I partially agree with you, but it depends heavily on the task. In certain kind of work AI really helps you, in other kind of work it just waster your time. Work where finding the answer is hard (but not too hard), but validating it is easy and fast, AI works quite well.

I think that the problem is that quite many try to use AI for work where AI is extremely bad at. E.g. I would not use AI to write production code, because it is much harder to validate and fix code than write it yourself, but I could use AI to write prototype code, because you can validate it well enough just by running it.

Perhaps not in the calculated percent but for example AlphaFold2 did about 100 000 000 years of scientist work by solving protein folding and AlphaEvolve improved several algorithms, which human scientists have failed to improve, so it is hard to say how much did it do.

That dependency has dropped in the past years from 11 million tons to less than million tons. It is not that Euope depends on it, some countries just prefer to buy cheap oil.

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.statista.com%2Fstati...

- Creating common currency (Remember that EU is not a single country)
- Border-free travel thanks to the Schengen area
- Keeping peace in EU (Europe used to have a lot of wars)
- Creating single market inside EU (benefits which Britannia noticed after brexit)
- Connecting the EU into the same electricity network.
- Common charger (Just adding this, because I like charging my devices with the same USB-C charger)

I have tested AI quite a lot and I know quite well how often AI makes mistakes and it is not a problem.

I would not use AI to generate production code, but I could use it to generate a prototype for the UI. I would not ask AI about books or songs, but I would ask it to solve a school math problem. I would not ask it to write a book, but I could give it a part of the book and ask if there are any errors in the page that should be fixed.

In general, ask things that are hard to answer, but easy to verify. Things where you can throw the answer to trash after you got the information you need. Don't ask for things you want to save and store.

I am personally impressed that LLM can solve school math problems, even if it makes small errors every now and then. But what I am really impressed about is the fact that AI has made huge scientific discoveries.

Even if you don't use AI at all. Are you not impressed that AI solved protein folding (a Nobel prize problem)? Are you not impressed that AI is next used to discover new drugs that work exactly to the problem and because we know all human proteins, we can also let the AI test for side effects. And once the system is up and running, we can just let it solve all known deceases.

Interesting thing is that if you 2000 years ago had enough money to hire a soldier and you turned that money into gold. If you today would turn that gold into a money, you could hire a soldier with it.

On the other hand. If you 5 years ago had enough money to hire a soldier and you bought NVidia stocks with that money. Today the stock would be worth of about 30 soldiers.

Prometheus was a titan who stole fire from the gods and gave it to humans. Because of this Zeus tied Prometheus to a rock where an eagle would eat his liver every day, a torture that never ends.

I am curious. Did they pick the name in honor of a thief, or is it a prophesy of how the project will end.

> Java's import statements bring in a single type.

import x.*;

I don't mind giving Google data, but I don't give them everything. I mostly ask trivial questions related to e.g. cleaning, cooking, computers, math, chemistry, things I wouldn't mind even if someone finds out I asked those. If they can use that to improve their AI that is even better for me.

AI has been a great help for me, but I wouldn't pay for it.

Natural time cycle (length of day) for humans is speculated to be 25 hours based on experiments where human lives in isolated area without sun and clocks.

> Solar panels require an accompanying energy storage system, hydro or gas turbines.

That is not exactly true. For example Europen super grid is an alternative solution. The idea is that you generate electricity in one country and then transfer it to another. For example "midday" lasts about 7 hours in Europe, so at least one country. You will have pretty good generation between 8:00 - 16:00, so that is 15 hours of good electricity generation. During night time the demand is lower. When you combine this with wind, you get pretty good summer/winter coverage.

You will still need for example hydro and gas turbines, but those don't need to be in your country as you can buy it from other countries. For example Eastern countries can buy nuclear from France during their evening time when it is already night in France. And hydro in Sweden and Norway can cover multiple countries around them.

I think that the best game ever is Minecraft. Not the vanilla one, and not entirely because of its creators, but because of those who write mods and modpacks for it. They get usually zero money from it, and I think this is why it works:
0. Some people write a framework, which makes writing mods easy.
1. A random person writes a good mod, e.g. Immersive Engineering that adds huge machines to the game.
2. Another person writes a mod that adds pollution. Similarly others write new simple mods.
3. Next person combines the mods and configures them into a working modpack. This also repeats several times.
4. Now you have a thousands of completely new games that has completely different goal than the vanilla Mnecraft. Some mods add magic, some add monsters, some animals, some space ships etc.

So I disagree with Nadella. I think that innovation needs a lot of people who enjoy doing things.

The more years women spent at school, the less children they will have, as shown here:
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwol.iza.org%2Farticles%2Ff...

Next, be reminded that Finland has free education system (for citizens) and many women (over 20%) spend 18 years in education and at minimum they will spend around 13 years, most spending something in between. Look at the diagram again and see where this range falls into.

I agree with you and I want to add some examples.

Here is a story from 2016 how a cucumber farmer automated cucumber sorting with AI. I want to mention this, because many think that only LLMs are AI, but there are much simpler implementations of AI also that replace tasks and those have been around for a long time. Replaced tasks mean less need for workers:
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fcloud.google.com%2Fblog%2F...

Bigger example would be Amazon. If Amazon didn't have robots and AI doing work in their warehouses, they would have to employ much larger group of people. Only reason why Amazon is not cutting jobs that much, is because they are growing at the same time, so job cuts happen in their competitors. Here is a story about that:
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.nytimes.com%2F2019%2F1...