But in this case following security norms would be a mistake. The right thing to take away is that you shouldn't dogmatically follow norms. Sometimes it's better to just build things if there is very little risk
Nothing actually bad happened in this case and probably never will. Maybe some people have their crypto or identity stolen, but probably not a rate rate significantly higher than background (lots of people are using openclaw)
Even if you believe that many are too far on one side now, you have to account for the fact that AI will get better rapidly. If you're not using it now you may end up lacking preparation when it becomes more valuable
But as it gets better, it'll also get easier, be built into existing products you already use, etc. So I wouldn't worry too much about that aspect. If you enjoy tinkering, or really want to dive deep into fundamentals, that's one thing, but I wouldn't worry too much about "learning to use some tool", as fast as things are changing.
I don't think so. That's a good point but the capability has been outpacing people's ability to use it for a while and that will continue.
Put another way, the ability to use AI became an important factor in overall software engineering ability this year, and as the year goes on the gap between the best and worst users or AI will widen faster because the models will outpace the harnesses
That’s the comical understanding being pushed by management in software companies yes. The people who never actually use the tools themselves, but the concept of it. It’s the same AGI nonesense, but dumped down to something they think they can control.
why does every AI skeptic assume that everyone is lying to them. theres millions of developers using AI to be more productive and you just keep plugging your ears and screaming, claiming its only dumb managers, meanwhile Linus Torvalds is vibe coding stuff.
Who said anything about that? The argument was "if you're not using AI RIGHT NOW, you will fall behind forever"
This is the nonsense management and CTOs are pushing. Use it now if you want, I do. Wait for things to cool down if you want. You'll be fine either way. The comical view that it'll be a "winner takes all" subset of developers who some how would have figured out secret AI techniques that make them 10Kx more productive and every other developer will be SOL is laughable.
> Put another way, the ability to use AI became an important factor in overall software engineering ability this year, and as the year goes on the gap between the best and worst users or AI will widen faster because the models will outpace the harnesses
Is it, lol? Know any case where those “the best users of AI” get salary bumps or promotions? Outside of switching to the dedicated AI role that is? So far I see clowns doing triple the work for the same salary.
I mean, right now "bleeding edge" is an autonomous agents system that spends a million dollars making an unbelievably bad browser prototype in a week. Very high effort and the results are jibberish. By the time these sorts of things are actually reliable, they'll be productized single-click installer apps on your network server, with a simple web interface to manage them.
If you just mean, "hey you should learn to use the latest version of Claude Code", sure.
I mean that you should stay up to date and practiced on how to get the most out of models. Using harnesses like Claude code sure, but also knowing their strengths and weaknesses so you can learn when and how to delegate and take on more scope
Okay yeah that's a good middle ground, and I'd even say I agree. It's not about being on the bleeding edge or being a first adopter or anything, but the fact that if you commit to a tool, it's almost always worth spending some time learning how to use it most effectively.
I mean if the capacity has outpaced people's ability to use it, to me that's a good sign that a lot of the future improvements will be making it easier to use.
The baseline, out-of-the-box basic tool level will lift, but so will the more obscure esoteric high-level tools that the better programmers will learn to control, further separating themselves in ability from the people who wait for the lowest common denominator to do their job for them.
Maybe. But so far ime most of the esoteric tools in the AI space are esoteric because they're not very good. When something gets good, it's quickly commoditized.
Until coding systems are truly at human-replacement level, I think I'd always prefer to hire an engineer with strong manual coding skills than one who specializes in vibe coding. It's far easier to teach AI tools to a good coder than to teach coding discipline to a vibe coder.
I wonder if psychology plays a role here. An engineer with strong manual coding skills might be hesitant to admit that a tool has become good enough to warrant less involvement.
It won’t, the state of the art is changing so quickly it is near impossible to stay on top of. Right now claude code is doing stuff for our team that was impossible with ai coding six month ago. Probably a year from now it will be something else. I think that if you are not staying on top of things though, you will discover that you should have stayed more on top of things the day you get fired.
I have noticed a troubling skill atrophy in some people who heavily use LLMs (this is particularly concerning because it renders them incompetent to review ‘their own’ PRs prior to submission). I’m… not keen to sign up for that for no reason, tbh.
