evanwise

I would hope this results in criminal charges and jail time. Honestly, even if it's by negligence and not deliberate. Fines are not enough to abate this kind of despicable behavior.

The only possible reason for barriers to entry is regulation? There are many industries where startup costs are high for reasons intrinsic to the business and margins are relatively low so you are unlikely to get investment from outside sources. Grocery stores for example. You are trying to apply a toy model from econ 101 to explain the behavior of complex real world markets, and when the model doesn't fit you invent bogeymen to blame. This is more like religious fundamentalism than any kind of science.

What was the controversy?

Every post that is even remotely critical of ChatGPT has someone posting this exact sentiment like clockwork. It's probably just mindless fanboys, but I'm genuinely starting to wonder if it's some kind of astroturfed ad campaign for OpenAI.

Worth pointing out that sometimes things really are impossible. If an MBA tells you something is impossible, it is probably just really hard. If a physicist tells you your idea violates conservation of energy, it is probably not worth wasting your life chasing it.

There should be criminal liability not just monetary.

This is a really fun use of language models. I wonder if more capable models would be better or worse in this kind of application? To some extent the charm comes from the whimsical aspect of it.

Did you read the article or just the title? They mention the specific models the researchers were testing and note that increasing model size did not seem to offer much improvement on this metric. It also ends with a discussion of research into methods for improving performance on queries involving negation.

Even if we accept their motivations at face value (which is naive), it's a totally incoherent morality. There are infinitely many future states with trillions of of lives. Which one is the one we attach moral value to? If it's all of them, then every single decision you make is equivalent to killing infinitely many future people. If it's one particular future state how do we decide which one? (I am not, of course, saying that you shouldn't value making the future better, just that applying this specific kind of utilitarian calculus to future humans is childish nonsense.)

It gets reposted because it confirms the biases of right-wing posters and is an easy way for them to farm positive engagement and feel like they have some special insight into current events.

If the OpenAI is using a machine learning system to flag comments as hateful there is a simple and parsimonious explanation: it simply had more examples of hate speech directed at minority groups and people with left-leaning politics. Interesting that the author never considers this and instead implies some nefarious political intent on the part of OpenAI.

I'm all for people pursuing amateur chemistry if they want to, but they should be doing so with accurate resources. The Anarchist Cookbook is not reckless. It's just straight up wrong in a lot of places in ways that very well could get you killed. Making homemade explosives is reckless, doing it poorly is just stupid. If you want to make explosives there are better guides on YouTube of all places. There is no reason at all to use this dated book full of inaccuracies.

I would tend to agree. The only people I saw really struggle with learning Fourier series and the Fourier transform were people who had struggled with more basic concepts and failed to internalize them. This kind of explanation might be helpful to build someone's intuition up a bit but is not a substitute for a proper education on the subject and definitely not some magic key to understanding.

The methodology they describe seems like it would be prone to the multiple comparisons problem. Time to dredge up the actual study and see what's going on.

I hope this is sarcasm.

While it's true that some of the simpler recipes will work, a lot of the others just won't work or are dangerous. From memory, the TNT and mercury fulminate recipes are wildly dangerous and omit key steps. Messing with stuff like this in any capacity is dangerous, even with proper PPE and good procedures, but if you must, there are much better resources available freely online these days.

Please, for the love of God, don't follow any of the recipes in here. Many are dangerously wrong (as in blow yourself and your neighbors up dangerous).

To be fair, sometimes a single idea is difficult enough that it deserves an entire book.

Nothing in this contradicts what I said. Prisoners have a right to informed consent which can be removed in certain circumstances, specifically to ensure the safety of staff or other inmates. You made a much broader categorical claim that prisoners simply do not have the right to informed consent and you are wrong. For example, here is a document from Tennessee outlining the scenarios in which an inmate can and cannot refuse care: https://www.tn.gov/content/dam/tn/correction/documents/113-5...

I'd suggest doing some actual research into the relevant case law, e.g. Washington v Harper, instead of just googling until you find something that can be misinterpreted to support your position. While prison administrators have wide leeway to force treatment they are still required to show that it's necessary, either for safety reasons or because the treatment was mandated as part of the sentence of the inmate.

The irony of complaining about cherry picking statistics while yourself cherry picking statistics is astounding. If you combine federal and state prisons about 40% of prisoners are reported to be in prison for violent offenses (in 2019 according to statistics prepared by the Bureau of Justice Statistics). So, yes, the majority of prisoners are in fact imprisoned for non-violent offenses. But this is not the end of the story, there is no universal standard for what is considered a "violent crime" so states are free to set their own standards (often under considerable political pressure from the local legislature). Some states consider any theft involving drugs or even embezzlement as violent crimes. Beyond this, these statistics are collected and maintained by the same institutions which have a direct monetary incentive to over report violent crime in order to obtain increased funding. This is a clear conflict of interest.

