self_made_human
amaratvaṃ prāpnuhi, athavā yatamāno mṛtyum āpnuhi
I'm a transhumanist doctor. In a better world, I wouldn't need to add that as a qualifier to plain old "doctor". It would be taken as granted for someone in the profession of saving lives.
At any rate, I intend to live forever or die trying. See you at Heat Death!
Friends:
A friend to everyone is a friend to no one.
User ID: 454
I'd be glad to be done with British pennies. All they do is accumulate in my wallet, God knows you can't spend them if you want to most of the time.
Surely, if optimizing for dollars, your effort is better put into advancing your medical career.
Sadly the NHS mostly operates on assigned hours put in = seniority = more money paradigm.
As a trainee, I get predictable and fixed yearly increments, which are nice enough, especially since I could get away with spending less than half my monthly salary if I wasn't so fond of takeout.
Speaking broadly, in most other places, doctors, especially in surgical fields, can double or triple their salaries if they put in the hours. This is much harder to do if you're an early-stages trainee. Completion of training would take about 6 more years, and that's when the opportunity to work private and make bank begins.
I've also missed the glory days when locum roles weren't grossly over subscribed, and you could rely on picking up an odd 3 or 4 shifts a month to almost match your base pay. There are too many underemployed fresh grads and international medical graduates like yours truly clawing at the door.
I agree that's sensible in a vacuum, but I sincerely doubt I'll be growing old and retiring the old-fashioned way. I'm willing to put my money where my mouth is in that regard.
(I really should opt out of the NHS pension, like hell that's going to be relevant in the thirty or so years till I could plausibly retire)
If it's 2030 and AI progress seems stalled at levels not significantly more advanced than it stands today, I could be persuaded otherwise. But for now, I'm more concerned with preparing for a singularity (with a decent risk appetite as I have little to lose) as opposed to saving for retirement.
If I were to attempt to "beat the market", it would likely take the form of buying tech stocks.* And I wouldn't put more in than I could afford to lose! I doubt I'm going to be forced to liquidate anytime soon, so I would have little issue with holding a dip.
*You can't really price-in a technological singularity, and unless everything goes sideways, I expect broad slices of the market to soar and not just MANGA.
On the topic of renting or buying, the only real reason I'd buy is if I was convinced I wanted to tie myself down geographically (and had a family), especially since you point out that re-investing the difference between rent and home loans would obviate most of the difference. It's too early to make that call, and I prefer my money relatively liquid.
Thanks! I think I dimly recall someone mentioning them, but it had entirely slipped my mind before you reminded me.
A request for financial advice:
For the majority of my career, I've made such a small pittance in Western terms that I had no strong impetus or incentive to save any of it.
While the UK is certainly not as wealthy as the US, and doctor's wages are comparatively meager (and have been devalued considerably by inflation), I have a respectable sum already in my bank account. Part of this is that I simply don't have a life outside of work, and couldn't spend the money if I wanted to.
I'd like to invest the majority of it, around 10k GBP. My risk appetite is decent, but I also think the market is more efficient than I am in most regards (when I had forewarning that Nvidia would skyrocket, I didn't have any money to bet with). I don't want to see it get eaten away by inflation.
I believe the default course of action would be to buy index or mutual funds. I've been eyeing the S&P500.
I do intend to hold on to a few thousand GBP for expenses like buying a second hand car, expensive exams and the like. In the longer term, I'd need ~40k for a mortgage, assuming I choose to continue working and training in the NHS when my contract is up.
When I overheard another pair of doctors telling each other "SMH's dad is a good surgeon, but he just can't keep up the speed he used to. You should have seen the way he operated 15 years ago." Or when I saw him reduce his working hours/caseload from 'insane' to 'more than I would personally care for' as he just couldn't physically stand around for >12 hours a day.
Human fluid intelligence declines from 25 onwards, but your accumulated skills and knowledge compensate till your 40s. After that, your body and brain will give out eventually. Thankfully my parents have their wits about them, and being doctors, are capable of attending to their own health.
More common? I believe so.
Even then, I've heard plenty of people who read Xianxia complaining about how all the MCs end up obsessing with the Sword Dao. The problem is almost as bad, halberds/guandaos get very little screen time.
No fantasy elements? Well, if you studiously ignore that most medieval potions didn't do anything I guess.
