I’m excited for this application of AI, because in many lower-income countries doctors are not trained particularly well and people do not have access to a western standard of diagnosis. Being able to have a consultation with an AI doctor who is more knowledgeable and accurate than local doctors would be life changing for millions, if that comes to pass.
We all know this isn’t replacing doctors. We do know that when applied correctly, this will assist doctors and other clinical specialists in their work. With doctor burnout at all time highs, I think stuff like this is amazing.
> We all know this isn’t replacing doctors. We do know that when applied correctly, this will assist doctors and other clinical specialists in their work.
Let's say a clinic currently has eleven doctors and is able to treat X number of patients per week. Let's say one of them retires, and instead of finding a human replacement, the remaining ten doctors choose an AI clinical assistant to free up 10% of their workload. Now it only takes ten AI-assisted doctors to continue serving the same number of patients as eleven doctors used to do.
This is just an example to show that there is no meaningful difference between "assisting" and "replacing". Any time an assistant, AI or not, takes some workload off somebody's plate, they have partially replaced them, and it adds up.
There is a difference, but with an unwritten assumption: that the AI can do all the things the doctor can do.
If the AI can't do all the things that a doctor can do, then even when it can take up the slack for one doctor retiring, that doesn't tell you anything at all about whether or not it can take up the slack for two retiring.
Right now there's more work to be done than there are doctors to do it; this means that the same number of doctors are getting more things done as the AI improves… but not infinitely more, because there's still stuff the current AI can't do, that only humans can do.
We have a lot of things where technology has fully saturated demand: in food, this is why we've got an obesity problem in much of the world[0]; in medicine, this is how we wiped out smallpox entirely, and are very close to wiping out a few other diseases entirely; in telephony, this is why video conference calls are basically free.
But in each of those fields, there are other things we still have demand for, they're not complete post-scarcity: restaurants, old age, and bandwidth costs are still a thing.
[0] not so for transport, which is one reason why we simultaneously have some people starving
Except there aren’t enough doctors in many places. Also physician performance varies widely.
So really this won’t replace doctors, but hopefully patient care will be affected positively. Doctors aren’t always brilliant outside their field and AI “assistance” can turn out to be a clever troll sometimes.
1. There is a meaningful difference between replacing and assisting. Replacing implies being able to take over the entire function. Technology that assists doctors with one part of their function, or process, can improve their output but is not capable of producing like-for-like output on its own. So the question of whether AI can replace or only assist doctors is very relevant to determining its impact on the role. Power tools didn't replace tradesmen, for example. If AI was able to replace doctors, then your clinic would be able to scale down to 0 rather than 10.
2. If a clinic can use an AI tool to make doctors 10% more productive, doctors become worth more rather than less. Firms are incentivized to hire more rather than less in this scenario. What you're invoking here is the "lump of labor" fallacy. There are market conditions where increasing efficiency really does reduce quantity demanded, but it's not clear that medicine really is one. As far as I can tell, far from there being a fixed lump of medical work, the general population in most of the West is under-serviced and struggles to get reliable, timely, cost-effective access to medical expertise.
Whether or not there is pent-up demand for healthcare, the emergence of AI clinical assistants, like any other form of efficiency increases, effectively expands the supply of healthcare services. In a free market, an increase of supply lowers the price at equilibrium. And as their salaries go lower, fewer people become interested in joining the profession.
We saw this play out in agriculture, in manufacturing, and now it is starting to happen in some services. I do not understand why would we think it will be any different this time around.
In theory, yes. But in practice, at least here in Eastern Europe, there is such a shortage of doctors that even if they became three times more productive, there wouldn't be any meaningful changes in demand.
For example, I haven't had a personal doctor for the past five years because they don't have any free capacity. Last month, I called the doctor twice because my child was sick, and they told me I shouldn't call them so often.
So I think we're a long way from that happening.
