UKH

Highlights from the neuralink update talk by Elon Musk.

As I've said before I'm quite into transhumanist ideas and so I'll be watching this closely.

youtube.com/watch?v=vxehbGLoar8&

http://news.sky.com/story/elon-musks-neuralink-puts-computer-chips-in-pigs-...

TWS

It's great to see how quickly this has progressed to animal tests - it looks ready for humans from the technological viewpoint.  

The data compression seems slightly too optimistic to me - for it to really work I think they'll need a lot of training data, potentially from the individual concerned.  This means a higher bandwidth route is needed off the transponder / interface chip; they must be using something with much higher bandwidth for research purposes.  I don't think it matters for pulling basic information out of the brain, but if they're going to eventually find the ghost in the shell...  Arguably they've been too ambitious in going for a Bluetooth/BLE connection from the skull implant instead of a shorter range wireless link to a head mounted transponder with a bigger battery and easier recharging; I'm distinctly uneasy about there being a battery embedded inside the skull compared to wireless power delivery from an external unit - this thing is going to be massively power hungry compared to normal medical implants and having a chemical battery replacing skull bone is an invitation for disaster in case of major head trauma.   Then again, they're going to need a chunky battery in there in the end - so if they start solving the problem now...  Also, Musk has engineering staff work between his companies (the "Starship" prototypes have several Tesla battery packs welded to them to run the various actuators in lieu of the final integrated vehicle fluids hydraulics systems) and all of Tesla, SpaceX and Boring make heavy use of rechargeable battery technology, and Tesla and SpaceX both do so in life-critical systems.  Still, are people going to have to walk around with an emergency chip pulling device in case of a battery fire?    

No new updates on the electrodes - which when revealed 2 years ago seemed really quite impressive compared to current academic research.  I imagine they're doing in-parallel development of next generation electrodes whilst building the interfacing component for the current ones.

The big issue I see is regulatory - the FDA only approve/licence medical products to combat specific diseases or injuries.  Neuralink should be able to get permission to install these in quadriplegics very quickly, because they can shunt nerve signals past spinal damage and restore motor function - I think the technical challenges to this are minimal beyond where they are now, although I don't know about passing sensory information from the limbs back in to the brain so that people can safely use them.  This makes such a massive change to quality of life (and cost of care...), and there is no alternative treatment so licensing should be rapidly forthcoming and I think volunteers will line up - some people in that situation make a one way trip to Dignitas, so if a company comes along as says "We will restore your limb control and pay all your medical bills for the rest of your life" I think they'll have a lot of takers.  Likewise they can follow this approach for deafness and blindness once quadraplegic trials have started to retire risks; this helps them test information flow in to the brain.  But.....  Plugging the internet or an AI into the human brain is Musk's stated intention, and is in no way a cure for a disease (far from it if you look at mobile device addiction and the way people get drawn into idiot-level conspiracy nonsense via the internet...).  So, whilst people with disabilities give him ample opportunity to thoroughly safety test these devices and gather a lot more data - and to do good in the process - there is no regulatory/licensing regime pathway to his end game.  So, he has to lobby, bribe or sue his way to a new class of licence that allows surgical/medical intervention for non-disease reasons.

This is massive, because now we have an individual with a net worth approaching $100 Bn and who is very well connected working to create a new regulatory regime for non-therapeutic medical interventions.   From a transhumanist perspective this opens the door to all sorts of other medical interventions not targeted to a disease from cosmetic gene therapy to pharmacological interventions to slow or reverse the ageing processes.  

Also - there's no FDA on Mars!  "He's having a schizoid embolism!"

In 510 BC my Greek ancestor, the little known  Birdocoles Num Num interfaced his neurons with an abacus. As a result the extraordinary extra computing ability of his already bulging brain gave him an insight into a relationship between the hypotenuse of a right angled triangle and it’s other two sides. That evening, after an urn of wine and a feast of kebabs, one of his so called ‘mates’ in whom he had confided, a bloke called Sid Pythagoras, with cynical intent, plied  him with Metaxa laced with Henbane rendering Birdocoles unconscious for three days. In which time the unscrupulous Sid had snaffled his papers and published the theorem as his own and thus robbing the Num Nums down the centuries of the recognition (and royalties, etc) that they richly deserve.

Elon needs tests run in animal models representative of the human illness first to show some form of device efficacy.  You are right about regulatory challenges, the medical device evidence and standards for testing are for the entire product, which also includes the packaging and labelling. It also includes the evidence for the lifetime of the product which in case of a neural implant will probably be the lifetime of the patient so those three young pigs we are shown (plus the 50 or so more in the back) will need to get significantly older before ISO10993 biological risk assessment will be considered satisfied. And that is if the product now in testing is considered the final version.

 The problem with many of those illnesses you mentioned is the underlying causes varies between people but the outward presentation seems the same or in the case of Alzheimer’s we have no idea what is going on. You could find yourself supercharging the engine to try and fix a broken axel.

Do you think this will extend to the apps people will eventually run on their phones to connect to the device?  People cry “walled garden” at Apple’s App Store approach, but imagine one with a 3-year review process for every minor update - it seems clear a new regulatory regime is needed.  There’s a strong argument that it should be to outlaw all non-therapeutic brain augmentation indefinitely, but then you just end up with enhanced criminals and the dystopian cyberpunk future gets even worse.

I think that compared to the implant, animal model efficacy for a shunt around spinal damage will not be difficult to show - but there is no public information on that; I wouldn’t like to guess at timelines.

This is where there’s a big culture clash coming.

Musk’s normal way of doing things - rapid prototyping, pivoting whole designs without considering sunken cost, having a consumer or industrial product whose design continually evolves literally week to week with improvements, tests and tweaks - clashes with the entirely sensible but long duration requirements for licensing of medical products.  He doesn’t take “no” very graciously (or maturely) but it’s hard to see how the usual processes can be shortcut.

Arguably that’s what the last decade of research into plaques has been doing.  We’ve got some invertebrate models of altzheimers... I think the next decade of research could be a lot more fruitful.  

It's is totally surreal and depressing the way science and engineering are progressing faster than ever and we have AI, neural interfacing, multi-species chimera, travel to Mars and so on starting to beome realistic possibilities while politics is regressing.   

We have the exact same bullsh*t in the handling of corona virus as we had in the 1918 flu epidemic and Trump and to a slightly lesser extent the Brexiteer Tories are following the same path as the fascists before the second world war.  We can't even abolish the f*cking royals or the House of Lords and we are rolling back the institutions that could respond to global challenges e.g. WHO.

It is about time that politics and economics got their finger out and started to evolve and improve to match the technical progress.

The FDA are moving towards accrediting companies rather than individual medical devices, Apple, Samsung and a few others are on a list to be the first to make this happen. As to what they achieve is unknown, I would imagine soft targets such as Apps moving towards Software as a Medical Device (SaaMD) clinical decision support software (e.g. AI based stuff) would be the first to get in under this new regime. High risk devices (e.g. pacemakers and a neural lace) which are mission critical to patient life will be much slower to get uptake.

Lean launch/minimum viable product strategy saves time and money however the argument would be at the patient benefit/ethical end where whipping items in and out of us like an old car is unlikely to be received well by clinicians, insurers and national health services who arguably have the final say on reimbursement or adoption. 

Regarding Alzheimer’s, all the Phase 3 amyloid plaque targeted therapies have failed to demonstrate effectiveness (Biogen, Merck etc). The recent FDA resubmissions are all based on finding people well before onset of clinical symptoms and using high doses so I don’t hold out much for them meeting the cost-effectiveness standards set by NICE or ICER.