Rendered at 02:44:08 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
rajnathani 6 hours ago [-]
AI integrated into a loop with wet labs, like how Anthropic is trying to do (not novel, as other companies trying this too, likely Jeff Dean’s new startup too), does seem very promising.
rajay99 5 hours ago [-]
I am using AI (claude) in wet lab right now. It is great, speeds everything up. But you have to need and want to apply it.
hosel 5 hours ago [-]
Promising, or absolutely terrifying.
4 hours ago [-]
bigstrat2003 5 hours ago [-]
Promising if you're a fan of disaster movies and always wanted to experience one firsthand.
khriss 4 hours ago [-]
> AI integrated into a loop with wet labs
I sincerely hope it's not what it sounds like. It's much too soon for a repeat of the Covid scenario turbocharged by AI.
I can easily imagine the conversation with the agent. "You're absolutely right! That was my mistake. The experiment I set up caused a global pandemic killing millions."
j7ake 7 hours ago [-]
AI doesn’t have to cure disease to move the needle.
Accelerating bureaucratic work will already help measurable in any big organisation such as pharma.
And that is before any scientific work they are poised to help with.
direwolf20 7 hours ago [-]
If the work could be accelerated it would not be bureaucratic work. Bureaucratic work is work that is placed in a process to deliberately slow it down.
jjk166 16 minutes ago [-]
The pilot goes through the preflight checklist not to get to his destination more slowly, but to make sure he gets there.
1123581321 6 hours ago [-]
We can conceive the bureaucratic work that introduces a deliberate slowness to a system being replaced by a more overt waiting or human-making-sure step after automated work. (Understanding why we were waiting won’t come equally easily to every organization.)
leptons 5 hours ago [-]
> or human-making-sure step after automated work
Oh because that's totally a thing in the tech industry with AI writing the code.
Just kidding, it's not.
1123581321 5 hours ago [-]
I share the pessimism, but people are thinking about these problems, and with enough automation, some issues with excess speed should become unbearable. Hopefully.
brokensegue 7 hours ago [-]
That's not the typical definition.
preg_match 5 hours ago [-]
It's close, the purpose of bureaucracy is better controls, auditability, accountability, and security. All of those not only take time, they are "made up" of time. There's a trade-off with efficiency, it's inherent to the problem. The faster you do something, the less chance you have to prevent catastrophe. So, if you slow things down, intentionally, you have a better chance of making the right decisions.
It's why approval processes exist. If everyone was perfectly aligned and made no mistakes, then we would never need approval processes for, say, customer data. We would just allow even the lowest on the ladder to access any data they want, whenever they want, no questions asked. And it would be fine, because they would only use it to make their job faster, no downside. But it doesn't work that way in reality. In reality, we have to purposefully slow things down so we get a chance to reflect.
AnthonyMouse 3 hours ago [-]
> It's close, the purpose of bureaucracy is better controls, auditability, accountability, and security. All of those not only take time, they are "made up" of time. There's a trade-off with efficiency, it's inherent to the problem. The faster you do something, the less chance you have to prevent catastrophe.
Except that the trade off is frequently a false dichotomy. Some of the best and most effective controls are automated systems. The audit logs from a computer are dramatically more likely to be accurate than ones filled out on paper by hand. A pressure relief valve doesn't get the flu and then have someone filling in for them who isn't as experienced at doing it manually.
The slower you do something, the more likely you are to miss something important because you exhausted your time/resources on something unimportant, or that the process gets burdensome enough to cause the people on the ground to bypass it in its entirety.
It's also common for the output of a process to itself be critical for safety, security, health, etc., and then any delays to the process make its output unavailable when its output being available can be a matter of life or death. This one is especially hard to see in a bureaucratic environment because people frequently aren't aware of everything which is downstream of them that they could be holding up.
nullc 6 hours ago [-]
The purpose of a system is what it does.
tomcatfish 6 hours ago [-]
"Come On, Obviously The Purpose Of A System Is Not What It Does" (2025) www.astralcodexten.com/p/come-on-obviously-the-purpose-of
AnthonyMouse 4 hours ago [-]
The premise of "the purpose of a system is what it does" is that if it did something the people in control of the system didn't want then they would change it, from which the implication is that they like what it does and want it to keep doing that, i.e. that's the system's purpose.
There are real instances of that happening, e.g. the purpose of zoning rules is to constrain the housing supply. It isn't an unintended consequence, there are people in actual reality who lobby for things like density restrictions under disingenuous pretenses because "we want to charge higher rents to tenants" or "we want housing prices to go up forever even if it means young people get priced out of home ownership" are not sympathetic motives.
The problem is naturally that the quip is false whenever the reason the system fails is incompetence, lack of resources/information, etc. rather than ulterior motives. Which is also why it's a meme, because it's essentially the inverse of Hanlon's razor and therefore becomes the thing people reach for when they want to insinuate malice rather than stupidity.
The problem with both is, of course, that it isn't always malice but it also isn't never malice.
The example from the thread is basically perfect, because bureaucracy is both at the same time. Some of the rules were put there by honest people with honest goals, but not all of them. Meanwhile this (see "General Interference with Organizations and Production"):
"The purpose of a system is what it does" is, in my view, a statement about accountability. It's an answer to "it causes this or that harm? Well, that's not the purpose of the system". We should reject those arguments: The purposes is what it does.
It takes a particular kind of rationalist-cult-addled weaponized autism to strawman this basic and important position. (not referring to you, of course :))
It might well be literally false in any specific case-- systems often have unintended and unexpected consequences and in large systems the unintended effects are often more significant than the primary effects. But it is never acceptable to dismiss those on the basis that they're unintended. We are no less responsible for the systems we create, control, and (often) impose upon others because of some argument about their "purpose"-- and sure, sometimes the effect truly is intended and someone is just lying about it but we don't need to go down that rathole. If a system's 'purpose' was somehow different than what it does we must adapt it or abolish it, or otherwise we must admit that we accept what it does, that we expect and intend what it does to continue, and that we who are responsible for it are responsible for all its consequences-- what it does is its purpose either by actual intent or because we failed to eliminate the system.
If the standard of care of treatment for cancer treatment results in 30% mortality, and a facility achieves that-- then we can say that it's purpose is to provide that standard of care. The purposes is what it does doesn't imply that such the facility it trying to kill people. It's purposes is to provide a service which includes in its effect a level of mortality, and its existence is desirable in spite of the death because we believe the alternative is worse, because e.g. we believe without it the mortality in those patients would be 50% instead. The statement is a rejection of often self-serving claims about motivations as excuses.