>increase of supply lowers the price at equilibrium
The thing you're missing here is that "healthcare services" and "doctor's labor" aren't the same unit. Ceteris paribus, efficiency increases allow the price of healthcare to decrease while the price of doctor's labor increases. The thing that makes this non-contradictory is that a single doctor can now produce "more" healthcare. Economics says the opposite of what you think it does here. Increasing productivity drives expansion in market size, which drives up the ratio of value in the market to its labor inputs which drives up salaries.
Like I said, there might be important real-world reasons why these scenarios won't play out in medicine the way the theory predicts. But so far, you haven't provided any.
Manufacturing has also seen the opposite of what you are saying here. Global manufacturing production value has exploded over the last century, quite literally lifting billions of people out of abject poverty. In particular the last 3 decades of enormous per capita income increases in China have been driven by industrialization. I'm guessing you're taking a US-centric view that is exclusively focused on the local collapse of US manufacturing. This is to do with globalization and free trade, not improvements to labor efficiency.
Yeah, I see this "not replacing, assisting" argument all the time, and it just doesn't work. If you can do more with fewer people, fewer people will be employed for the same task. In the past, nearly everyone was a farmer. We're not all farmers with, like, really easy jobs now because the machines do the boring bits. No, there are far fewer farmers.
I don't think I want to have diagnostic support from an LLM. Perhaps it works most of the time but then you wake up with 13 fingers. The "battle" of Gemini vs GPT-4 doesn't really gain too much.
There are quite good specialized systems for medical applications that were thoroughly tested and vetted against quite high barriers for entry.
I hate the approach of ad companies to approach medical problems. Of course you need patient data for clinical studies, but far more interesting would be to collect data that hint to medical indications and offering up this knowledge to doctors that cannot know about all of them.
LLMs probably will just grow a new generation of hypochondriacs because they certainly will never say that you are healthy if diagnostic supports ever make it into production.
IMO if anything will get us to longevity escape velocity, it's AI. This is why I'm so fundamentally against the anti AI folk. Their jobs simply are not worth the chance that we can evolve this into AGI and achieve escape velocity, saving billions of lives.
I feel the same, and not only for medicine. But like always, we have to be careful to not land in dystopia first and we are barreling toward that rather fast. Internet and fake news was bad before gpt, now it’s spiralling into a nightmare. Bad ‘support’ was bad before gpt, now I often feel that I am in some sort of different universe with companies simply not caring if their support ‘bots’ tell nonsense or keep going in circles; as long as it’s cheap!
So yes, full speed ai, no neutering, but we need laws against profit-at-all-cost (I mean, why is a for-profit company, even if their income is ads, not by law required to provide human support within x hours or get fined? Reasonable I would say) or actual UBI soon otherwise longevity is the last thing you want as you will be living in pure hell; most of the population not being able to pay for anything, enslaved by AIs and the healthcare AIs trying to minimise cost even more effectively than US insurers. I don’t want to live in that world and we are not that far from it.
The "anti AI folk" covers a huge range of different things, for different reasons, with different timescales and concerns.
I feel sympathetic towards all of these groups, even the ones whose jobs will disappear. I felt sympathetic towards truck drivers 15 years ago, as I was expecting self-driving to arrive before 2019.
That aspect of AI alone may well lead to a re-run of "workers who felt their livelihood threatened by automation flung their wooden shoes called sabots into the machines to stop them. Hence the word 'sabotage.'"[0] — and while AI communism could be better or much worse than industrial communism, I'm fairly sure that Overthrowing The Government While Angry is more likely to lead to a bad outcome than a good one.
There are also "anti AI folk" who fear various kinds of dystopia; the stuff in this thread being amongst the options, but also "tiling the universe with smiley faces", and (please excuse fictional references, take the core ideas but don't take them too literally): The Matrix; Westworld; "I Have No Mouth, and I Must Scream"; and every story about an evil genie granting the wish you actually made rather the one you meant.
[0] Star Trek, not sure how much artistic licence they took with that etymology, though the point remains valid in the form of luddites and literal Marx' literal Manifesto.