How so? Why would a couple of months break in employment (worst case, if I truly become unemployable for some reason until I learn the tools) harm or destroy my career?
Even if the models stopped getting better today, we'd still see many years of improvements from improving harnesses and understanding of how to use them. Most people just talk to their agent, and don't e.g. use sub-agents to make the agent iterate and cross-check outcomes for example. Most people who use AI would see a drastic improvement in outcomes just by experimenting with the "/agents" command in Claude Code (and equivalent elsewhere). Much more so with a well thought out agent framework.
A simple plan -> task breakdown + test plan -> execute -> review -> revise (w/optional loops) pipeline of agents will drastically cut down on the amount of manual intervention needed, but most people jump straight to the execute step, and do that step manually, task by task while babysitting their agent.
Nothing gets worse in computers. Name me one thing. And if the current output quality of LLM stays the same but speed goes up 1000, quality of the generated code can be higher.
Hot keys. Used to be, you could drive a program from the keyboard with hotkeys and macros. No mouse. The function keys did functions. You could drive the interface blindfolded, once you learned it. Speed is another one. Why does VSCode take so long to open? and use so much memory and CPU? it's got a lot of features for a text editor, but it's worse than vim/emacs in a lot of ways.
Boot time.
Understandability. A Z80 processor was a lot more understandable than today's modern CPUs. That's worse.
Complexity. It's great that I can run python on a microcontroller and all, but boring old c was a lot easier to reason about.
Wtf is a typescript. CSS is the fucking worst. Native GUI libraries are so much better but we decided those aren't cool anymore.
Touchscreens. I want physical buttons that my muscle memory can take over and get ingrained in and on. Like an old stick shift car that you have mechanical empathy with. Smartphones are convenient as all hell, but I can't drive mine after a decade like you can a car you know and feel, that has physical levers and knobs and buttons.
Jabber/Pidgin/XMPP. There was a brief moment around 2010? when you didn't have to care what platform someone else was using, you could just text with them on one app. Now I've got a dozen different apps I need to use to talk to all of my friends. Beeper gets it, but they're hamstrung. This is a thing that got worse with computers!
Computers are stupid fast these days! why does it take so long to do everything on my laptop? my mac's spotlight index is broken, so it takes it roughly 4 seconds to query the SQLite database or whatever just so I can open preview.app. I can open a terminal and open it myself in that time!
And yes, these are personal problems, but I have these problems. How did the software get into such a state that it's possible for me to have this problem?
Native GUI libs look like shit out of the box, and are terrible to work with when you want to make something that doesn't look like out of the box tkinter/swing/qt/winforms Windows 95 looking crap.
Software has gotten considerably worse with time. Windows and MacOS are basically in senescence from my point of view. Haven't added a feature I've wanted in years, but manages to make my experience worse year to year anyways.
CPU vulnerability mitigations make my computer slower than when I bought it.
Computers and laptops are increasingly not repairable. So much ewaste is forced on us for profit.
The internet is a corporate controlled prison now. Political actors create fake online accounts to astroturf, manipulate, and influence us.
The increasing cost of memory and GPU make computers no longer affordable.
Doctors here are cognitively captured by a system designed to limit cost (and that's mostly a good thing)
But scanning frequently is overwhelmingly good for the patient. The problem is the doctors. Imagine two possibilities.
1. You scan every six months and a doctor reviews your scans but never tells you anything no matter what
2. You scan every six months and a doctor reviews your scans and only tells you results if you have an obviously growing mass that has a probability greater than 95% of being cancerous
Obviously #2 is better for the patient than #1, but #1 is equivalent to never testing if you ignore cost.
So the actual reason we don't have effect frequent scans combined with effective diagnostic techniques is cost, and doctors cope with this reality by saying clearly wrong things about "over diagnosis". It's a local minimum of the payer/provider dynamic that has nothing to do with scans per se.
Why is it good for the patient? I think that to claim this, you'd need to show a difference in outcomes.
Here, you have a tool with a ~100% false positive rate, so if we start administering it to everyone, it will almost certainly cost lives. Botched biopsies, unnecessary treatments, other complications. Not to mention the huge cost that would divert money from other welfare programs. So you need to show that when it actually detects something, it saves at least as many lives. And I doubt that's the case.