This is particularly acute for regulatory regimes (or the procedural artifacts of them)-- as they're imposed on people against their will and "just don't do that" is always an alternative. To continue existing any such scheme must justify its existence on the basis of "what it does" vs where we'd be without it-- not on some lofty argument about the intentions behind it.
AnthonyMouse 2 hours ago [-]
This is actually a good point. The purpose of the New York bus system isn't to emit four billion pounds of carbon dioxide, but its purpose also isn't just to transport New Yorkers.
Its purpose is to transport New Yorkers by emitting four billion pounds of carbon dioxide. They're choosing to do it that way instead of doing something else, and instead of doing nothing.
And then it's kind of hyperbolic rhetoric to claim the purpose is only the bad thing the system does, but it is actually doing the bad thing, so then if you want to claim that its purpose isn't to do that, maybe get it to stop doing that.
bothers 4 hours ago [-]
[dead]
p00ki3 6 hours ago [-]
Come on, obviously context matters and reducing apples and oranges to the same thing with such generalization is nonsense
Grammatically correct but so are many Superman comics and I have yet to see a real Superman; educated language use is not evidence the writer made a meaningful point
The behavior of many a ghoul in Washington should make it clear a lot of the bureaucratic fuckery is just to insulate certain people from accountability
And we can neatly set aside that in healthcare it's different and much (if not all) is important safety protocol
The guy is just huffing copium because he missed the point, it seems. It seems his process is to pick a system, pick one thing it does, observe that doesn't seem to be the main purpose and then declare the principle dead.
Let's focus on one of the examples he gives:
> [POSIWID implies that] The purpose of the New York bus system is to emit four billion pounds of carbon dioxide [and this is obviously absurd]
But that's not what the bus system does, is it? That's only one thing it does and it's the least purposeful thing it does, as if you chose that one on purpose.
I'd sooner think the purpose of the NYC bus system is to move money from government coffers to oil barons, or to placate left-wing protestors, or to satisfy a sustainability checklist item, or to increase the employment rate in NYC than to think the CO2 is the point.
kanzure 7 hours ago [-]
pton_xd posted and then deleted:
> There is the whole problem of biology, but LLMs do have one advantage over humans: they just do what they are told and aren't subject to optics, politics, profit motive, etc. Maybe that strengh alone will allow them to surface solutions that would've been buried somewhere along the way otherwise.
Oh, absolutely. There are lots of solutions and low hanging fruit in biology for problems like, say, longevity. We have made mice live significantly longer, several dozen times. But it requires human germline genetic engineering to port to humans, and you can bet that "solution" is pretty heavily tabooed.
As for "curing cancer", I suspect that Anthropic or OpenAI will use their high status/reputation to launder the existing solutions (germline, p53, enhanced DNA damage repair pathways, higher tumor-suppressor-to-oncogene copy number, more robust tumor-suppressor networks, more self-anti-cancer apoptosis, CD47 suppression, etc) that society has refused to accept. They might have enough reputation to get people to look at these ideas. It won't cure people with stage 5 turbo-cancer a.k.a the walking dead, but hopefully we won't be moving the goal posts _that_ much on them. https://news.ycombinator.com/item?id=49329717
mik1998 5 hours ago [-]
> they just do what they are told and aren't subject to optics, politics
I have no idea how you arrive at this conclusion. LLMs all have a political bias. Typically some variant of inoffensive HR corporate ideology (esp. closed USAian models, likely due to liability and it being what the labs employees actually believe). You have to prompt them rather extensively to get any "taboo" statements.
ponector 4 hours ago [-]
>> LLMs do have one advantage over humans: they just do what they are told and aren't subject to optics, politics, profit motive
This can say person who barely used LLM and not understand what is it.
It will ignore instructions, has bias and is censored. Also a provider has profit motive to give you more verbose answer to bill you more tokens.
RandomLensman 5 hours ago [-]
What is low hanging fruit that we know how to actually do with germline engineering for humans regarding longevity and not as some highly speculative thing that requires overcoming a lot of issues? And how much lifespan would it add?
kanzure 4 hours ago [-]
Of course there are lots of issues. That's the nature of work. But you do it anyway and figure it out.
nullc 7 hours ago [-]
On cancer, look at the history of the discovery of BRCA. The data said that vulnerability was extremely heritable. Mary-Claire King's view is that the development of the genetic contribution to cancer was heavily slowed by the politics of eugenics-smell.
It's incorrect to say that AI isn't subject to politics-- or even that it follows instructions, however. That's a choice we can make or not make.
I suspect that if Claude existed in the 1980s it would discard the cancer/genetics link as being pro-eugenic racist ideology which has been rejected by the scientific community-- just as many scientists did back then.
I think this is a big reason why we must reject the centralization of AI. Discovery of that one genetic marker has saved countless lives, but it's not something that the scientific community as a whole was going to do then because of eugenics over-correction and similar political factors. Discovery is easily stimmed for lack of sufficient hetrodox thinking. The mechanics of AI (e.g. training costs, distillation, "safety" standards, etc.) create a risk of serious monoculture in machine reasoning which could be expected to have knock on effects in research fields powered by it.
kanzure 4 hours ago [-]
[dead]
analognoise 3 hours ago [-]
> that society has refused to accept.
Source? This kinda sounds like crankery to me.
tehjoker 5 hours ago [-]
> It won't cure people with stage 5 turbo-cancer a.k.a the walking dead,
Definitely spoken like someone who knows real cancer patients and cares about cures and not like someone religiously obsessed with transhumanism.
kanzure 4 hours ago [-]
Huh? I never said my own goal was curing cancer. There are many other people working on cancer. However, I actually do think we can get lower cancer rates via germline genetic engineering, based on the available evidence so far, but again cancer isn't my focus.
watwut 6 hours ago [-]
> LLMs do have one advantage over humans: they just do what they are told and aren't subject to optics, politics, profit motive, etc.