I don't know which of various meanings you ascribe to "AGI", but the most general sense has no implications beyond the initials: is it artificial, is it "intelligent" (and that in the Turing or Dijkstra senses of what can it do not is it really thinking), and is that intelligence "general purpose".
Surely, if it is AGI it won't do stuff just because it is * told * to same way you as a company don't do exactly what your customer *tells* you to do but what you *think* it *wants or needs*.
So I expect AGI to do stuff we might not understand but further in the future we will look back with enough hindsight and we will understand why it did what it did and we will agree that it was the best action to take at that time. Just like children lacking understanding of why their parents tell them to eat bland veggies or do stuff they don't want to do even though it is good for them in the long-term. I expect this to be the main source of conflicts with AGI in the future.
This is one of the reasons I, overall, agree with the OpenAI position that models should be kept secret until proven safe even though there are examples of safety researchers only being able to perform some of their safety research due to open models.
It's like… OK, so regardless what you think of it, you've probably heard some people blame COVID on a lab leak: I'd say that giving out increasingly advanced models for everyone to play with at home is like giving out increasingly advanced genomes in a world where everyone needs to have their own home gene synthesiser, including highschool kids for their homework — it's absolutely fine until that gene synthesiser reaches a certain level, and then suddenly everyone catches an airborne version of HIV.
Labs may still leak (IIRC, wasn't GPT-2 leaked?), but that's how I think we should be treating models: dangerous until proven safe, not safe until proven dangerous.
"""
I also want our future AI to be literally incapable of wanting to harm humanity, akin to in software "Make invalid states unrepresentable". But, an AI which cannot distinguish good from evil is definitionally Machiavellian; and if a model can distinguish them, and if the model is distributed, then it's almost trivial to find and flip that specific axis. But I did say "and if the model is distributed": if it isn't distributed, then the "switch" can be kept out of sight and away from fiddling hands.
There is another possibile safety mechanism, which is at least worth considering: if all possible moral issues are squashed down into a single dimension (something I've previously noticed humans did, but only by going past it) then anyone with a copy of distributed model will only have one single good-evil axis to modify — they would still be able to make a demonically evil moustache-twirling villan, but they wouldn't be able to accidentally destroy the world by turning off all the dials except the one saying "maximise shareholder value by maximising the number of paperclips", nor would they be able to run a dictatorship by saying "I am the only person who matters, everyone else is my slave".
The downside, and yes there is a downside, is that this will be the morality of whoever makes the model, and basically everyone disagrees on what "good" is. For example, even limiting the world to those who follow the Ten Commandments, "thou shalt not kill" is often regarded as so incredibly obvious that it shouldn't need to be stated, and yet that specifically has sometimes included and sometimes excluded each of: meat, abortion, war, the death penalty, and suicide — can any of us truly imagine how much broader the range and depth of disagreement is in agregate, or is this, too, an unthinkable thought?
This has been said about most absurdly expensive therapies that are now cheap: monoclonal antibodies, gene therapy, stem cell therapy, robotic surgery, proton beam therapy. The evidence is not on your side.
Longevity is neither necessary nor required for "running out of space".
If the mean person has more than 2 kids, the population grows until it can't, no matter how long or short our lives.
If we cure aging and all disease, our half-life due to accidents and homicide is (currently) somewhere around 1000 years — the same "2 kids" rule applies.
Alzheimer's, leading to not drinking enough water (not that anyone around her was aware of that until the coroner's report), leading to kidney failure.
Her last day was very obviously painful, and given the nature of the condition, I suspect that the instantaneous experience of pain was her whole universe.
My dad died from bowel cancer; took a year of gradually wasting away. That was also natural.
I would much rather spend another 900+ years with my loved ones (or however long I want) rather than being forced to die so ridiculously early.
Put another way: if you knew for a fact that you would get murdered 40 years from now, would you choose end it all via peaceful lethal injection today? I very much doubt it.
