https://youtu.be/dsI3MEDZTYI I think these guys are more on top of what the big news in AI has been... I doubt McDonalds makes next year's list...

Seems to suggest Microsoft.

Really good new video team covering programmers getting AI'd out of their jobs:

https://youtu.be/DBJNZOLWPgQ

If you are afraid, democracy is your friend, because nothing can happen without a majority.

If you are afraid democracy will fail, voting is your friend.

If you are afraid of guns and autarky, the state is your friend... because the state can coordinate responses.

If you are afraid of being useless and poor, then having land/agriculture/gardens is your friend.

Whatever is safest will be most targeted because it will garner the most resources. Ironically, it is things that are not hard that will be slow to be automated. In STEM, I would guess items in pure science like astronomy will garner fewer resources and thus will be last to automate. Things needed and wanted that are in short supply, like nurses and physicians will see great focus on automation to make them universally available.

Depends on your view of overblown. I lived before smart phones. I still can't quite get over them.

I lived before the Internet. People who have not done so really can't fathom it. Just like I cannot fathom the absence of cars and airplanes.

My general range is 5-10 years. Next year's breakthroughs are already prototyped and carry much greater secrecy, but I'll have a go. These are on or before end of 2023:

1. Google will announce an entertainment division that will be mostly AI-based content.
2. The US military will announce a completely crewless capital ship or submarine.
3. CAR-T will become the fastest-growing treatment for cancer.
4. Autonomous electric trucks will be fleet vehicles in mainline service with at least one mainline logistics company in Europe.
5. Large-language model derivatives will cause a major rethink in the format and process of academic publishing across virtually all journals.
6. A major nation--perhaps Korea or Japan or France will require all medical diagnostics to be reviewed by an AI.
7. Trains will be produced to be driven without crews.
8. Airbus will announce plans for an aircraft that will have no functioning flight crew--merely fall-back status for humans.
9. NASA will announce plans to build a drone/robotic deep space station to facilitate trips to Mars and beyond.
10. Humanoid robots will be used in policing.
11. A full-length feature production will show at major movie theatres conceived, designed and made exclusively by AI.
12. AI will write songs that enter the top 10 in sales--though performed by humans.
13. AI will write a critically acclaimed novel.
14. Most major car companies will announce their future is exclusively electric vehicles and that they will be largely on-demand autonomous by 2030.
15. Governments will announce smart pylons that enable any device to know the exact time and location of the beacon signal from the pylons. These will become common place--perhaps on city light polls.
16. Governments will announce smart pylons that enable any device to know the exact time and location of the beacon signal from the pylons. These will become commonplace--perhaps on city light polls.
17. Universities will begin to reform teaching programs to drive costs lower by using at-home/AI methods to teach most basic STEM courses--e.g. Calculus, chemistry, physics, etc.

I think people overestimate medicine as a field. It is incredibly slow and conservative--sometimes out of fear of ethics, more often out of rather poor incentives and scientific practices. Data falls on the floor with every patient. Medical training requires calculus *useless and ignores statistics and probability. It is still a rare medical school that teaches anything about AI or patient data systems. Hospitals swim in bacteria. I have personally witnessed data handling that would shock a good undergraduate-trained data scientist. On and on it goes. I seriously doubt progress in this field will arise soon.

Truckdrivers make more now than ever before--because any 19-year-old knows it is unreliable to enter the trade. You will get paid for a time to be obsolete.

Most of this is cost. Robotics is easier in a standardized environment. Humans will, for a time, be better at one-off speciality stuff. But the cost variance to have "custom" will be 10-fold. Humans are stupidly expensive. That's why robots win. Not because they are at first better, but because they are nearly always hugely cheaper. They don't get tired. They only get better. They don't have mental breakdowns, get old, or go postal. And they get an upgrade in skills every 30 days. Hard to compete.

It really doesn't matter. The question is, is it feasible? If it is feasible, it will be built regardless of public opinion. Once built, you can't un-invent it.