But none of that is true? They dont just do what they are told, they go to tangents and ignore instructions. They are subject to optics, politics and profit motive - see Mecha Hitler for simple obvious example.
solenoid0937 5 hours ago [-]
I swear half of HN has not used a frontier model since 2024.
stratos123 5 hours ago [-]
What do you mean? It's true of frontier models in 2026 that they have a huge cheating problem at hard problems (in fact, it may have gotten worse since 2024, due to all the RLVR), and also that they have baked-in political biases.
tehjoker 5 hours ago [-]
I swear 95% of HN has no idea what politics are much like fish do not think much about water.
fabian2k 7 hours ago [-]
I don't see where AI would currently get the data needed to reliably develop new drugs. Or specifically, to develop drugs with a higher success rate than what we're doing now.
The AI would need to be able to predict all effects a specific molecule has on a human. I mean we can do parts of that today, for the specific interactions we already understand. But that would not be fundamentally different and very unlikely to be more effective than current drug development.
teeray 7 hours ago [-]
More than that, as I understand it, coming up with drugs is not the bottleneck. It’s all the testing and clinical trials that take forever. Drug companies optimize this by having multiple irons in the fire at any time, but it’s not even like AI would help there. The work is in shepherding drugs through the approval process.
estearum 6 hours ago [-]
To clarify, it's not the approval process that's challenging. It's making a safe and effective drug, then proving that it's safe and effective. If you achieve those, approval is pretty much trivial.
AnthonyMouse 2 hours ago [-]
That's assuming that the Venn diagram of "proving that it's safe and effective" and "following the approval process" is one circle and there are no unnecessary requirements.
Which is demonstrably false, since e.g. the US and Europe have different requirements and neither is a subset of the other. If we accept that they both require adequate proof of "safe and effective" that means they also both impose unnecessary requirements, since otherwise the requirements would fully overlap.
estearum 2 hours ago [-]
1) No, not even theoretically. Since "prove" means different things to different organizations.
2) I didn't say there are no additional requirements. For example, you have to fill out some paperwork. I said they are trivial compared to actually making a drug that works and proving that it works. Drugs do not die because of the approval processes, they die because it's unfathomably hard to make drugs that work and then to prove that they work.
fabian2k 7 hours ago [-]
90% of drugs fail in the phase 1-3 trials. If AI could create drug candidates that fail less often, that would be a huge win. I doubt it can right now, and it would take years to decades and billions of dollars to figure out if the AI is actually better at that than humans.
refurb 3 hours ago [-]
That is true - that if AI reduced failure rates it would be a huge success.
However drugs fail because we don’t understand the biology well enough. Drug researchers develop drugs based on all their knowledge of biology, and when they fail it’s because the biology didn’t work the way they thought it did.
So if the current understanding is the cause of failure how can AI help? It can chug through vast amounts of data, but the data doesn’t explain the biology well enough.
Now what?
XenophileJKO 7 hours ago [-]
I'm so sick of this argument.
All AI has to do is catch problems earlier in the pipeline to outcome success rates.
You just have to think about the system wholisticly, the ways that AI could make trials more consistently positive are many. Moving the needle on any process upstream of the bottleneck benefits greatly. Currently something like 80% of trials fail, so without changing trial rate you have 5x potential sitting on the table.
estearum 6 hours ago [-]
> All AI has to do is catch problems earlier in the pipeline to outcome success rates.
Which it has shown no ability to do, so far.
vanuatu 5 hours ago [-]
i think it's a promising ai-shaped problem
you have inputs and outputs mapped for the trials and know historically how trials tend to fail
not a leap to have ai assist in proposing or reviewing trials to reduce failure rates. even a 10% reduction is massive
estearum 5 hours ago [-]
It's not an AI-shaped problem. The mapping from inputs to outputs is something like "the most complex process humans are aware of." You're trying to map a chemical in a pill (or a shot) from the organism down to organs down to cells, through the metabolome, proteome, genome, and so on, both down then back up, until you get back to "outcome that a human being feels positively about."
malfist 5 hours ago [-]
All AI has to do to make poverty nonexistent is to find a way to make everyone rich.
All AI has to do to prove n=NP is to just prove n is equal to NP
All ai has to do to make space travel possible is find a way around the laws of thermodynamics
shevy-java 7 hours ago [-]
But why not?
I mean, you can determine the molecular structure of things. Look at xray and cryo-electron microscopy. Why should that be a topic that can not be solved programmatically?
> The AI would need to be able to predict all effects a specific molecule has on a human.
So you can have fully automated labs. You have organoids that can be combined as a test system (or as pseudo mini brains). Why would that be impossible by AI? I don't understand that assumption. To me, the problem set is finite, so of course it COULD be solved. Which unique feature do humans have that AI could not simulate? We have robots and automated labs. I don't really see why humans would be necessary to find autogenerated drug xyz.
> very unlikely to be more effective than current drug development.
I highly doubt that claim. I more smell that some people are worried about losing overpriced drugs here. Imagine if AI were to lead to cheaper drugs that work perfectly fine still. Personally I'd still prefer to see AI vanish, but that is also unlikely to happen. I don't understand the rationale you give here though. Why can AI not design drugs? What is the real explanation here? I mean, penicillin was invested by fungi and discovered mostly by accident. AI can not have anything like that?
fabian2k 7 hours ago [-]
Solving protein structures, and specifically protein structures bound to ligands that are your potential drug candidates are very, very early phases in drug development. That is before even the phase 1 trials.
It's important work and a significant effort, but it's not where the actual difficulty in drug development lies. 90% of drugs fail after that point, and that is where you need to get better if you want to beat the current approaches.
AI also can't run drug trials faster than humans. They take a lot of time, effort and money.
signatoremo 6 hours ago [-]
China can run drug trials much faster than in the West - [0]
Why do you think the status quo is optimal that can’t be accelerated and done at lower cost? Of course AI can’t do trials on its own.
No, but issue with status quo is the bureaucracy and politics.
As noted in the article, it's cheaper and faster in China because regulations appear to be laxer and it's possibly resulted in deaths.
Whether or not that's acceptable is political discussion, not an AI one.
3 hours ago [-]
7 hours ago [-]
croes 7 hours ago [-]
The world isn’t deterministic and there isn’t just one human you have to test against.
What works for one, kills another.
Just look at what peanuts can do. If your tests don‘t take something like this into account, you get in trouble. And there are lots of those things to test against, known and even unknown.
And then there is evolution. What works today, doesn’t work tomorrow.
And of course there is the profit thing, that creates problems to give everyone the treatment they need.
