Well, it is arguable that it is "laughable" that there is a strong push to try and regulate/contain these systems to avoid human extinction when they can't tell you the number of weekdays in a month (and fail in this manner).
I'm no AI skeptic (and am aware the above does not seem to apply to GPT4), but the general point stands - finding these limitations is an interesting endeavor, and when they're particularly severe or unexpected, newsworthy.
Uhm - maybe train a secondary NN that scores summaries on their factual accurateness/quality? Anything under a given threshold is either sent for manual review or re-ran through the LLM until it passes.
I think it's interesting to see the ease with which a few commenters on this thread assume they either 1) know how to answer this question or 2) are comfortable hand-waving it away.
I don't get what point you're trying to get across with this comment. The question at hand is essentially one of causality, but you seem to be avoiding it? Addressing your points one by one:
>Survivor bias. If nothing existed [1], we wouldn't be here to ask the question.
Survivor bias leading to the question is not the reason anything exists, in the same way you being alive to ask this question is not the reason you're alive. In both cases existence is simply a pre-requisite to asking the question, but it does not answer it.
>It's like asking what is P(A|A).
No, that is misinterpretation of the question. The question is not "What is the probability that anything exists, given that anything exists?", it's asking "Why does anything exist?", as per the title.
>The question "Why do I not exist" has never and will never be (seriously) asked.
Haven't given much thought to this, and at face value, yes, you're right. But "seriously" is ambiguous and there are some interesting questions here regarding the possibility of an evil demon like entity, or, more interestingly, GPT-3 posing this question.
I've tried to restrict the suggestions below to sources that I perceive as fairly unbiased.
While it's not exclusively focused on the economics side of things, one recommendation I would have is to read "The World: A Brief Introduction", by Richard Haass.
On the Financial side of things, the Financial Times is probably the best publication for what you're after.
On YouTube I'd recommend tldr news for current news and Caspian Report for geopolitics.
Unproductive bit is true mostly due to the low value added nature of most jobs.
The lazy thing is a tired, prejudiced, stereotype. Unless you also consider countries like Japan who work similar (but less) annual hours to be 'lazy'.[0]
This comment is not only based on anedotical evidence, but demonstrably wrong [0] (compare hours worked between Portugal and the UK - these are significantly higher in Portugal) and perpetuating stereotypes with regards to Southern European countries.
Just a layman's guess here, but I strongly doubt the bystander effect is strong enough to overcome the chances of a single person helping. Note that for the bystander effect to happen, all the people in the vicinity have to not intervene, while for it to be defeated, it takes a single individual to act.
I doubt that as N increases the "network" effects are strong enough to overcome the chance that someone that will help, regardless of what the effect of "N-1" bystanders is, shows up.
That would be my guess as well, at least for large N. It might still show interesting behaviour for smaller group sizes, for instance a group of 2 vs a group of 6, or when the probability of a single person helping is comparatively low to begin with (for instance, in a bad neighbourhood known for scams).
The effect is informative regarding what to do when you need someone's help, though. Since the basis of the effect is the diffusion of responsibility, it helps to address a person directly when asking for help. Instead of asking whether anyone in the group can help with something, point to a particular person and ask them whether they can help. In this way, the refusal to help becomes explicit and people will be less likely do that. This is something I've observed from personal experience too.
It's an interesting thought but I think there's little basis to believe there is a concerted effort here, as the average consumer is very aware that the quality of, say, a screwdriver from China, is very different from that of a German made one.
I think it's just an unintended consequence of a race to the bottom in cheap, disposable manufacturing. Even then, I'm unsure if on an aggregate level the progress in materials engineering and overall manufacturing processes does not eventually offset this focus on cheapness, resulting on better value than the more expensive alternative.
It's not a concerted effort, but over the course of forty or more years it may seem like a trade surplus built on disposable goods is somehow a strategic move?
I'm no AI skeptic (and am aware the above does not seem to apply to GPT4), but the general point stands - finding these limitations is an interesting endeavor, and when they're particularly severe or unexpected, newsworthy.