trention

Your "real question" ultimately resolves itself because the moment the novelty wears off (and it will happen very fast), nobody will be interested in chatting with "robots".

Net zero worldwide will never be achieved with pure emission reduction because neither China nor the industrializing third world will ever accept trading carbon usage for lower (than potential) living standards.

The only way this works is with massive investment in carbon capture. Reducing emissions is nice but alone will never solve the problem.

My first work visit to silicon valley was during the democratic primaries in 2016. Most people there were supporting Bernie Sanders. While me, being from Eastern Europe, could see through him quite easily.

I pretty much made all the inferences I will ever need about the valley-type demographics back then. Nothing since has disproved any of them.

>create fact-checking and supposedly neutral "services"/media

>due to selection bias, those "services" and the people providing them end up being heavily on one side of the political spectrum

>deus ex machina (no, not really)

>wonder why people on the other side of the political spectrum refuse to trust your fact-checking "services"/media

>while the aging population is larger than the younger population who are working at first that’s just because during the elder generations birth rate was higher. As that ages out of the population you’re back to an equilibrium.

This is absolutely wrong. The "aging population" will be larger than the "younger population" in any situation in which there is sub-replacement fertility + stable (or increasing, obviously) life expectancy. It's just that the ratio may stabilize but it will be <1.

>can-kicking at best

Covid deaths per 100k population

USA 311.7

New Zealand 40.27

Source: https://coronavirus.jhu.edu/data/mortality

This is what happens when you "can-kick" infections after vaccination (actually, not even after vaccination but after offering vaccines to the population - and a lot of them stupidly refusing them).

As I said:

>Humanity is really (mostly) unfit to deal with ... nuanced information

No, in that both are usually repeated by clueless charlatans.

Got it, "higher-order thinking" is the new reincarnation of "doing your own research". Not that surprising, actually.

Is there an authoritarian government in the history of humanity that managed to kill 50% of its population? What's the average?

Yes, there are situations in which "higher order thinking" doesn't matter.

The first-order effect of ebola is (almost) certain death. No need to consider anything of higher order.

>an authoritarian surveillance state which is far worse than any virus

I'd take any authoritarian state over ebola, thank you very much.

Excess deaths worldwide are more than 20 million and most of them are COVID deaths since excess deaths track covid cases well pretty much everywhere.

So, yes, covid did indeed kill way more than 6 million.

This is a hypothesis based on a sample size of 9 and already in the comments people are treating it like gospel while pushing their political agendas.

Humanity is really (mostly) unfit to deal with both nuanced information and collective action problems.

I can't comment about India but in Eastern Europe software engineers have a comparable purchasing power than in the US - it's just that it's achieved at ~50% of the wages because of cost of living differences.

>Poland and Ukraine pay between $22,000 and $23,000 per year

Good luck finding a developer in Eastern Europe with more than 2 years of experience willing to work for 23 thousand a year. And if that's a gross number (as I suspect it's supposed to be) - good luck finding anyone with even 1 year of experience.

>Meeting the acceptance criteria with guess and check isn’t hard.

Acceptance criteria often don't exist or are wrong. Incidentally, when someone says "business requirements are easy", that's a pretty big tell they have never worked on a complicated project.

>The fraction of samples being verbatim reurgitations is low

Copilot spitting out a function out of its training data with changed variable names to match those in your file(s) - and no one is actually testing what proportions of results those are - is still regurgitation.

>The existence of the sentiment ‘lol you trained a big model on X so obviously it produces a good probability distribution on X’ only exists because big models proved to be extraordinarily more effective than anyone expected.

No, the "popularity" of that sentiment exists because of their "effectiveness". That sentiment was existing and being voiced 10 years ago.

The proportion of comp sci tasks that have practical importance for which a quantum computer can shave off a factor of ~000s seems to be sufficiently small to doubt it will "change the world".

> and all their content and capital is vanishingly tiny

Only humans are currently allowed to own capital and that's not going to change in the near future.

>judicial systems with parallel construction and deepfake “evidence” (Fake or real) pictures and video are in a lot of circumstances not admissible evidence in court, with deepfakes the set of inadmissible evidence will simply expand.

There is very real and very big concern about fake news and job loss and (in a slightly longer term) GAI going rogue but that's about it.

>Their training data is also part of their output.

More deflection and going off point.

>is making progress in others - like programming

I already explained that programming is not amenable to the approach that allowed alphago to learn from the games it played against itself.

This is already getting pointless and tedious, should in the future know better than to respond to people that are obviously in the "scale is all you need" nonsense camp, nothing meaningful comes out of it.

>I would say these versions did exceed their training data

Well, this is obviously incorrect because the training data included all the games it played against itself (good luck replicating that in the domain of programming though).

We were not talking about "exceeding human abilities in at least some domains", we were talking about whether ML (DL really) systems were able to "go beyond" their training data - but your deflection and going off point is noted.

>Alphago, and similar, are all examples of ML systems exceeding their training data

Alphago basically generated its own training data (hence, "exceeding" is not correct) because the domain was amenable to a specific form of reinforcement learning. Programming isn't such a domain.

Given the fact that those models seem to improve logarithmically and that they pretty much trained codex on ~all available public code (maybe divide by 2) meaning the training data will at most scale with the growth rate of public repos in Github, I doubt that whole "1 million times better". Even the number of parameters won't improve with a factor of 1 million (it will not improve by a factor of a 1000 and may even not reach a factor of 100) by the end of the decade.

Calling this "engineering" is just beyond parody.

>most AI people acknowledge that the timeline for AGI is unknown and that it will most likely involve yet to be discovered algorithms, not just throwing more data/processing power at it.

Try telling that to Sutskever or Karpathy (or indeed a bunch of "rationalists" on twitter, at least a subset of which works in AI) because both of them have made the claim you assert you haven't seen.

>We could claim the human brain is just 100 million neurons arranged in a mesh were each neuron has their own 'activation potential'.

You actually can't because, unlike neural networks (we know exactly what they do), we don't have anything even close to an exhaustive view of the functioning of the human brain.

Your example is not at all as trivial as you may think when you delve into the details. And if you don't want to delve into details, there is already stuff like shopify.

Do you realize that writing the code is maybe ~25% of what a senior software engineer does?

He meant fixing the output(s) not the model that produced them.