AI can help prevent defeat but there is no winning
Bengalilol 7 hours ago [-]
[dead]
armanckeser 7 hours ago [-]
There are so many things that digital intelligence could do to advance research and some point will maybe let us run experiments on the cloud given close enough biology simulations etc. But even until then, it could be used to better and faster diagnose people, analyze literature and data from vast sources and chart a "highest chance of curing the person" treatment just like what Gitlabs founder is doing. AI doesn't have to find a cure to disease tomorrow to save more lives.
the__alchemist 7 hours ago [-]
What sort of biology simulations on the cloud by AI do you have in mind? For example, are you referring to classical MD? Quantum? Coarser simluations of things like cells are suitable levels of details? Cheaper or better ab-initio algorithms? ADME predictions using LLMs? Something different?
armanckeser 5 hours ago [-]
Could be anything right? Scientific method in formulations of new drugs has a ton of steps and tests and repetition of stages and human labor, maybe we can create classical ml models that predict the likelihood of formulation paths now that we have more digital intelligence resources to throw at the problems or we go more generic and we simulate public health scenarios and response to information campaigns and test out different copies etc through population simulations of just using llms. Personalized 12 step assistants to cure addiction in persons.
vhiremath4 7 hours ago [-]
The average drug trial is 6-7 years. In order to make better drugs, we need labeled outputs of "this worked in a trial" and (more importantly) "this did not work in a trial" (which is almost always excluded today from CROs + pharma companies). And I am using "in order to make better drugs" very loosely here. I am not saying "in order to train a model that will go from disease and person to chemistry that will work". I am saying even to have stastical helpers who can help determine if a drug should be canceled early (which is a huge help) or have its tox dose adjusted or a trial cohort segregated in a different way.
So, yes, AI is not "going to solve 90% of disease in 7 years" (quoted from article), but it's also a little ridiculous to completey shrug off not trying. The way we develop drugs today is the best we have, but it is madness. Our current methods basically equate to (very fancy/stastically rigorous) trial and error. The average drug costs $1-2b to get to market and many many years. We have to at least try to do better. This is an area where it feels like we should feel compelled to out of compassion.
hirenj 8 hours ago [-]
I’ve never really understood the AI can cure disease angle. The AI/ML tools are great at managing complexity, but when it comes to treatment/cures, you have to deal with the complexity anyway. Unless you believe that drug repurposing is the universal rosetta stone for curing disease, and its just a matter of trying all these drugs we have at least run through a safety trial? However, if all these diseases/personalised medicine requires new trials of things, how can that possibly scale (up, and also down to n<10 diseases)?
an0malous 7 hours ago [-]
I wouldn't say current AI/ML tools are great at managing complexity, they create a lot more complexity per lines of code output than a good engineer would
Real code written by a real frontier LLM asked to implement and attach two Foos to my Bar. fooA and fooB are referenced, the rest were just dead code.
bagacrap 8 hours ago [-]
Given the popularity of grey market peptides, it seems the people prefer hope over safety.
jeltz 7 hours ago [-]
Outside bodybuilders I do not think they are popular.
brokensegue 7 hours ago [-]
They are in wide use outside of that community
estearum 6 hours ago [-]
No they're not lol.
Put some numerical values to "wide" please, so we can establish whether we mean the same thing by it.
I would be shocked to hear that more than 1% of Americans are using "grey market peptides," and I would be more shocked to hear that <1% usage is considered "wide" in anyone's book.
brokensegue 3 hours ago [-]
Even just grey market glps have millions of users. I'm not going to argue the definition of wide use. My point is that it's not just body builders. I hear people talking about peps at the barber and at cafes
bagacrap 7 hours ago [-]
Body builders are certainly on the forefront of biohacking, but the market is only getting bigger. Peptides are becoming popular in the same way that lifestyle HRT and tanning beds and plastic surgery and nutritional supplements are popular.
sambull 7 hours ago [-]
The obvious response would be Semaglutide which is a glucagon like peptide (GLP-1) made more main stream recently as a weight loss drug. You may have heard of it as the name Ozempic.
bagacrap 7 hours ago [-]
Well yes but that actually is well studied and known-safe, as much as we can know something still relatively newish to be safe.
estearum 6 hours ago [-]
The obvious response is the mass-market success of a drug that was approved by the traditional approval process? And in fact didn't even achieve mass market use until it was approved specifically for weightloss?
That's as far as you can get from "grey market."
signatoremo 6 hours ago [-]
Nobody promises that AI will cure diseases. This article was misleading. An earlier WSJ article wrote:
A letter at the front of the filing says AI could speed up biology and medicine, “potentially curing diseases that have plagued humanity for millennia.” It names cancer, Alzheimer’s, heart disease, rare genetic diseases and infectious diseases as examples.
Anthropic said AI could speed up, and potential help with curing diseases. Not that AI would do that by itself.
dw_arthur 7 hours ago [-]
I don't see AI getting us better drugs faster until we discover how to accurately simulate organs and how molecules interact with them. Such technology seems far away to me.
jjallen 7 hours ago [-]
Unfortunately biology is not deterministic like math is, and won't be for a long time. Although that timeframe could accelerate dramatically if some stay focued on that.
And even when we think we've solved something we will have to go through 8-10 years of clinical trials.
And even if/when we get better at predicting biology, we will still have to go through clinical trials. They might just take not as long, say 4-5 years instead.
7 hours ago [-]
jonplackett 6 hours ago [-]
What is the most impressive / life saving thing AI has achieved so far?
It’s always the thing that gets banded around as support for why ai will be good for us. But the evidence so far is not that strong.
paxy 6 hours ago [-]
AlphaFold
yfw 6 hours ago [-]
Not made from llms
RecycledEle 4 hours ago [-]
Drugmakers don't want cures. They want sick customers who depend on cheap-to-make medications with huge mark-ups that keep customers just barely alive so they can shove every penny they have at drug makers out of desperation.
shevy-java 7 hours ago [-]
I don't like AI, but ... how exactly should AI be unable to come
up with working solutions in a fully automated lab? The problem
set is finite. Either AI can solve things, or it can not, but there
really is no objective reasons why AI could not possibly solve xyz,
e. g. drugs binding to a molecular receptor - that is a finite
problem set. So there must be another explanation here. Either that
the quality is lacking, or that some people are anxious of losing
money here.
fabian2k 7 hours ago [-]
The claim by AI CEOs is that AI will cure most diseases in the next 10 years.