But even if violent offenders were the majority, violent offenders in general are, perhaps surprisingly, less likely to reoffend. This is because the best predictors of violence in the general population are age and secondarily gender. The overwhelming amount of violent crime is committed by male adolescents and young men. While we may be fascinated with macabre cases where serial killers and other extreme offenders were released, these are outliers. We should not base policy decisions on our emotional responses to crime but on measurable outcomes.

Verbal communication may be the default option but there is a reason we often defer to (carefully constructed) written communication in complex domains. While emotional nuances may be clearer (to some) in verbal communication, the details of complex ideas often are not. Have you ever tried to "talk code" at someone without the aid of written code? It's woefully inefficient and prone to miscommunication. Writing allows you to externalize state and forces you to make explicit unstated assumptions that may simply go uncommunicated in verbal exchanges. Indeed, in your examples, the screen share on which the code can be read by both parties seems to be doing most of the clarifying work, that is, you are cheating and using written communication anyway.

I said most people learn to speak before they learn to write. Is that untrue?

This is true but completely irrelevant. Most people also learn how to dance before they learn how to write, that does not mean interpretive dance is a suitable medium for conveying technical information.

How about the idea that you can communicate (crudely, sure) with gestures even with people you don't share a common language with?

This is an edge case, and again, kind of cheating as you are attributing the benefits of visual aids to the verbal format. It does not provide any evidence that spoken communication is universally better.

How about the idea that both writing comprehensively and reading comprehension are high level skills that many people don't possess enough to truly communicate effectively via writing? Do you disagree?

Reading comprehension and the ability to write cogently are basic skills of any knowledge work. I think people who are poor communicators are probably poor communicators regardless of medium, so this is a red herring. In general, the things you are saying are true to some extent but do not constitute an argument for verbal communication being universally better than written (your claim). Rather, even if I am charitable and ignore the clearly fallacious parts of your argument, at most you've shown that in some circumstances verbal communication has some advantages (a much weaker claim which does indeed seem rather uncontroversial).

I don't think it's unfair to say that verbal communication is a cornerstone of the (near) universal human experience.

This isn't just "my preference" is what I'm getting at. I'm pretty sure it's just the human default.

So, those of us with hearing problems or speech impediments are simply inhuman? This is the arrogance I was talking about. Your assumption that your preferences and the way that you work most efficiently is universal when it is clearly not. Again, you are free to conduct yourself in this fashion but it won't make you many friends.

At this point it's worth pointing out that this user has conspicuously ignored this and other replies that point out that, beyond any moral qualms, the user is woefully uninformed about the medical rights of prisoners. Dunning-Kruger strikes again.

He's also just talking out of his ass. Prisoners have informed consent rights just like you and I do.

Well, not only are you evil, even worse: you're wrong. The only scenario in which the state can compel you to receive medical treatment while in prison is to ensure the safety of staff or other inmates (e.g. they can force a violent inmate to take anti-psychotics or force inmates to receive vaccinations to prevent the spread of disease). Otherwise, prisoners have the same rights to informed consent as you and I do.

A root canal is also a fairly safe procedure but I would be rightfully upset if it was done to me without my consent.

I was giving you the benefit of the doubt in the previous reply but this is so arrogant it's quite silly. If you want to force your preferences on others be my guest but don't be surprised when some people respond negatively. Verbal communication is often imprecise and vague. I can't count the number of times I've had to write up a detailed summary of a meeting to ensure that two different stakeholders didn't just take away what they wanted to hear from the conversation. There have been many times where having a paper trail of my conversations has me allowed to prove that, yes, I did indeed inform so and so about that problem weeks ago. There are also many people, myself included, who have difficulty processing and producing spoken language (either through speech and hearing impairments or ASD or any number of other conditions). Your preference is just that, a preference, not a universal truth.

I can understand the gripe about people who don't respond promptly during a chat and I too sometimes just call people like that, but in my experience much of the time when someone wants to "hop on a call" they essentially want me to hold their hand through something that can be easily and clearly explained in a few sentences of text (maybe with a link to an article). It seems to be more about personal comfort than any efficiencies in communication. Which is fine ultimately, but it should be recognized as a preference rather than some panacea for communication woes.

Yes, that kind of arrogance can definitely come from the front-end side as well. I've just noticed a pattern with a certain type of back-end developer. Similar to the old XKCD physicist comic, there is a personality type that seems to think they can spitball an ad hoc solution to a problem that is better than the work of specialists in the field.

Having worked on both sides of the divide, I can say it's sadly common for back-end devs to underestimate the complexity of front-end development and discount the advice and opinions of their front-end colleagues. I've even seen scenarios where back-end devs were allowed to dictate standards to their front-end peers just because they had more seniority in the company.

Another situation where geometric algebra makes everything clearer.