Microsoft’s 3 billion bet on Copilot is assessed to be a disaster

[Mnementh]

SvennoJ said:

Rationality doesn't take the personal experience into account. It's very rational to say the world would function better with far less people, release Covid 2.0.

Rationality wouldn't be rational if it avoids taking things into account.

And the conception the world would be better off with less people is just simply wrong. That is called Malthusianism, after Thomas Robert Malthus who laid out the idea in an 1798 (!) essay of his. I can't understand why this idea persists, because according to Malthus we should've seen societal breakdown a hundred years ago. The thing Malthus didn't understand is, that more people also produce more ideas and therefore a lot of progress, which in turn is good for us. There are other factors involved to make it work, like education and basic living, but in contrast to what you claim, Malthusianism is exactly *not* rational, as it clashes with reality.

SvennoJ said:

Rationality also argues torture for information to prevent worse is acceptable, imagined or not.

This also isn't at all rational. That is because the relation you lay out is very short term thinking and avoids facing long term consequences. A society that accepts torturing results in a lot of mistrust and would overall work a lot worse. So again, torturing to prevent short term bad things is irrational, it is after all not drivven by Ratio, but by emotion, one of the strongest emotions there is: fear.

Also evidence shows that torture often results in false information, so it would be mostly useless to add.

SvennoJ said:

Rationality is also excluding deviations from the group to optimize the functionality of the average.

Again, anything that excludes valid things from your considerations isn't rational at all. Rationality is exactly that you look behind your biases and blind spots, and that means it is the rational stance to include all variations in considerations, otherwise how could it be rational if the thinking doesn't reflect the reality of the world (and excluding people that are not average is denying the realities of the world).

SvennoJ said:

I fear rationality! Rationality without humanism and empathy.

Again: humanism is a rational movement. So it is rational to include humanism.

The reality is, that the crimes of the history and the world there the result of misapplied emotions: fear, anger, despair, greed, hubris, contempt, condescension.

Emotions are a product of our evolution and as such work well enough for basic survival — in the world they evolved in. Mostly that means small groups. We aren't made for big societies with thousands and millions of people living together. This is used against us: PR of big companies and propaganda permanently uses emotions against our best interests.

That said, we are after all emotional beings. So the rational thing is also to consider this.

SvennoJ said:

Current AI is also a reflection of us. It's an accelerationist factor on the terrible path humanity is on. AI will only increase inequality, not solve it. AI will also make things less safe when more people rely on it, use it as a 'shortcut' and forego understanding and deep knowledge of complex system for higher productivity. As well as deflecting accountability, the machine did it.

I fully agree on accelerator. That already implies that it is up to us how we use it.

I can take another example of a technology: cars. Edward Bernays working for oil companies convinced the US government to invest in road infrastructure instead of public transport. This transformed the US into a country made for cars, not for people. You have cities there you can't walk, letting your kid walk or bike to school seems crazy to an american. The americans invented the crime "jaywalking", as in being a nuisance for cars. In comparison european cities are walkable, many kids walk to school or take the bicycle. There is plenty of public transport. And many european cities have core areas, where cars are banned altogether. The same technology - cars - different outcomes. It depends on how we use it.

I cannot recommend the story Manna by Marshall Brain enough. Crazy to think it was written in 2003 already, more than 20 years ago. The software described in it - Manna - is not at all like LLMs. But the societal impact is spot on. But this story has a very important subtitle: "Two Views of Humanity’s Future". It lays out that it is our use of the technology, the societal parameters around it that decide the outcome. The technology itself is neutral, it supports either societal structure. In this time of AI usage I cannot recommend reading this more, it is a relatively short read and completely free to read on the internet: https://marshallbrain.com/manna

So yes, LLMs and current AI is rapidly accelerating our societal development. This puts pressure on us to fix the problems we have. But these are our problems and the developments would happen with or without AI - just slower. We need to fix our world, and now!