Don't see too many cars on walking paths and sidewalks. The number of e-bikes on walking paths and sidewalks has skyrocketed. It's almost as if someone decided being a pedestrian is a sinful activity, and that every walkway must now be infested with morons on wheels.

Then let me get started on mobility scooters.

I'd just like them banned from walking paths. At least once a day I'm getting some crazy asshole ringing his bell as he comes flying up behind me. I'm not a fan of any kind of bike on walking paths, but at least the people on regular bikes have more control. The worst are probably older riders who often seem like they're barely in control. And the three wheeled ones take up outrageous amounts of space on smaller paths, regularly forcing other users on some of the narrower paths I frequent to get to the side of the road.

It's hard to imagine, short of motor vehicles, anything more hazardous to a pedestrian than some stupid prick on an e-bike.

What you describe is essentially a form of bootstrapping, which is a legitimate statistical method. However, there are important limitations that cannot be overlooked.

First, the constructed data are still being created from real data. Ethics is not just about preserving patient privacy, although that is a very important aspect. It's also about taking into consideration how the data will be used. Does the patient consent to this use, and if they are unable to consent, how should this be taken into consideration? Medical science has not had a stellar track record with respect to ethical human experimentation (e.g., Henrietta Lacks, the Tuskegee syphilis study, MKUltra--and that's just in recent US history). There is a documented history of patient collected data being used in ways that those patients never even conceived, let alone anticipated or consented. Caution must be exercised whenever any such data is used, even indirectly.

Second, this kind of simulated data is problematic to analyze from a statistical perspective, and any biostatistician should be aware of this: there is no such thing as a free lunch. The problem of missing data--in actual patients!--is itself difficult to address, since methods to deal with missingness invariably rely on various strong assumptions about the nature of that missingness. So to make inferences on data that is entirely simulated is, at the very least, as problematic as analyzing partially missing data.

Third, the current state of LLMs, and their demonstrated tendency to distort or invent features from noise (which is arguably the primary mechanism by which they operate), is such that any inferences from LLM-generated data would be questionable and should not be considered statistically meaningful. It could be used for hypothesis generation, but it would not satisfy any kind of statistical review.

It all comes back to what I said in another comment: you can't have it both ways. If you can draw some statistically meaningful conclusion from the data, then that data came from real-world patients and must pass ethical review. If you don't need ethical review because the data didn't come from any real patient, then any inferences are dubious at best, and are most likely just fabrications that cannot pass confirmatory analysis.

The purported claim is that "because the AI-generated data do not include data from actual humans, they do not need ethics review to use."

But if the data only represent actual patients in a "statistical" sense (whatever that means), how can the research be CERTAIN that it has captured appropriate signals or effects that are observed in such data? And I say this as a statistician who has over a decade of experience in statistical analysis of clinical trials.

There is a fundamental principle at work here, one that researchers cannot take the better part of both ways of the argument: any meaningful inference must be drawn on real world data, and if such data is taken from humans, it must pass an ethics board review. If one argues that AI-generated data doesn't need the latter because it is a fabrication, then it doesn't meet the standard for meaningful inference. If one argues that it does meet the standard, then no matter how the data was transformed from real-world patient sources, it requires ethics board review.

In biostatistics, we use models to analyze data to detect potential effects, draw hypotheses or make predictions, and test those hypotheses to make probabilistic statements--i.e., statistical inferences--about the validity of those hypotheses. This is done within a framework that obeys mathematical truth, so that as long as certain assumptions about the data are met, the results are meaningful. But what "statistically naive" people consistently fail to appreciate, especially in their frenzy to "leverage" AI everywhere, is that those assumptions are PRETTY FUCKING IMPORTANT and using an LLM to generate "new" data from existing, real-world data, is like making repeated photocopies of an original--placing one model on top of another model. LLMs will invent signals where none originally existed. LLMs will fail to capture signals where one actually existed.

And it may prove to be the most prophetic of all ST:TOS episodes. Unfortunately, the C-suite is currently jacking off to it. "No, don't let that pointy-eared freak control the guard's mind! And pass my cocaine!"

Climate change will render all of that moot. It is the one thing that not even bloviating halfwits can alter. You see, people are morons.

Just remember that the laws of physics don't give a fuck, and will crush us with as much ease and as little concern as a tidal wave crushing an ant.

That we imagine we can alter physics by denial is a sign of what an utterly idiotic species we are.