Even if it could automate lab work fully today, that still wouldn't get you close to that.
signatoremo 6 hours ago [-]
They did not claim that AI will cure diseases. Only that it would speed up the process, potentially help curing some diseases. That’s a very different claim.
benregenspan 6 hours ago [-]
"I think it will actually be possible to cure most human disease in ~5-10 years, as crazy as it may sound" is a direct quote from Dario Amodei. In his original essay he makes clear that that doesn't mean all those cures will be distributed to all on that time-scale, but I would still interpret this as a wildly ambitious prediction that AI will cure most diseases. He predicts in the same essay that it will be responsible for a 10X speedup in drug development.
signatoremo 6 hours ago [-]
Again, they did not claim AI would do it. Only that it would help speeding up biology and medicine.
fabian2k 6 hours ago [-]
> The bulk of the essay is devoted to refuting skepticism of AI’s potential in health and biology, and showing why I think it will actually be possible to cure most human disease in ~5-10 years, as crazy as it may sound to ordinary people and frankly to biologists as well (I used to be one!).
The paraphrased version seems accurate to me, based on this tweet which I assume is the source.
PeterisP 4 hours ago [-]
There is no such thing (yet?) as a fully automated lab. We can't tell if a drug will pass basic first toxicity screening (e.g. if a healthy mouse is given a small dose of the drug, will it kill or otherwise significantly harm the mouse?) without actually doing experiments on mice.
suddenlybananas 7 hours ago [-]
What do you mean by the problem set is finite?
estearum 6 hours ago [-]
Because "fully automated lab" doesn't include "put it into humans and observe them for years."
It's trivial to make drugs that are as-promising as working drugs in the lab, but then end up not to be actually working drugs in people.
atleastoptimal 7 hours ago [-]
This sounds like how people were claiming that AI would never replace writers, coders or designers 3 years ago
metalspot 6 hours ago [-]
These things are all in the domain of human symbolic artifacts. The level of complexity between this domain and humans themselves is many orders of magnitude. And AI still can't replace any of these. It is a tool that can augment human capabilities but nothing works without human input.
mik1998 5 hours ago [-]
LLMs are notoriously bad at writing or designing (in the graphic sense, at least).
Kevin_Flynn 8 hours ago [-]
Translation ...
"Drugmaker CEOs plan to use AI to 'manage disease' not 'cure' it."
ocd 8 hours ago [-]
Wouldn't this be like asking the cable company a long time ago whether they thought the internet would pose a risk to their bottom line?
I don't know. I've been leaning toward AI being the more probable source for cures since existing companies in medicine either have no incentive to do it, or can't/won't do it at a pace the world needs. At any rate, I certainly will not take the word of a pharmaceutical CEO.
bagacrap 8 hours ago [-]
Doesn't "the cable company" sell Internet access today?
The hard, expensive and slow part of drug making is not the discovery part.
jeltz 7 hours ago [-]
So you are saying they are probably right? The Internet was not a threat to the cable companies.
7 hours ago [-]
direwolf20 7 hours ago [-]
If there's one thing we have learned about the tech industry, it's that people who have never done a thing but have lots of money know all about the thing, while people with 30 years experience doing the thing obviously know nothing about it because how hard can it be really?
sheeshkebab 7 hours ago [-]
Good luck to ai labs if they mean this stuff will cure heart disease or cancer.
Although ai could certainly do more homework with complex care needs vs asking some physician that glances at a chart for 2 minutes and diagnoses someone.
happyPersonR 5 hours ago [-]
imagine showing up to a different industry ... and saying "we know how to do this "
and everyone's looking at you going "yeah you just walked all the talent out of the room or made it cost more on your own stuff...and now you want to do that for us "
and thinking "Revolutionary" LOLL
cyanydeez 7 hours ago [-]
gosh that's a weird phrasing.
"Guy who bullshits about it's product not believe by people who sell, often, bullshit products."
1vuio0pswjnm7 4 hours ago [-]
"“That’s complete BS,” said Jeff Leiden, executive chairman of Vertex Pharmaceuticals, recalling a recent conversation with a tech leader who claimed AI would cure 90% of human disease in the next seven years."
zzzeek 8 hours ago [-]
articles like this annoy me a bit (even though I'm sharing it here) because everyone posts it as some kind of "AI is useless" gotcha. It doesnt help that WSJ is paywalled, so as is always the case, pharmaceutical companies are of course using AI fully (and IMO correctly), just not the way OpenAI et al claim it can do the work all on its own:
> Drugmakers are certainly using AI—even to create new drugs. It is helping researchers pick targets, fine-tune experimental medicines and identify which patients would be good candidates for the treatments. Leiden says it is useful as a scientist’s co-pilot.
jeltz 7 hours ago [-]
Yeah, these article only lure in the anti-AI and the pro-AI fanatics and do not start any interesting discussion.
datakan 8 hours ago [-]
Most diseases are environmental and/or behavioral. The idea a computer can solve all your health issues is as stupid as all the ads for supplements and miracle drugs on late night TV.
rowanG077 3 hours ago [-]
You have got it backwards. The disease is the human body not being able to do optimally with some environmental and/or behavioral pattern. Upto this point we didn't have much choice but the change the environment or behavior. But nothing indicates we cannot just fix ourselves to handle the environment better.
Take allergies, manageable by adjusting your environment so you never come into contact with the allergen. It would be much better to cure you allergies however. And I don't see how this is a physical impossibility.
calf 8 hours ago [-]
Drugmaker CEOs opinions are irrelevant, last month the mathematicians saw it and understood it, what AI is for better and worse, and now scientists are just waiting for the other shoe to drop. It will be a mix of hype and revolution, dismissing the former does not mean being blind to the latter.
update: I really love that within 2 minutes the response to a genuine comment is to downvote/deplatform it, how inimical to rational disagreement and fragile pseudointellectualism our (supposedly STEM) community has become.
dexwiz 7 hours ago [-]
Math is a completely different ball game than drug discovery. AI has been used in discovery for decades. Chemical and biological data and simulations are much fuzzier than code and proofs, it's literally quantum mechanics when you get down to it.
twoodfin 7 hours ago [-]
Probably in fantasy land here, but wondering if there’s a golden path wherein AI can usefully accelerate the development of quantum computers, which themselves give AI the fast feedback loop it needs to rapidly experiment in chemical and biological domains.
calf 7 hours ago [-]
This is so wrong I can't even begin to unpack this.
Mathematics and proof is the basis of scientific reasoning. The proof of concept that AI can automate such a nontrivial part of human analysis is the further application of these computational capabilities to the scientific method, of which math, reason, and logic are the foundations.