I wouldn't argue we should roll this out to everyone. But I am glad it exists. I commented earlier in this topic about how it caught cancer in my wife at the age of 44. She didn't have to go through chemo or radiation treatment because it was caught so early. Surgery removed the whole cancer.
Additionally for me, I have a scan that shows what my body currently has. I had something show up that I did get a scope to check out that was a pancreatic rest. No big deal. Now, if I ever have another MRI and somethings is somewhere else, we have a baseline to compare against. Everything is a risk calculation. When I did my MRI, I also had other procedures done like a heart calcium score.
I will get a little more personal. We didn't do it out of the blue. My wife and I decided we want to live on a sailboat. That was a big purchase for us and boats take a long time to sell. We didn't want to commit to such a purchase then 1 year later find out either of us had cancer then we have the stress of cancer and the stress of trying to sell a boat.
I would never suggest everyone do it, but I am happy we did.
I’ll give you a counter example. I had an MRI of my neck for unrelated reasons. It found a thyroid nodule with suspicious characteristics. Incidentally I had had an MRI of the same areaa few years before and it wasn’t there.
So I had a biopsy. Which was equivocal also.
So I had it out which involved removing half my thyroid. Turns out it was a cancer but like the least serious kind, in fact the classification of it as actual cancer has gone back and forth over the years
But my other half of my thyroid couldn’t produce enough thyroid hormone, and now I have to take thyroid replacement the rest of my life to start alive
Also the surgery affected my voice and I sound like RFK jr now.
I clearly suffered some harm, and even after having the thing out, it’s unclear if that was beneficial at all. A large proportion of these kind of tumors quit growing and never do anything bad. But some do. So who knows.
No. To argue for the benefit of the procedure, you need to show a difference in outcomes. Not that it can detect something, even if it could (which whole-body MRIs clearly don't). That the detection improves your chances of survival.
If you have an growing mass in your body, then if it's cancer, after a year, it might be too late for treatment. Or it may turn out to be nothing: a benign tumor / cyst / fat deposit in an unusual place. Or it may be slow-growing prostate cancer that you can live with for another 20 years, and maybe it's the chemotherapy that will do you in. It's really not that clear-cut in medicine.
To give you have another example: let's say that the risk of appendicitis in people who have an appendix is 1%. And the risk in people who had an appendix removed is 0%. Does this justify proactively removing the appendix? No, because the consequences of complications are much higher than the harm you're preventing. The same applies here: detection, even if 100% accurate, doesn't mean anything. You need to show that what you do with the result actually helps.
With my change: 95% of people who are shown scans have cancer and are treated earlier. 5% of people do not have cancer and get CT scans. 0.5% of people get useless biopsies
Without my change: many of those 95% die, the 0.5% do not get useless biopsies
And the beauty of this is you can pick the percentage!
> If you have an growing mass in your body, then after a year, it may very well no longer make a difference whether you treat it or not. Or it may be that you would have lived another 20 years just fine
This is just wrong for many parts of the body. In your brain? Your lungs? Growing for a year between 3 scans 6 months apart? Extremely unlikely to be benign
> The same applies here: detection, even if 100% accurate, doesn't mean anything. You need to show that what you do with the result actually helps.
This is wrong. If you had a 100% accurate cancer detector, fewer people would die of cancer with no downside
> With my change: 95% of people who are shown scans have cancer and are treated earlier. Without my change: many of those 95% die
Why? What happens if the cancer still doesn't respond to treatment even when detected early? Or, to the contrary, if the cancer also responds to treatment when it starts becoming symptomatic?
That's why we have studies to understand if screening is a good practice or not. It's not that clear cut.
> With my change: 95% of people who are shown scans have cancer and are treated earlier. 5% of people do not have cancer and get CT scans. 0.5% of people get useless biopsies Without my change: many of those 95% die, the 0.5% do not get useless biopsies
You assume that treating cancer automatically improves the outcome. Treating cancer often kills you, so treating a non-fatal tumor can easily be a bad decision. And a lot of the tumors found by agressive scans are like that, but we don't know yet how much exactly and how to tell one from the other. It's a new question that requires decades-long observations to answer.