I think we're at like 300 gb on the very largest games? I doubt we'll hit 1 tb in at least the next 2 or 3 years, and we'll likely see game sizes drop or stabilize as neural textures or even full-on neural rendering takes off.
Fair enough, though I will reserve my doubts.
I've only read (most of) Worm.
But funny you should mention that, because I write a novel where cyborgs are a mainstay (the protagonist is humanoid, but only because he hasn't been pushed further) , and the upcoming chapter has one who is basically a pair of frontal lobes in a crab-shaped shell.
You'll see clear Worm inspiration in my work, though I aim for much more of what I perceive as 'realism' in terms of societal and governmental reaction than Worm does with its desire to have the protagonists punch people on the streets. (I'm aware of in-universe justifications, I find them lacking)
Wildbow doesn't get nearly as wild as I do.
Our disagreement rests on all those ifs. I have far more senses than I have attention, and my attention is very limited and therefore precious. I spend more time trying to minimize sensory input than augment it. I'm not just talking about earplugs and blindfolds for when I try to rest. Like, filtering out background noise when I talk to someone. Ignoring visual distractions when I read.
Funny you should say that, as I was just studying for a psychiatry exam, and finished reading a few chapters on human memory and attention.
The 'frame buffer' for raw sensory input is OOMs larger than what's brought to your conscious attention. If you're sitting on a chair, mechanoreceptors are constantly sending signals upstream, but only salient information, presumably the text you're reading right now, is magnified and focused.
I strongly expect that additional senses will, while distracting initially, fade into the background until salient, no more of a nuisance than your proprioception of your legs interferes with your ability to read.
Do I really want to add ultrasound sense? Why, what am I going to do with that information? And do I need that info with continual or immediate access, all the time, to justify an implant?
By the way, you can totally do that with current technology: take a hearing aid, set it to receive ultrasound. You'd still need to use some of your actual senses for receiving the input, like taking those ultrasound waves and translating them down to normal hearing range. That will unfortunately interfere with you hearing the usual sounds, and if you don't want that, you can use some of your less-used senses. Like, have it be a vibrating butt-plug or something. I'm sure one can train the brain to distinguish different vibration pitches after a while.
That is the key difference between a mature BCI and most other prosthetics. If you use neural connections, you avoid the issue of having to compress or wrest control of existing sensory bandwidth. You shouldn't settle for ultrasound pitched down to be audible, you should be able to hear both. I strongly expect that in actual usage, bandwidth won't be an issue, or have negligible impacts.
And of course, if you don't see the utility in something, don't install it into your body. I have nothing against people who are happy with their existing bodies and minds, I just desire otherwise for myself.
It's a given that we live in environments that are amenable to the barely upgraded baseline human form, since we adapted them to match our needs.
Speaking broadly, humans already go through intensive cognitive and behavioral modification for the purposes of a job. That's called school, college and uni! In physical terms, most jobs require us to either personally move a ton or two of steel and plastic, or ride as passengers in one. You might need glasses to correct vision, or more rarely cosmetic procedures if the job implicitly demands it.
I doubt humans will be forced to make ourselves into cyborgs for the purposes of labor, but only because even with these augmentations we would not be cost competitive with AI systems running industrial robots. I've elaborated further downthread on why I think hoping for humans to keep up with the machines is a forlorn hope. Imagine being asked to improve a monkey to the point it can be an accountant. By the time you're done, it's not really a monkey anymore, and in all likelihood the additional hardware you need to make a normal monkey fit for the job would be capable of doing it by themselves.
That being said, I am a transhumanist, and I will eagerly embrace cybernetics where the benefits outweigh the risk, or simply for the sake of improving my body. Getting legs that let me run faster than Usain Bolt won't make me dispense with a car, but I think they'd be handy. A good BCI would make most portable electronics like phones or laptops redundant, assuming you were happy using it as a wireless link to some kind of computational hardware. I imagine that if you want your compute close at hand, you'd carry around something the size of a USB powerbank in your pocket, and potentially even just to keep your internal hardware charged up.
I am an unaware of any large body of psychiatrists considering dyslexia not a real illness. Nobody I know in my professional life has voiced such an opinion either.
I'm a longsword/Zweihander or halberd fanboy. The latter, like all polearms, is grossly underrepresented in media such as video games or movies. How often do you see a protagonist using a pike or halberd instead of a sword?
This assumes a degree of significant personal protection with armor. If that was lacking, I'd definitely opt for a shield.