For every person dying "peacefully" on their deathbed, truly happy to succumb to death, there are 1000 more people that would do anything in their final moments to physically reset to the age of 20.
Also, think bigger - if technology continues to progress, we won't be biological a few hundred years from now. We'll be digitized, each consciousness likely spanning multiple redundant nodes throughout the planet or solar system. So these freak accidents are not relevant - nothing short of a cosmic event would kill you. It's highly unlikely that you'll die before whenever you choose to die - our lives would be indefinite for all practical purposes.
---
Edit because of the ridiculous HN rate limit:
> The "short" life span is one reason we take risks.
No, it's why you take risks. I'm happy to take risks - the fact that I'll die one day has no bearing on, say, my decision to go paragliding this weekend, or to push hard for a promo.
Risk does not necessarily arise out of the certainty of death, it can arise out of any number of other factors.
> I bet we would get a dystopia with long living dicators and short living peasants.
People have said this about most therapies historically but this has never come to pass.
Anyways, why throw up your hands and give up immediately? That attitude never got our species anywhere.
> Altered Carbon style, eh? Cool, sign me up for a cortical stack.
A bit more clever than that :) Individual consciousnesses could be distributed to avoid death from a single cortical stack being crushed.
1000 years would give me time to climb all the mountains, walk all the rivers, hike all the trails, and start a family, write those books, and learn to play those instruments, and still have time for a round-trip to another star.
Why does Gates give away so much money? Because the money itself makes no difference to his life any more.
Why do people leak documents? Because they think other people's secrets should be public information.
Why do we create ownership rights over information, such as patents and copyright, in the first place? To incentivise people to make more of those things. Why are these rights time-limited? Because nobody can, with a straight face, continue to claim they need any greater incentives than they already have.
Why go governments have Eminent Domain? Because the best interest of the people (or the nation) isn't the same as the best interest of the existing owners.
Why did the October Revolution and the Cuban Revolution happen? Because enough of the right people in the right place at the right time, decided that the old system of combined ruling-and-owning was bad.
Flack for his epstein connection? Fear of remote illness? Lot's of ways to reflect your self-interest.
> Why do people leak documents? Because they think other people's secrets should be public information.
Because some people are noble. Some people are corrupted by remote interests.
> Why do we create ownership rights over information, such as patents and copyright, in the first place?
To incentivize publishing of discovery. There's no obligation to publish discoveries.
> Why go governments have Eminent Domain?
Is Eminent Domain justly applied?
> Why did the October Revolution and the Cuban Revolution happen?
Why did my apartment catch flees? I don't know but we had that shit bug bombed.
I think you and I both would be part of some revolution but we differ on the likely conclusion. I personally find a lust for eternal life (except for your fruity loins) repugnant, lest the whole worlds resources become fuel for a gigantic cock sucking machine.
Do any of those alternative responses to my already-answered rhetorical questions change the fact that the fruits held by rich owners ended up on the tables of the poor?
I have literally no idea what you're doing with that metaphor in the last paragraph.
What I'm saying is that if you don't have good reason to think some vastly wealthy thing isn't building a giant cock sucking device then you should assume it is in fact building supreme cock sucker and not you or anyone will get in the way.
What does this metaphor have to do with "will the rich let poor people have medical treatment"?
Given that the rich are, as I have demonstrated via the example of Gates, donating money to make sure the poor have medical treatment, along with all the other examples I gave of the rich being forced to help the poor even when they don't do so voluntarily.
Longevity discovered by AI, implies[0] AI which is so good that none of us can work for money anyway — it comes with an economic change too large to make a reasonable projection of what happens next, as it's more different from what we have today than either Communism (any of them) or Neoliberalism is from Ancient Greek city states using electrum.
[0] but we can't say for sure until we have it, thanks to all the other things we assumed would need human-like AI and didn't, like Chess, Go, writing code, etc.
Maybe it's already discovered but they won't tell you.