> This is wrong. If you had a 100% accurate cancer detector, fewer people would die of cancer with no downside
You're saying it as if detection somehow cures cancer, it doesn't.
> You're saying it as if detection somehow cures cancer, it doesn't.
No, I didn't say the detector would cause cancer to be cured.
I said fewer people would die with no downsides. If treatment is sometimes harmful then the detector also fixes that, you'd never treat people without cancer
No, the detector doesn't fix that, that problem is not treating people without cancer. The problem is treating people with cancer that won't kill or harm them during their lifetime. In this case even a low risk treatment becomes harmful, let alone cancer treatments.
Whether treatment is net harmful or not depends on the level of the risk with no treatment. If you apply treatment with 15% chance of severe side effects to a tumor that will kill the patient with 50% chance in the next five years, of course it's net positive. If you apply it to a first-stage cancer that has 10% chance of progressing to the second stage, the very same treatment will be net harmful.
So no, most cancer treatments aren't net harmful now, but they likely will be in a world where your programme is implemented. Even something as "simple" as biopsy has mortality rates far from zero. Applying it at scale may not have the effect you expect. And surgeries and chemo are much worse.
But do you see how crazy that sounds? If we know the numbers, we would just not do the treatment in those cases.
And in reality it's not actually close like this. Early treatment is so overwhelmingly better than it completely dwarfs all considerations of biopsy risk.
Late stage lung cancer has less than half the 5 survival of early stage, and around 50% of lung cancer detection is late stage. That's around 30000 lives you could save every year from just lung cancer.
Let's say you don't protocol for all adults. Typically post CT or PET at most 10% get a biopsy. So let's say we increase the false positives that lead to CT or PET by X% of adults. That's 270M * X additional biopsies. Lung biopsy mortality is under 2%.
So the worst case you need a false positive rate under 5% for this to he net beneficial for lung cancer. That's low but with 3 scans 6 months apart and a good radiologist, it's not unreasonable
I believe I've assumed the worst case for all the quantities. In reality the addition screened people would have much lower CT detection rate than the background population, and biopsy mortality has been decreasing. Plus you can do this for only people above 40 where the benefits are higher
Yes, IF we know the numbers, that's the whole problem. Unprompted full-body MRI scans turn up a completely new probability distribution. We CAN'T know the numbers as of today, it takes decades of controlled studies to get them. (Which is exactly what doctors advocate for)
Even for something as studied as colon cancer it's still controversial whether mass colonoscopies are better than occult blood tests. US sticks to the former, Europe to latter. US have higher rates of polyp removal, higher rates of cancer detection, higher rates of surgeries and ... higher mortality rates than EU. Why this happens is being studied today, there are some early results. What is absolutely clear though is that higher detection rates don't automatically lead to better outcomes, it can easily be the opposite.
Or take something as simple as sunscreen. It's well-known that sun exposure for light-skinned people can cause cancer and sunscreen is advised. A few multi-decade studies have shown that while it indeed eradicates the skin cancer it also shortens the life expectancy by a few years.
Primum non nocere, doctors and scientists are unsure whether sunscreen should be applied at scale and you effectively call for mass biopsies and surgeries.
So if this is true, it seems that we must accept that many people will die of cancer we could have detected and cured with frequent scans, because doing frequent scans will overall cause more harm to people who didn't need treatment. So the overall death/harm rate would be worst with more frequent scans?
Isn't that then just a problem with the scan and diagnosis? With more frequent scans it seems highly unlikely that we wouldn't improve this process and end up in a better place.
When there is low prevalence of a condition, but a non-zero false positive rate of a test, the false positives generated by universal testing can in fact be a net dis-benefit (worry, invasive further procedures, etc) to the patient population as a whole, regardless of cost. This is a well understood statistical phenomenon, and is carefully considered by healthcare systems when advising on testing.
As if there was an easy, foolproof and precise way to calculate these probabilities. "Just only alter the patient when it is appropriate" he said. You can solve all the world's problems this way, just "always do the right thing". Reality, unfortunately, is more complicated.
What if in reality the doctors can only say if the probability is above 60% or not? What if some doctors are better than others at estimating probabilities? What if estimates are influenced by financial reasons by some entity like the hospital or insurer?
> What if estimates are influenced by financial reasons by some entity like the hospital or insurer?
