Marking this important moment in history on my blog.Last week we had humanoid robots setting new marathon records (power/speed) and beating professional ping pong players (dexterity/reflex).

We have noted important AI moments in the past like IBM defeating a chess grandmaster in 1997, Google defeating Go in 2015, and OpenAI passing the Turing Test in 2025.

Well, today we have 20 different AI models (LLMs) that are smarter than the average human (ranked as 100 on this chart).

More importantly, we now have 11 models running beyond genius level (130 on this chart) according to Mensa Norway.

Futurist Ray Kurzweil predicted that by 2029, computers will achieve human-level intelligence, passing a valid Turing test (already happened last year), marking the arrival of Artificial General Intelligence (AGI).

He expects this AI to possess human-level cognitive capabilities, leading to significant human-machine integration and the start of a technological/biological singularity by 2045.

He may have been too conservative on this forecast by at least 15 years.

A humanoid robot just surpassed the human world record (in half-marathon) by 7 minutes.Having worked through 32 years (and living through 53 years) of the technology industry's growth, it always amazes me what it feels like "in the moment" versus the way we wax poetic about it 20 years later.

We all know those moments.

When a young Bill Gates thought every home will have a PC, and the IBM people in the same meeting felt like the total opportunity was 10,000 hobbyists.

When a young Jeff Bezos wondered what to do with tech infrastructure they had to build for their busiest day (but sits idle for the rest of the year), and magazines start making fun of him on the front cover.

Here is the deal.

The first PC was VERY limited. But it did have slots, bays, and ports that enabled innovators and entrepreneurs to build trillions of dollars of value in the future (Thanks Steve Wozniak for winning this battle over the other Steve).

The first iPhone was VERY limited. But it did have a future vision of what an App Store could do and unleashed millions of inventive minds into the ecosystem.

The same criticism came of early LLMs.

They hallucinate. They can't be trusted. It is a consumer noveltly.

Then an LLM passed the legal bar exam (bottom 10%). Six months later it passed (top 10% - Harvard/Yale level). And then 6 months later it was writing, grading, and proctoring the exam for humans.

All in 12 months.

We have difficulty in projecting what fast growing technologies will become. Bill Gates said in 1996: "We overestimate the first 2 years, and underestimate the first 10".

So here we are with humanoid robots. They are clunky, they look funny, their fingers don't have the dexterity.

Until they do.

--> Last year, humanoid robots finished the half-marathon for the first time. The result? Twice as long as the human winner. It was funny to watch.

--> This year? They set a new world record by a staggering 7 minutes.

--> Next year? They will probably finish a half marathon in 5 to 10 minutes total - the race will have to move to a Formula One track to facilitate.

Future:

The world has been built (for thousands of years) by humans FOR humans. The idea that we need hundreds of "smart devices" in our lives may turn into the fact that we only need a very few - shaped as humans and as capable (both physically & intellectually).

Examples:

- Your smart vacuum (sorry iRobot on bankruptcy), doesn't need to be smart when a humanoid robot can use a dust pan and mop.

- Your fancy smart cat litter box can go back to the normal one when a humanoid robot cleans it after every use.

- Your car with tens of thousands of dollars in sensors, cameras, and Lidar can go back to normal when a humanoid robot can take the wheel (with safety already 10X distracted humans).

- Even as simple as your smart thermostat. It can go back to the old dial when humanoid robots (which will be in every home within 20 years) can turn the dial with much better accuracy.