Jon Phillips
September 2, 2026
If you looked at what AI’s were doing a handful of years back, it was largely specialized deep learning machines pulling off some specific scientific tasks or other intellectual feats that humans couldn’t hope to match. For example human grand champions for chess and GO cannot hope to win against the top game playing machines. Same is true of video games in general. Nobody bothers with it anymore unless the AI’s have strictly limited capabilities. Otherwise the machines always win.
 |
| AlphaGo, developed by British computer company DeepMind beats GO Master Lee Se-dol, 2016. |
In science, Alphafold demonstrated that it could basically perform decades of human scientific theoretical and modeling efforts in hours or days. That radicalized anything having to do with predicting protein folding — producing high tech medicines as a single tech use example. What is AI doing now? It’s poised to revolutionize entire economies — and likely to disrupt human labor markets in the near term. Human society cannot adapt so quickly to keep up with the radical changes that are going to occur with AI right now and for the foreseeable future.
In 1963, American biochemist Christian Anfinsen proposed that a protein’s one dimensional sequence of amino acids should give away its full 3D structure. This work earned him a Nobel Prize in 1972. Since that time, scientists have been exploring this route using less expensive and more accessible computational methods. The problem is that there's an almost infinite number of ways that a protein could fold; identifying them all can take a lifetime. Enter AI.
These current humanoid robot Olympic games seem silly, but what was going on in humanoid robotics just a handful of years ago in a comparative sense. Not much frankly. A revolution is happening right before our eyes, year over year, with humanoid robotics (it’s linked to the AI revolution — it’s not just about advances in materials, motors, batteries, etc, it’s about a revolution in optimal controls and a great deal of that happens in a brain in biology).
The real point is to develop and demonstrate humanoid robotic technology because it will be able to “fit” into the kinematic and dynamic space that humans “fit” in physically. A great deal of the societies we live in are deeply engineered to accommodate human bodies and how they move and what they can generally do (not high athletic prowess). Does an average human outrun the fastest humans on Earth? Not remotely. But now, humanoid robots can outrun that fastest human and out jump the highest jumping human and... This acceleration of physical capability of humanoid robots happened in just a few years. What will happen over the next decade? I suppose it’s difficult to imagine.
The truth is that we don’t need all that many humanoid robots. Specialized robots would be better for many repetitive purposes — and probably less expensive. Self driving vehicles are examples of specialized robots — they currently operate commercially as taxis in some of the most demanding traffic in the world (San Francisco for example — have you driven in San Fran?). Right now, the freight trains in the US are being converted to operation by wire. I talked to a current train engineer that I know about what’s going on at work. That’s all he talked about. He’s worried about his job; will he get 'excessed' by a machine?
Train engineers are often at home driving their train these days. They have a system like a fancy video game with lots of camera feeds and a console that lets them drive a train hundreds of miles away. Someone might say, gee… they could live wherever they want. But think a little further, if it’s all drive by wire at a distance, can’t you just turn freight trains into self driving robots? Of course you can. Another human profession bites the dust.
Passenger trains would likely have human engineers because of liability of moving humans as opposed to freight. I suppose the passenger trains will have fewer engineers but they’ll be there as a sort of backup monitoring the machine (perhaps one or two staff?). They’re already putting self driving tractor trailers on US Interstates for late night runs where traffic isn’t so demanding. No more truckers will pull those shifts in the future. Workers are still required at distribution centers to load and unload trucks, but the amount of automation there is growing very quickly in those gigantic warehouses. Cargo is palletized and has bar codes. The few jobs that are difficult to fully automate, since they require human dexterity, may eventually be filled by humanoid robots.
Can you build a specialized robot that can outperform any human professional athlete? Of course that could happen, and especially if you don’t require them to move the same way that humans move. I also point out that human athletes are highly specialized if they want to be the best in some particular event. But the fact of the matter is that the human body is organized to do many general activities — not just one. Being a great athlete is actually contrary to the main point of the design of a human body — which is generalized ability to do many varying daily tasks.
If you can ditch the legs and replace them with wheels, can a robot go much faster than a human? Absolutely and already. A robot could outrun the fastest running animals on Earth right now (self driving cars already do every day on the Interstate highways in the US). A robot can fly faster than bird if you allow it to use propulsive motors, propellers, etc. It could break the speed of sound. Self flying cargo planes (they’re already in advanced testing stage) are going to be a reality soon. Wide body jets that are robotic, flying all over the world. The passenger planes will probably drop the number of human pilots a little (the staffing per plane isn’t that large to begin with). Modern cargo ships are already fairly automated. Passenger ship will likely retain lots of social and wait staff since the human touch is the only reason anyone travels that way anymore.
You can do very clever things if you don’t try to copy animal body dynamics (including humans). A great deal of limitation on animal body motion is linked to the need to have the internal systems of the body completely sealed to avoid biological infection all the way down to viruses. Rotational motions and wheels on machines are great inventions, but they’re not perfectly sealed devices. Legs and arms are general purpose motion devices that can be microscopically sealed at joints. We don’t have legs and arms because they move the best. We have them because they support general forms of motion and they can be sealed against infection.
They move up and down stairs, over midsized broken rocks with large cracks, etc., at high speed. They can also have two more appendages (arms) that can carry and use tools (6 appendages or more — bugs seem to have some advantages in body form — but still no wheels, drive shafts, motors including props and jet engines). These robots can easily outrun humans. They will inevitably be on future battlefields.
All the blah blah about humanoid robots is missing the point. In just a few years all this will be entirely clear. When we start "thinking outside the box," the possibilities are endless, and frightening.