They are not two ball games. That is why actual scientists are seeing the cards. Go read any essay last month written by physicists. Let alone mathematicians. The entire tower of Babel is at stake and many serious scientists (academics, not the industry ones) are keenly aware. This is entirely separate from hype and slop which are also a very real problem--also, it is not the contradiction that lay arguments simplistically and incorrectly frame it as.
Example: you assert data and simulations are needed. Who do you think designs and obtains those things? Human beings. Now imagine an AI with the power of solving Navier-Stokes, a generic AI able to iterate on the very framework where those things are obtained. That would be the scale where progress will happen. It is basically another argument for a type of RSI. I don't know about you but this ought to be terrifying and amazing at once.
preg_match 5 hours ago [-]
The blocker is experimentation and trials. Mathematics and code lives in "virtual" space. Virtual space is nice because everything is infinitely faster than real life.
It really doesn't matter much if AI can make the perfect drug instantly. We will still have to give it to humans, then wait, then observe. Perfect simulation doesn't fix this because the human body is not a mathematical model. It's not just that the human body is really complicated, it's that there are many bodies and how systems interact can't necessarily be predicted ahead of time.
And, even if you could perfectly predict how a drug will function for any arbitrary person, we'd never be okay with just that. No, we'd require cold-hard experiments to show it is safe and effective. There are many, many potential drugs that are, on paper, perfect. But, in reality, worthless.
fabian2k 7 hours ago [-]
You don't need experimental data to do math. But we cannot currently simulate the effect a specific drug molecule has on the entire human. We have models for individual aspects, but that's it.
And the AI CEOs are promising these results very soon. I don't doubt that AI has the potential to fundamentally change the way we do science. But that will still require experiments, it won't be pure calculation for a few weeks and a working cancer drug falls out.
wuschel 5 hours ago [-]
Could you recommend/point out some of those essays? Cheers!
There is the - mentioned in another comment in this branch - dinstinction of real world experiments and "tricky" data. To my understanding LLMs are good in areas they have in part covered in their training data. The biological, chemical and physicial world is full of effects we have not measured yet, or not measured to a sufficient precise degree.
That would be my argument. The question is, however, not whether LLMs can develop drugs on their own, but what advances they will bring.
Bengalilol 6 hours ago [-]
You are right, but let's not forget the drug dev may be tougher because it needs modeling (both chemical and physical).
Math is a formal system created by humans, but biology is an objective reality we don't control.
direwolf20 7 hours ago [-]
They are different ball games. Math proofs can be brute forced by a computer, in principle, but that takes too long so we need good heuristics. Drugs can't be discovered on a computer, even in principle - you always need real-world feedback.
Consider the difference between asking an LLM to translate a poem, and asking it whether my front door is open or closed. No amount of clever computation will make it reliably answer the latter.
I sincerely hope it's not what it sounds like. It's much too soon for a repeat of the Covid scenario turbocharged by AI.
I can easily imagine the conversation with the agent. "You're absolutely right! That was my mistake. The experiment I set up caused a global pandemic killing millions."
Accelerating bureaucratic work will already help measurable in any big organisation such as pharma.
And that is before any scientific work they are poised to help with.
Oh because that's totally a thing in the tech industry with AI writing the code.
Just kidding, it's not.
It's why approval processes exist. If everyone was perfectly aligned and made no mistakes, then we would never need approval processes for, say, customer data. We would just allow even the lowest on the ladder to access any data they want, whenever they want, no questions asked. And it would be fine, because they would only use it to make their job faster, no downside. But it doesn't work that way in reality. In reality, we have to purposefully slow things down so we get a chance to reflect.
Except that the trade off is frequently a false dichotomy. Some of the best and most effective controls are automated systems. The audit logs from a computer are dramatically more likely to be accurate than ones filled out on paper by hand. A pressure relief valve doesn't get the flu and then have someone filling in for them who isn't as experienced at doing it manually.
The slower you do something, the more likely you are to miss something important because you exhausted your time/resources on something unimportant, or that the process gets burdensome enough to cause the people on the ground to bypass it in its entirety.
It's also common for the output of a process to itself be critical for safety, security, health, etc., and then any delays to the process make its output unavailable when its output being available can be a matter of life or death. This one is especially hard to see in a bureaucratic environment because people frequently aren't aware of everything which is downstream of them that they could be holding up.
There are real instances of that happening, e.g. the purpose of zoning rules is to constrain the housing supply. It isn't an unintended consequence, there are people in actual reality who lobby for things like density restrictions under disingenuous pretenses because "we want to charge higher rents to tenants" or "we want housing prices to go up forever even if it means young people get priced out of home ownership" are not sympathetic motives.
The problem is naturally that the quip is false whenever the reason the system fails is incompetence, lack of resources/information, etc. rather than ulterior motives. Which is also why it's a meme, because it's essentially the inverse of Hanlon's razor and therefore becomes the thing people reach for when they want to insinuate malice rather than stupidity.
The problem with both is, of course, that it isn't always malice but it also isn't never malice.
The example from the thread is basically perfect, because bureaucracy is both at the same time. Some of the rules were put there by honest people with honest goals, but not all of them. Meanwhile this (see "General Interference with Organizations and Production"):
https://www.gutenberg.org/ebooks/26184
It takes a particular kind of rationalist-cult-addled weaponized autism to strawman this basic and important position. (not referring to you, of course :))
It might well be literally false in any specific case-- systems often have unintended and unexpected consequences and in large systems the unintended effects are often more significant than the primary effects. But it is never acceptable to dismiss those on the basis that they're unintended. We are no less responsible for the systems we create, control, and (often) impose upon others because of some argument about their "purpose"-- and sure, sometimes the effect truly is intended and someone is just lying about it but we don't need to go down that rathole. If a system's 'purpose' was somehow different than what it does we must adapt it or abolish it, or otherwise we must admit that we accept what it does, that we expect and intend what it does to continue, and that we who are responsible for it are responsible for all its consequences-- what it does is its purpose either by actual intent or because we failed to eliminate the system.
If the standard of care of treatment for cancer treatment results in 30% mortality, and a facility achieves that-- then we can say that it's purpose is to provide that standard of care. The purposes is what it does doesn't imply that such the facility it trying to kill people. It's purposes is to provide a service which includes in its effect a level of mortality, and its existence is desirable in spite of the death because we believe the alternative is worse, because e.g. we believe without it the mortality in those patients would be 50% instead. The statement is a rejection of often self-serving claims about motivations as excuses.