It's this one
It doesn't matter that some cancers are hard to diagnose. What matters is that some are not hard, especially if you have multiple scans spanning several years
Did you know optometrists now scan your retina during ordinary eyeglasses appointments? They didn't do this in the past because it was too expensive. But optometrists cannot diagnose or treat based on these scans. They refer you to an opthamologist.
Some day your PCP will do the same with MRIs, but only once they are cheap. For now only rich people have this luxury
Tons of diagnostic interventions have been scaled without strong outcome data. For example many clinics now do fractional exhaled nitric oxide (FeNO) tests because they are safe, fairly cheap, and patients often pay directly out of pocket so they are easy to make money on.
But the evidence for the diagnostic usefulness of this test is extremely low, multiple meta analysis have concluded.
The reason FeNO tests are done but not MRIs is because FeNO tests are $40
Your case #2 doesn't have nearly enough information to say whether it's obviously better for a population of patients. There are a lot of other variables you would need to know:
- The accuracy of detecting a mass
- The true distribution of masses in the population
- The likelihood that of falsely detecting a mass in the same place twice (you seem to implicitly assume that false detections are uncorrelated with each other)
- The likelihood that a real mass is cancerous (you stipulate that this is 95% in your scenario, but you don't say what other factors are used to determine this - as opposed to just knowing that there's a mass that grew.)
- The positive effect of treatment in the case of true-positives.
- The negative effect of treatment or further diagnostics in the case of false-positive.
Saying that doctors are lying about over diagnosis to cope with the fact that diagnostic techniques are too expensive is absurd. They have to actually make decisions in the real world, where your two neat little categories can't be known even if they hypothetically exist.
No? The point of the article, and of the preceding comments, echoing a pretty common tenet of evidence-based medicine, is that frequent full-body MRIs are a bad idea for the patient.
This guy has never heard the term 'scanxiety'. Go ask what it means on a cancer forum. The real OG's are the VHL folks. Bet we have a few here on this thread. Respect.
I have, it's the fault of how medicine is practiced to reduce cost. It's completely avoidable, you can just not tell people their scan results if they have no symptoms and the detection is less than 95% likely to be cancer. This is strictly better than the status quo because the only difference is some people who almost certainly have cancer learn that they have cancer and nothing else changes
Right, the problem in your model is that the doctors do! I am pointing out that this is a problem with the doctors, not the scan.
Scans are a tool, doctors are not allowed to use them rationally because it would be too expensive, so they don't use them. With an ideal doctor, patient outcomes would be better with a scan than without one, but my claim is that doctors are not ideal.
No doctor would order a full body MRI just to throw out the result in 99% of cases, because *it's too costly*
I agree, but they are failing to order scans when the probability of finding anything worth biopsying is low because of cost.
This is so obvious it's crazy, our entire world view is misshaped around saving money.
When you walk into a doctor's office, why do they take your blood pressure and temperature? Why do they look at the back of your throat and listen to you breath?
These are all diagnostic procedures with extremely high false positive rates. The reason doctors do these but do not do MRIs is because MRIs are expensive.
If MRIs were free you would get one automatically every time you go to the doctor.
It’s a great document, I’m an MR tech and we now have something to lean on when we say no to these scans. We can then scan people with problems rather than people with too much money.
Great document? I just read the entire thing, it contains no evidence or justification for their claims.
> We can then scan people with problems rather than people with too much money.
Right, this is what it's really about. MRIs are a scare resource and providers need to manage cost. Fortunately I can afford to pay out of pocket, but I'm just annoyed that doctors are so irrational about this
To be clear, it's up to a doctor whether or not to do a "needless biopsy". That has nothing to do with a scan.
You're imagining that the doctor is required to act irrationally or against the patient's interest. I understand that they do act this way, but I'm pointing out that they don't have to!
You walk into a doctor's office and meet for the first time.
Case 1. You have 3 full body MRIs taken 6 months apart in hand and give them to your doctor. Nobody has opened or looked at the scans yet.
Case 2. You do not have a scan yet
You are claiming case 1 will lead to worse outcomes in expectation. I claim that's impossible because the doctor can simply do the following:
Without opening the scans, email them to the world's best radiologist. Tell that radiologist "only reply if it is nearly certain the patient has an operable cancer with the information available"
Now there are 4 possibilities.