Or it's an expensive procedure.
We already produce enough to feed everybody and still people starve to death. The problem isn't production but distribution and that isn't a technical problem but an economical and political one.
Money is power and the ones with power will do everything to prevent losing that power.
> Maybe it's already discovered but they won't tell you.
It would have leaked. People will buy literal snake oil in the hope of life extension, so even just a hint of this would have everyone capable of industrial espionage on the trail.
> Or it's an expensive procedure.
Given all the dead rich people, I doubt it.
> We already produce enough to feed everybody and still people starve to death. The problem isn't production but distribution and that isn't a technical problem
but an economical and political one.
Indeed, and I mentioned the food dichotomy in another comment.
But we also had the thing a few years ago where people were calling on Musk specifically to use his wealth to end world hunger. This resulted in "[…] tech entrepreneur Elon Musk challenged the United Nations last year to show him a $6 billion plan that would end world hunger, he got in return a proposal that would save 42 million people in 43 countries from starvation." - https://www.oxfamamerica.org/explore/stories/how-much-money-...
And in case you blame him for not doing this, it goes on to say "Current estimates suggest that as of this year, we need donor governments to invest around $37 billion every year until 2030 to tackle both extreme and chronic hunger." which is more than his net worth at the time the article was written.
> Money is power and the ones with power will do everything to prevent losing that power.
For now.
AI, regardless of the impact on healthcare, is going to break that in much the same way that the industrial revolution broke the "land is power" apparent-tautology of feudalism.
> That's ca. a quarter of the US yearly military budget.
The USA != Musk personally, and he personally was getting flack for not personally fixing it, and he personally said (not a quote) "if you can give me a plan to do it that cheap, I will", only to be given a plan that wouldn't fix it and the actual cost of fixing it would exceed his theoretical net worth.
> But we still have feudalism.
As a response to the specific bit you quoted, that's just an argument-as-a-soldier, not an interesting point.
It's also an example of "tell me you don't know what ${X} means without saying you don't know what ${X} means" — given that Scotland got rid of the last vestiges between 2000-03 and Sark in 2008, the closest thing in the current world to that is North Korea, or specifically the 500 people who still have Common Rights in the New Forest and the use of an open field system in the village of Laxton in Nottinghamshire.
But none of that matters, as the industrial revolution still broke the connection between land and power; and human-level general AI will break the connection between anything currently resembling money and power.
Sure, it might also figure out energy before longevity or neither of them. I think what we are looking for is the vague ‘AGI’ goal and that, by its (vague) definition , is capable of finding multiple solutions to many problems.
Solutions clearly exist for managing diseases (we keep finding ways to give years to cancers that were an instant death sentence just 15-20 years ago); I have no clue if they exist for making us 180 in good health; I am not that smart.
And space flight is also solvable if energy is solvable and that also seems possible, but again, I am not smart enough otherwise I would be in that business.
Human conflict is solvable in a simple way; just eradicate us all. That solves all but longevity anyway for some definition of solve. But seriously speaking; if a human had a solution for these conflicts, they wouldn’t be there so either a solution without peril to humans is not there or we are too dumb to come up with one and need a 400 point iq increase to do so. I am not saying AI will ever do it but so far it seems to be the our only hope. Then again, it’s only been a few 1000 years of modern human so we might figure it out without finding AGI.
Anyway, currently this is more ‘religion’ (aka hope) than science on my side; there is no clear sign agi or super intelligence will happen soon; I am now more deadly afraid of ‘intelligence-ish simulators’ (LLMs) used everywhere to replace people and making everything vastly worse for almost everyone. Which is what’s happening at the moment.
Candida auris is a really bad example, given it predominantly affects people with weakened immune systems.
To make that point better, you could've gone for COVID-19, what with the number being the year it popped into existence.
But if you had, then the point would be: yes, the world is dynamic, that means our world needs to continue developing new things, not that we can't get eternal youth.