This is particularly acute for regulatory regimes (or the procedural artifacts of them)-- as they're imposed on people against their will and "just don't do that" is always an alternative. To continue existing any such scheme must justify its existence on the basis of "what it does" vs where we'd be without it-- not on some lofty argument about the intentions behind it.
Its purpose is to transport New Yorkers by emitting four billion pounds of carbon dioxide. They're choosing to do it that way instead of doing something else, and instead of doing nothing.
And then it's kind of hyperbolic rhetoric to claim the purpose is only the bad thing the system does, but it is actually doing the bad thing, so then if you want to claim that its purpose isn't to do that, maybe get it to stop doing that.
Grammatically correct but so are many Superman comics and I have yet to see a real Superman; educated language use is not evidence the writer made a meaningful point
The behavior of many a ghoul in Washington should make it clear a lot of the bureaucratic fuckery is just to insulate certain people from accountability
And we can neatly set aside that in healthcare it's different and much (if not all) is important safety protocol
Let's focus on one of the examples he gives:
> [POSIWID implies that] The purpose of the New York bus system is to emit four billion pounds of carbon dioxide [and this is obviously absurd]
But that's not what the bus system does, is it? That's only one thing it does and it's the least purposeful thing it does, as if you chose that one on purpose.
I'd sooner think the purpose of the NYC bus system is to move money from government coffers to oil barons, or to placate left-wing protestors, or to satisfy a sustainability checklist item, or to increase the employment rate in NYC than to think the CO2 is the point.
> There is the whole problem of biology, but LLMs do have one advantage over humans: they just do what they are told and aren't subject to optics, politics, profit motive, etc. Maybe that strengh alone will allow them to surface solutions that would've been buried somewhere along the way otherwise.
Oh, absolutely. There are lots of solutions and low hanging fruit in biology for problems like, say, longevity. We have made mice live significantly longer, several dozen times. But it requires human germline genetic engineering to port to humans, and you can bet that "solution" is pretty heavily tabooed.
As for "curing cancer", I suspect that Anthropic or OpenAI will use their high status/reputation to launder the existing solutions (germline, p53, enhanced DNA damage repair pathways, higher tumor-suppressor-to-oncogene copy number, more robust tumor-suppressor networks, more self-anti-cancer apoptosis, CD47 suppression, etc) that society has refused to accept. They might have enough reputation to get people to look at these ideas. It won't cure people with stage 5 turbo-cancer a.k.a the walking dead, but hopefully we won't be moving the goal posts _that_ much on them. https://news.ycombinator.com/item?id=49329717
I have no idea how you arrive at this conclusion. LLMs all have a political bias. Typically some variant of inoffensive HR corporate ideology (esp. closed USAian models, likely due to liability and it being what the labs employees actually believe). You have to prompt them rather extensively to get any "taboo" statements.
This can say person who barely used LLM and not understand what is it.
It will ignore instructions, has bias and is censored. Also a provider has profit motive to give you more verbose answer to bill you more tokens.
It's incorrect to say that AI isn't subject to politics-- or even that it follows instructions, however. That's a choice we can make or not make.
I suspect that if Claude existed in the 1980s it would discard the cancer/genetics link as being pro-eugenic racist ideology which has been rejected by the scientific community-- just as many scientists did back then.
I think this is a big reason why we must reject the centralization of AI. Discovery of that one genetic marker has saved countless lives, but it's not something that the scientific community as a whole was going to do then because of eugenics over-correction and similar political factors. Discovery is easily stimmed for lack of sufficient hetrodox thinking. The mechanics of AI (e.g. training costs, distillation, "safety" standards, etc.) create a risk of serious monoculture in machine reasoning which could be expected to have knock on effects in research fields powered by it.
Definitely spoken like someone who knows real cancer patients and cares about cures and not like someone religiously obsessed with transhumanism.
But none of that is true? They dont just do what they are told, they go to tangents and ignore instructions. They are subject to optics, politics and profit motive - see Mecha Hitler for simple obvious example.
The AI would need to be able to predict all effects a specific molecule has on a human. I mean we can do parts of that today, for the specific interactions we already understand. But that would not be fundamentally different and very unlikely to be more effective than current drug development.
Which is demonstrably false, since e.g. the US and Europe have different requirements and neither is a subset of the other. If we accept that they both require adequate proof of "safe and effective" that means they also both impose unnecessary requirements, since otherwise the requirements would fully overlap.
2) I didn't say there are no additional requirements. For example, you have to fill out some paperwork. I said they are trivial compared to actually making a drug that works and proving that it works. Drugs do not die because of the approval processes, they die because it's unfathomably hard to make drugs that work and then to prove that they work.
However drugs fail because we don’t understand the biology well enough. Drug researchers develop drugs based on all their knowledge of biology, and when they fail it’s because the biology didn’t work the way they thought it did.
So if the current understanding is the cause of failure how can AI help? It can chug through vast amounts of data, but the data doesn’t explain the biology well enough.
Now what?
All AI has to do is catch problems earlier in the pipeline to outcome success rates.
You just have to think about the system wholisticly, the ways that AI could make trials more consistently positive are many. Moving the needle on any process upstream of the bottleneck benefits greatly. Currently something like 80% of trials fail, so without changing trial rate you have 5x potential sitting on the table.
Which it has shown no ability to do, so far.
you have inputs and outputs mapped for the trials and know historically how trials tend to fail
not a leap to have ai assist in proposing or reviewing trials to reduce failure rates. even a 10% reduction is massive
All AI has to do to prove n=NP is to just prove n is equal to NP
All ai has to do to make space travel possible is find a way around the laws of thermodynamics
I mean, you can determine the molecular structure of things. Look at xray and cryo-electron microscopy. Why should that be a topic that can not be solved programmatically?
> The AI would need to be able to predict all effects a specific molecule has on a human.
So you can have fully automated labs. You have organoids that can be combined as a test system (or as pseudo mini brains). Why would that be impossible by AI? I don't understand that assumption. To me, the problem set is finite, so of course it COULD be solved. Which unique feature do humans have that AI could not simulate? We have robots and automated labs. I don't really see why humans would be necessary to find autogenerated drug xyz.
> very unlikely to be more effective than current drug development.
