably a diabetic with a
| glucose monitor and pump, and they want a smart device to
| analyse the data with natural language (but again, why a
| choker specifically? Wouldn't a smart watch or something
| make more sense?)
| dylan604 wrote:
| > or just "wearable"?
|
| wearable where? A choker is not some niche garment. It is
| a well understood accessory that maybe not everyone
| wears, but it's not because it is not well known.
| Describing it as a man-choker while a made up word, get
| over it that's how language evolves, is very descriptive.
| Calling it a choker means that it will be around the
| neck. Wearable could be a watch, and that's not what was
| meant. Using the word choker explicitly tells you where
| it was proposed to be worn.
|
| > and then there's a reference to "both devices" - what
| devices?
|
| one is the speech device, the other the glucose monitor.
| separate devices, but worn on the choker which is now
| becoming a tool belt. But I'm guessing you'd have a
| problem if it was described as a tool/utility belt worn
| around the neck like a choker???
| LastTrain wrote:
| Apparently GP prefers to be more playful with their
| language.
| hombre_fatal wrote:
| Yeah, I find it weirder for someone to claim that it's
| perfectly legible.
|
| The CGM example was so out of place that I thought they
| meant to post in yesterday's CGM thread. And I spent a
| moment wondering why "costlier rice" was qualified. Maybe
| they are hoping their more expensive yuppy rice is
| absorbed more slowly?
|
| And man-choker sounds like a word that Frank Herbert made
| up for Dune.
|
| The sibling reply to mine is getting lost in the weeds:
| you can admit that something is written in a confusing
| way while also being able to understand what they meant
| or even appreciating it. And telling someone to "Get over
| it" sounds like you aren't tracking what the convo is
| about.
| patmorgan23 wrote:
| Massachusetts' New England merch only names Irish citites.
| Is also legible but equally incoherent.
| x187463 wrote:
| Curious how much intentionality is required from the user to
| produce sounds. It would be unfortunate if this device just
| started firing off speech for what would otherwise be thoughts
| one would not say out loud. I suppose that depends on the
| mechanism required to activate the neurons to which the device is
| connected.
| dylan604 wrote:
| I hope it's something better than "Hey Siri, say..."
|
| Otherwise, yeah, that would be a new sort of hell where you had
| no private inner monologue
| connicpu wrote:
| The brain is incredibly adaptive, I guarantee eventually it
| would learn to avoid firing the neurons the device is probing
| when you don't want your inner monologue spoken aloud as long
| as there's a feedback loop where you experience negative
| emotions when something you didn't want spoken aloud was
| broadcast.
| Winsaucerer wrote:
| Without having RTFA, I'd guess/predict that it will be possible
| to learn to only do this intentionally, much like we can think
| about raising our arm without actually raising it.
| AnotherGoodName wrote:
| Even without this device there's been some consideration to the
| thought that the conscious brain is merely an observer since it
| appears to activate after the unconscious brain takes actions.
| You just go along with what the unconscious mind did in actions
| and speech and you convince yourself you meant to do that after
| the fact.
|
| https://pmc.ncbi.nlm.nih.gov/articles/PMC3746176/
|
| So here it could indeed just fire off speech and you know what?
| We'd probably convince ourselves that we absolutely meant to do
| that. In fact it could be a very interesting experiment (with
| willing participants). Mess with the inputs the device receives
| so it's not really the person activating it, let it do it's
| thing and see if they notice when they do/don't have control of
| it.
| stronglikedan wrote:
| Funny, I was just pondering this last night - how I often
| realize this phenomenon immediately after it occurs, and feel
| helpless when I realize it. It's like two minds, one that is
| moving my hands to pile mangoes into my face, as the other is
| telling me to save some mangoes for later (they're very hard
| to stop eating). But on a more serious note, now that I
| realize I realize it, I realize it _a lot_.
| adzm wrote:
| I once got too enthusiastic about eating mangoes and my
| face learned that the skin contains urushiol like poison
| ivy.
| lambdaone wrote:
| I find this simultaneously fascinating and disturbing; once
| you have someone using this, they become a hitherto-
| impossible human-AI hybrid, where their mind is now a fusion
| between the two, completely unnoticed to the user.
| EEBio wrote:
| It's intentional and requires quite a lot of focus.
|
| The original paper [0] mentions electrodes are placed over
| Broca's area (speech production, translates words to mouth
| movements) and motor area (adjusts the mouth movements). It's
| attempted speech, not thoughts.
|
| There is a lot of fear in mainstream media and populace of
| devices decoding thoughts, but that is a significantly harder
| problem, at this moment on the level of sci-fi of Civilisation
| Type II on Kardashev scale. There is a reason why the
| electrodes are not over Wernicke's area instead (language
| comprehension and production).
|
| 0: https://doi.org/10.1038/s41586-023-06377-x
| N_Lens wrote:
| The main challenge appears to be the neural-computer interface -
| the electrodes. As the article states, there are several startups
| in this space all bottlenecked by the same constraint, and
| accurately translating neural impulses into digital (Or even
| analog) signals is the key to unlocking a whole arena of
| transhuman development.
|
| Most such startups are scaling up the number of electrodes
| interfacing with the neurons to overcome this bottleneck, but I
| wonder if an unconventional approach could overcome the limit
| more gracefully. I may be a dreamer, but a high fidelity
| synthetic neural fiber is the holy grail here. I do remember
| reading people partially healed of paralysis due to spinal
| injury, because of electrical conduits that bridged the injured
| neural gap.
| swader999 wrote:
| "a high fidelity synthetic neural fiber is the holy grail here"
| I'm fairly confident there's nothing Holy about this.
| spauldo wrote:
| Depends if you worship a god that encourages mankind to reach
| its full potential instead of one that's strangely obsessed
| with people not doing butt stuff.
___________________________________________________________________
(page generated 2025-07-02 23:01 UTC)