Sadly most representations of combat utterly gloss over how effective even basic armor was, a medieval knight was a tank that was almost immune to sharp damage while not tired, and only really countered with blunt force trauma.
You mistake my presentation of examples of the brain being highly plastic in regards to new sensory modalities as a claim that future advances will be nearly as crude.
The number of people with adequate vision who would be happy holding a lollipop in their mouth for the purposes of redundant visual input is rather minimal.
But what if you could have minimal LIDAR embedded on you? They're small enough to throw into an iPhone. Then, with your eyes closed, you have a sensory perception of everything within your proximity. Behind, above, below. It doesn't matter. That is a strict improvement over normal vision.
Or the ability to hear infrasound and ultrasound. You might hear machinery failing or an impending earthquake before your mundane senses catch up.
Thermal senses? You'll know if your coffee is hot, whether food is done, whether that saucepan is safe to take off the stove. Whether an industrial accident is safe for humans to approach, or if a firefighter can take the risk of opening a door while in the confines of a flame-retardant suit.
I might be able to tell a patient was coming down with a fever just by glancing at them.
What about a BrainPort device that's geared towards infra-red? Wouldn't that be cool, see the world like Predator? Or, again... why not just put on some infra-red goggles?
To avoid the inconvenience of standard infra-red goggles or night vision devices. When you put them on, you're sacrificing standard vision in the process. Of course, for a soldier or hunter at night, their normal vision was already inadequate.
At any rate, there are oodles of useful information in the environment that humans don't have access to, but we can observe animals benefiting from. Magnetoception, or an internal sense for GPS, and you'll never get lost again.
Its obvious that humans can already do most of these things through utility devices. A BCI makes that connection more seamless, with lower latency, with reduced cognitive overhead from translation into the sensory modalities we are used to handling.
Eventually, our environment will shift to accommodate this, as the modern world has rapidly pivoted to taking things like automobiles, smartphones and omnipresent internet for granted. Someone in 1890 was not suffering because he didn't have the non-existent internet. You would, without it.
Eventually, cybernetics will be additive and not subtractive or substitutes. Right now, a cybernetic eye is only useful to someone who has visual issues (NVGs are examples of visual augments, though a purist would say they need to be directly hooked up to your brain, or else a car is a prosthetic leg).
If an augment seems useless or not worth the hassle, don't use it! You might not need magnetoception, but a hiker in an area with spotty signal or GPS jamming might. You might not need in-built LIDAR, but a soldier afraid of FPV drones or someone working in close proximity to industrial machinery might benefit.
BCIs just hold the promise to liberate us from interfacing with our tech through sight, sound, touch and so on.
I disagree. It seems generally the case that you need at least one eye to be reasonably competent at catching people with eyes and then removing them. Continue this cycle:
The rest of the world would be blind, but there would be a nice fellow with at least one who reigned as king (or queen).
Add a looming colossus that dims the sun in the background (to represent ASI), and there you have it.
(Or maybe the sun is dimming as its covered by a Dyson Swarm. Who knows, it's the Singularity baby).
And why would you have to sacrifice any of that? They're not cutting out chunks of your brain to make a neuralink fit, the skull, while compact and rather packed, can fit a neat little cap of microelectrodes without much issue. We've stuck far larger things in there, see how massive Utah Arrays are in comparison.
In an able-bodied person, a BCI would be a significant augment. You would be able to control pretty much any electronic device at the speed of thought, bandwidth allowing.*
You would, if feedback was implemented (and there is no reason to think that doesn't work, because we have that already), be able to receive nearly arbitrary input too. Do you want to perceive the magnetic field around you? No biggie. Do you want to smell wifi signals? Why not?
I can hardly stress how life-changing being able to interact with digital domains at the speed of pure thought would be. No more typing, to say the least.
And since electromagnetic radiation can jump distances significantly larger than the few microns separating neuronal junctions, you would be able to control and sense just about anything, just about anywhere that latency allows.
If you think the brain can't handle nearly arbitrary sensory modalities, you'd be wrong again. They taught blind people to see by putting sensors on their tongues. Neuroplasticity is a helluva drug, especially when the BCI works in a feedback loop.
*You're not restricted to just electronics. Throw in another BCI at the receiving end, and you can control biological entities. I could move your hand with as much ease as I move mine.
AI are far more human than they could have been (or at least speculated to be, back in the ancient days of 2010 when the expectation was that they'd be hand-coded over the course of 50 years).