Yeah sure, by instant I meant ‘get your affairs in order’ type of message from the oncologist after detection. Aka an instant death sentence. For some kinds of cancer where you got that in the 90s, now you have a good chance of walking out cancer free. Like a familiar cancer gene my family has: agressive, nasty, death sentence after detection in the 90s. Now multiple family members have been treated in the last few years and walked out cancer free. That was incredibly rare 25 years ago even with early detection; now it happens even in later stages. If it comes back they are dead but still; every few years added might result in yet another bump.
But sure others pop up, not sure if we ever can stop that; this is where AI could help; tireless and very fast moving wack-a-mole with diseases, new and old.
"[Everything] always seems impossible until it's done" - Nelson Mandela
There's no longevity escape velocity yet. But when I was a kid, there was no cloning of mammals, no Google Translate real time video mode on a device you could carry, and the best chatbots were ELIZA clones.
A better question is if it is a fitness advantage to die. And it is. If you have two species and the first dies when something happens and the second doesn't the first will "win". The first species will have its old die out and so the young will be able to evolve faster without having to compete with everyone else that is alive, only the current generation.
We have this "kill switch" in our brains already. It is the PVN. Very much oversimplifying, but as bad things build up "inflammation", it kicks of production of cortisol to deal with it. The cortisol reduces neurogenesis and little by little our brain dies. The PVN also controls LH levels so sex hormones so rising cortisol downregulate CRH and the PVN. Puberty and menopause are just the speed changing of GnRH to match. Pain, wounds, immune system reaction, they all go through this. Imagine, at one time there was a species that wouldn't die of natural causes when it was in constant pain.
Vertebrates have the PVN, invertebrates don't. Vertebrates (nearly always) die, invertebrates generally don't.
There is a reason why all the stuff to increase life span is (generally) about healing or reducing inflammation.
I am skipping over a lot of details, but you get the point. Dieing is an evolutionary advantage, not to you, but to the species.
You are invoking an argument that's called "group selection" [1], it's the idea that genes propagate based on whether they benefit the species as a whole. This argument has been heavily criticized and it is not used much in the mainstream evolutionary science. Perhaps group selection has some effect in some niche, limited circumstances but it doesn't seem to be a significant driving force of evolution.
You don't need to invoke group selection to explain dying though. Dying might be an evolutionary advantage, not to you, but to the genes that make you. Which is not the same as the species, at all [2]. Organisms die not to make space for others but because building the mechanisms to keep young indefinitely is not worth the price compared to spending these resources on reproduction mechanisms. Or because the right combination of genes that allow you to have free lunch just wasn't reached yet. If there was a magic mutation that prevents mammals from ageing without any other effects, I'm pretty sure that gene would spread. Quite possibly to the detriment of the species.
(Beware, not an evolutionary biologist, just somewhat interested in the topic.)
1) Nature intends nothing. There is no reason, in any logical sense.
2) If nature actually intended anything (again, it doesn't), it would also intend killing most of your children before they are 5. Using medicine to keep infants alive is just about as unnatural as extending our life span towards the end. In terms of morality, doing the later in a way that also keeps up quality of life is obviously important, but nothing that we can not chip away at, just as we did with child mortality.
Nature is red in tooth and claw, it doesn't care about us.
I also find it curious how many people continue to appeal to nature, while tapping a finger against a sheet of synthetic sapphire or glass that protects a delicate sheet of semiconductors, themselves made to glow by minerals taken from all over the world refined in complex chemical processes, and that glow guided by an unnaturally pure crystal of silicon cut and marked in ways too small to be perceived with any natural vision, in order to create a set of glyphs which silently convey meaning to others worldwide via a network of similar devices intermediated by space lasers and a planet spanning grid of cables laid at the bottom of the world's oceans.
Nature gave us our lifespan for the same reason it gave wolves a decade, the same reason it has elephants starve to death when their final set of teeth fail, and the same reason it made octopods go into a self destructive death spiral as soon as the eggs have been laid.