Case 1 without cancer: nothing happens
Case 2 without cancer: nothing happens
Case 1 with cancer: you maybe survive
Case 2 with cancer: you die
The only reasonable objection to this is "that's expensive"
A whole body MRI is low resolution and thick slices, they are crap.
Instead you could get thin slices at high resolution of a body site that has issues or is suspected to have them. Do that instead.
Whole body MR is 5-8mm thick slices at low in-plane resolution. A whole body scan has about 512 pixels over a 50-60cm field of view. Usually it’s even less pixels than this.
Something like a knee, or brain is 2-3mm slices, and high in-plane resolution. A knee is 512 or even 720 pixels over 14cm. It’s vastly better. The difference is stark.
A liver scan or other abdominal organ is lower resolution than a joint or brain, but unlike whole body MR is scanned in multiple planes with multiple image weighting (t1 in/out/fat sat, diffusion, t2, t2fs, gadolinium contrast). A liver scan has thin slices.
Whole body scans generally do two coronal images (stir and t1) then call it quits.
Yes the machines should definitely get better, but I recommend you do get it if the cost it truly zero. Based on the data in the article it is still worth it despite the low resolution!
Sorry, I edited my above comment to remove references to myself.
For clarity, I’m an MR tech and I can get scans of myself if I want to.
I generally don’t scan myself as it gets messy fast. If I had concerns and for some reason couldn’t get a proper imaging referral, I’d get a scan with small, good coils with high element counts (not body coils like whole body imaging uses) and scan individuals body regions.
It is not only money, machines and doctors are also a limited resource.
The other bit is that no one follows your protocol, everyone looks at the scans, finds reasons to worry.
Then you have the choice between unnecessary biopsies and psychological stress.
I kinda agree with you and I understand your point but I also think there is a social-ethical reality that if a doctor finds something they must treat it. The two options are 1. doing nothing, or 2. reducing cancer risk and you get unnecessary biopsies.
You're thinking of this as there being an objective positive utility for not dying of cancer and a objective negative utility for biopsies, and there being an objective optimal "rational" tradeoff that the best radiologist can optimize for to get their "nearly certain" detection threshold.
But frankly - the tradeoff for the value of a human life is perhaps the most uncertain thing one could choose. It lies in the eyes of the patient if the worry and time associated with a false detection is worth their reduced chance of death. The ethical uncertainty expressed in the OP - are these unnecessary biopsies worth it - is warranted.
You a-priori do not know if you have cancer. The case 1 branches are:
Case 1 without cancer: you're a guy that knows more than doctors because you're the "I have 3 scans on hand guy", so as soon as you open the scans and the doctor says there's a mass, you will say 'OK what can we do???' And then you had a worthless biopsy.
Case 1 with cancer: you saved the 1 week it would take to schedule the new scan and get results and you're basically in the same situation except one week earlier.
That's not how it works. Without the history of scans spaced out in time, they would typically need a biopsy. Multiple scans over time shows change and growth and improves the likelihood of an accurate triage
And for the first part that doesn't happen in my scenario before the doctor doesn't look at the scans
This entire chain of reasoning takes for granted that there won't be a singularity
If you're talking about "reforming society", you are really not getting it. There won't be society, there won't be earth, there won't be anything like what you understand today. If you believe that a singularity will happen, the only rational things to do are to stop it or make sure it somehow does not cause human extinction. "Reforming society" is not meaningful
But unfortunately it does sometimes work, for example in Yugoslavia. And it would have worked in Iraq if we hadn't dismantled the entire civilian infrastructure.
The killing in Bosnia and Kosovo were stopped by Bill Clinton. The bombings were what brought all sides to the table to broker the Dayton agreement. The siege of Sarajevo ended. The peace has held for nearly 30 years now.
Seems clear they were talking about NATO intervention in 1999 and the ouster of Milosevic. Also notable as a military intervention that was at the time widely seen as a "Wag The Dog" scenario.
You said "like that", ok but there may be some truth to reduced model intelligence. Also how AWS deployed Anthropic models for Amazons Kiro feel much dumber than those controlled entirely by Anthropic. Can't be just me
reply