I do very much understand what you're saying and it certainly adds to the complexity. One cannot put sociological or psychological factors on a box.

That aspect of the problem is indeed going to be much harder to deal with than, say, salt, trans fats, or known carcinogenic compounds.

Honestly, I'm not sure what you can do about those aspects - financial incentives help a little, but honestly I don't believe they make a huge difference - which is why I've concentrated on unsafe levels of ingredients, because although we don't know exactly what those should be, we've at least got a rough idea for some of them. It's going to be a delicate one, though -- you don't want to overly restrict sources of sugar because diabetics can suffer from crashes due to excessively low sugar just as badly as excessively high levels, and some items get unfairly maligned (chocolate, per se, isn't bad for you, it's the additives, and indeed particularly high percentage chocolate can be helpful for the heart).

But, yes, I absolutely agree with your overarching point that the problems are primarily psychological and sociological. I just don't have the faintest idea of how these can be tackled. Jamie Oliver tried (albeit not very well, but he did at least try) and the pushback was borderline nuclear, and that was where there was clear and compelling evidence of significant difference in health and functionality. If you can barely escape with your life for saying eating better reduces sickness and improve concentration, and pushing for changes where these two factors essentially dictate whether a person is functional in life, then I don't hold out hope for change where it's more ambiguous or the economics are much tougher.

I'm going to go out on a limb and suggest her statement was fair and accurate.

There are papers arguing that smoothies aren't as good as eating real fruit because it seems that there's actually a benefit to having to break down cell walls, even at the expense of not getting 100% of the nutrients from it. However, cooking food breaks down cell walls, although obviously not to the same degree. It's not clear that breaking down cell walls is harmful, even if it's not beneficial.

A lot of ultra-processed foods have been accused of having unhealthy levels of certain ingredients (usually sugars or salt) and certain styles of cooking can add harmful compounds.

It would seem reasonable to say that there's a band at which a given ingredient is beneficial (analogous to a therapeutic threshold), with levels above that being increasingly harmful, eventually reaching a recognised toxic threshold. In terms of the harmful compounds from cooking, it seems reasonable to suggest that, below a certain level, the body's mechanisms can handle them without any issue, that it's only above that that there's any kind of problem.

So it would seem that we've got three factors - processing that can decrease benefits, ingredients that follow a curve that reaches a maximum before plunging, and processing that can increase harm.

Nobody wants to be given a complicated code that they need to look up, but it would seem reasonable that you can give a food a score out of three, where it would get 3 if you get maximum benefit and no harm, where you then subtract for reduced benefit and increased harm. That shouldn't be too hard for consumers, most people can count to 3.

Yeah, understood, food is going to vary, since it's all uncontrolled ingredients and processing itself is very uncontrolled. So take two or three examples as a fair "representative sample". Further, most manufacturers can't afford to do the kind of testing needed, and our understanding of harm varies with time. No problem. Give a guidebook, updated maybe once every couple of years, on how to estimate a value, which can be used, but require them to use a measured value if measured, where the value is marked E or M depending on whether it's estimated or measured.

It's not perfect, it's arguably not terribly precise (since there's no way to indicate how much a food item is going to vary), and it's certainly not an indication of any "absolute truth" (as we don't know how beneficial or harmful quite a few things are, food science is horribly inexact), but it has to be better than the current system because - quite honestly - it would be hard to be worse than the current system.

But it's simple enough to be understandable and should be much less prone to really bizarre outcomes.

A pretty significant proportion of the world's IT infrastructure runs on single source products; from Microsoft to Broadcom to Oracle. Believe me, I'm trying to move our office over to open source where I can, but it's no mean feat.

The real lesson, unless the courts start holding licensors accountable, is that a perpetual license may actually be meaningless, no matter what you paid for it.

Too bad. Your shitty political system, with its impotent checks and balances and imperial powers handed over the to the President, made this possible. Your all frogs boiling in the Framers' failed experiment.

I have zero sympathy at this point. Supposedly the Framers put in a really top tier solution to tyranny, but instead it's mainly being used to take out schoolchildren. Everything about your country has become repulsive and self-destructive. I just hope my country can hang on.

The voted for someone who was unbelievably awful in their first term, knowing they were older and even worse. So I hope Americans suffer horribly. They have earned that.

Because Brazil is unique in creating the conditions of its own devastation...