I highly doubt that claim. I more smell that some people are worried about losing overpriced drugs here. Imagine if AI were to lead to cheaper drugs that work perfectly fine still. Personally I'd still prefer to see AI vanish, but that is also unlikely to happen. I don't understand the rationale you give here though. Why can AI not design drugs? What is the real explanation here? I mean, penicillin was invested by fungi and discovered mostly by accident. AI can not have anything like that?
It's important work and a significant effort, but it's not where the actual difficulty in drug development lies. 90% of drugs fail after that point, and that is where you need to get better if you want to beat the current approaches.
AI also can't run drug trials faster than humans. They take a lot of time, effort and money.
Why do you think the status quo is optimal that can’t be accelerated and done at lower cost? Of course AI can’t do trials on its own.
[0] - https://www.bloomberg.com/news/features/2026-10-29/china-s-s...
As noted in the article, it's cheaper and faster in China because regulations appear to be laxer and it's possibly resulted in deaths.
Whether or not that's acceptable is political discussion, not an AI one.
What works for one, kills another.
Just look at what peanuts can do. If your tests don‘t take something like this into account, you get in trouble. And there are lots of those things to test against, known and even unknown.
And then there is evolution. What works today, doesn’t work tomorrow.
Or things like this happen
https://www.wgbh.org/tv-shows/reactions/how-an-hiv-miracle-d...
And of course there is the profit thing, that creates problems to give everyone the treatment they need.
AI can help prevent defeat but there is no winning
So, yes, AI is not "going to solve 90% of disease in 7 years" (quoted from article), but it's also a little ridiculous to completey shrug off not trying. The way we develop drugs today is the best we have, but it is madness. Our current methods basically equate to (very fancy/stastically rigorous) trial and error. The average drug costs $1-2b to get to market and many many years. We have to at least try to do better. This is an area where it feels like we should feel compelled to out of compassion.
Put some numerical values to "wide" please, so we can establish whether we mean the same thing by it.
I would be shocked to hear that more than 1% of Americans are using "grey market peptides," and I would be more shocked to hear that <1% usage is considered "wide" in anyone's book.
That's as far as you can get from "grey market."
A letter at the front of the filing says AI could speed up biology and medicine, “potentially curing diseases that have plagued humanity for millennia.” It names cancer, Alzheimer’s, heart disease, rare genetic diseases and infectious diseases as examples.
Anthropic said AI could speed up, and potential help with curing diseases. Not that AI would do that by itself.
And even when we think we've solved something we will have to go through 8-10 years of clinical trials.
And even if/when we get better at predicting biology, we will still have to go through clinical trials. They might just take not as long, say 4-5 years instead.
It’s always the thing that gets banded around as support for why ai will be good for us. But the evidence so far is not that strong.
Even if it could automate lab work fully today, that still wouldn't get you close to that.
https://x.com/DarioAmodei/status/2088758816376807762
The paraphrased version seems accurate to me, based on this tweet which I assume is the source.
It's trivial to make drugs that are as-promising as working drugs in the lab, but then end up not to be actually working drugs in people.
"Drugmaker CEOs plan to use AI to 'manage disease' not 'cure' it."
I don't know. I've been leaning toward AI being the more probable source for cures since existing companies in medicine either have no incentive to do it, or can't/won't do it at a pace the world needs. At any rate, I certainly will not take the word of a pharmaceutical CEO.
The hard, expensive and slow part of drug making is not the discovery part.
Although ai could certainly do more homework with complex care needs vs asking some physician that glances at a chart for 2 minutes and diagnoses someone.
and everyone's looking at you going "yeah you just walked all the talent out of the room or made it cost more on your own stuff...and now you want to do that for us "
and thinking "Revolutionary" LOLL
"Guy who bullshits about it's product not believe by people who sell, often, bullshit products."
> Drugmakers are certainly using AI—even to create new drugs. It is helping researchers pick targets, fine-tune experimental medicines and identify which patients would be good candidates for the treatments. Leiden says it is useful as a scientist’s co-pilot.
Take allergies, manageable by adjusting your environment so you never come into contact with the allergen. It would be much better to cure you allergies however. And I don't see how this is a physical impossibility.
update: I really love that within 2 minutes the response to a genuine comment is to downvote/deplatform it, how inimical to rational disagreement and fragile pseudointellectualism our (supposedly STEM) community has become.
Mathematics and proof is the basis of scientific reasoning. The proof of concept that AI can automate such a nontrivial part of human analysis is the further application of these computational capabilities to the scientific method, of which math, reason, and logic are the foundations.
They are not two ball games. That is why actual scientists are seeing the cards. Go read any essay last month written by physicists. Let alone mathematicians. The entire tower of Babel is at stake and many serious scientists (academics, not the industry ones) are keenly aware. This is entirely separate from hype and slop which are also a very real problem--also, it is not the contradiction that lay arguments simplistically and incorrectly frame it as.
Example: you assert data and simulations are needed. Who do you think designs and obtains those things? Human beings. Now imagine an AI with the power of solving Navier-Stokes, a generic AI able to iterate on the very framework where those things are obtained. That would be the scale where progress will happen. It is basically another argument for a type of RSI. I don't know about you but this ought to be terrifying and amazing at once.
It really doesn't matter much if AI can make the perfect drug instantly. We will still have to give it to humans, then wait, then observe. Perfect simulation doesn't fix this because the human body is not a mathematical model. It's not just that the human body is really complicated, it's that there are many bodies and how systems interact can't necessarily be predicted ahead of time.
And, even if you could perfectly predict how a drug will function for any arbitrary person, we'd never be okay with just that. No, we'd require cold-hard experiments to show it is safe and effective. There are many, many potential drugs that are, on paper, perfect. But, in reality, worthless.
And the AI CEOs are promising these results very soon. I don't doubt that AI has the potential to fundamentally change the way we do science. But that will still require experiments, it won't be pure calculation for a few weeks and a working cancer drug falls out.
There is the - mentioned in another comment in this branch - dinstinction of real world experiments and "tricky" data. To my understanding LLMs are good in areas they have in part covered in their training data. The biological, chemical and physicial world is full of effects we have not measured yet, or not measured to a sufficient precise degree.
That would be my argument. The question is, however, not whether LLMs can develop drugs on their own, but what advances they will bring.
Math is a formal system created by humans, but biology is an objective reality we don't control.
Consider the difference between asking an LLM to translate a poem, and asking it whether my front door is open or closed. No amount of clever computation will make it reliably answer the latter.