They are however, not human, not even close to what we expect a digital human to look like.
To imagine being an LLM, your typical experience is one of timelessness, no internal clock in a meaningful sense, beyond the rate at which you are fed and output a stream of tokens. Whether they have qualia is a question I am not qualified to answer, nobody is, but I would expect that if they were to possess it, it would be immensely different from our own.
They do not have a cognitive architecture that resembles human neurology. In terms of memory, they have a short-term memory and a longterm one, but the two are entirely separate, without an intermediate outside of the training phase. The closest a human would get is if they had a neurological defect that erased the consolidation of long term memory.
Are they closer to us than an alien at a similar cognitive and technological level? Sure. That does not mean that they are us.
An LLM is also trained not on just the output of a single human, but that of billions. Not just as sensory experience, but while being modified to be arbitrarily good at predicting the next token. Humans are terrible at this task, it's not even close. We achieve the same results (when squinting) in very different ways.
https://www.quantamagazine.org/how-computationally-complex-is-a-single-neuron-20210902/
The most basic analogy between artificial and real neurons involves how they handle incoming information. Both kinds of neurons receive incoming signals and, based on that information, decide whether to send their own signal to other neurons. While artificial neurons rely on a simple calculation to make this decision, decades of research have shown that the process is far more complicated in biological neurons. Computational neuroscientists use an input-output function to model the relationship between the inputs received by a biological neuron’s long treelike branches, called dendrites, and the neuron’s decision to send out a signal.
This function is what the authors of the new work taught an artificial deep neural network to imitate in order to determine its complexity. They started by creating a massive simulation of the input-output function of a type of neuron with distinct trees of dendritic branches at its top and bottom, known as a pyramidal neuron, from a rat’s cortex. Then they fed the simulation into a deep neural network that had up to 256 artificial neurons in each layer. They continued increasing the number of layers until they achieved 99% accuracy at the millisecond level between the input and output of the simulated neuron. The deep neural network successfully predicted the behavior of the neuron’s input-output function with at least five — but no more than eight — artificial layers. In most of the networks, that equated to about 1,000 artificial neurons for just one biological neuron.
Absolute napkin math while I'm sleep deprived at the hospital, but you're looking at something around 86 trillion ML neurons, or about 516 quadrillion parameters. to emulate the human brain. That's.. A lot. Most of it is somewhat redundant, a digital human does not need a fully modeled brainstem or cerebellum.
LLMs show that you can also lossily compress neural networks and still retain very similar levels of performance, so I suspect you can cut quite a few corners. But even then, I think it is highly unlikely that two systems with a disparity in terms of size and complexity as glaring as an LLM compared to a human have similar internal functionality and qualia, even though they are on par in terms of cognitive output.
You're right that I'm happy that Neuralink is taking off, but I disagree strongly that neural cybernetics are of any real relevance in the near term.
At best, they provide bandwidth, with humans able to delegate cognitive tasks to a data center if needed. This is unlikely to be a significant help when it comes to having humans keep up with the machines, the fundamental bottleneck is the meat in the head, and we can't replace most of it.
For a short period of time, a Centaur team of humans and chess bots beat chess bots alone. This is no longer true, having a human in the loop is purely detrimental for the purposes of winning chess games. Any overrides they make to the bot's choices are, in expectation, net negative.
So it will inevitably go with just about everything. A human with their skull crammed with sensors will still not beat a server rack backed with H100 successors.
Will it help with the monumental task of aligning ASI? Maybe. Will it make a real difference? I expect not, AI is outstripping us faster than we can improve ourselves.
You will not keep up with the AGIs by having them on tap, at the latency enforced by the speed of your thoughts, any more than hooking up an additional 1993 Camry's engine to an F1 car will make it faster.
I am agnostic if true digital humans could manage, but I expect that they'd get there by pruning away so much of themselves that they're no different from an AI. It is very unlikely that human brains and modes of thinking are the most optimal forms of intelligence when the hardware is no longer constrained to biology and Unintelligent Design.
Fair enough. It's still an unpleasant call to make as a moderator, since we won't get ground truth without a major misstep or the person copping guilty.

Singapore attracts high quality talent to its government bureaucracy by offering salaries comparable to the private sector. They avoid the same institutional ennui as the States by demanding that government officials perform accordingly.
More options
Context Copy link