TL;DR -- We pause to refresh the technology focus for the future. Simulation will be a recurring theme for many reasons.
---
Given that we are now into the age where everyone has heard of "AI" and have three years of data on both sides of the argument, the future will be more interesting from several angles. Technology is great, when it works. It can be expensive.
Saw this image today and had to show the particulars. It's a robot that went through millions of simulated miles with no problems. But, then stumbled on something unexpected when let out into the real world. On LinkedIn, see the article: The Simulation Myth.
 |
| Spot, of Boston Dynamics |
Now, it's not just robotics. The question of how well advanced simulation can mimic Nature has been on the table for decades. Some are more inclined to think that a good simulation can be used instead of something natural. We will be discussing that.
Rodney Brooks of Boston Dynamics has been at this for a long while and has demonstrated remarkable results. For instance, he got to where his robots (of a bipedal type) could somersault without a tether. That is remarkable mathematics and computing.
It was interesting watching this development. Last year, Rodney was quoted as noting that the advancement was not quite as remarkable as he would like. He wants his bipedal robots to dance. And, after that comment, plenty have tried to show this. With varying amounts of success.
Dexterity? Like humans have? Rodney's site (Robots, AI, and other stuff) goes over that topic, quite thoroughly.
It is great to see industry leaders take the thoughtful approach rather than harp with marketing or PR messages. Science needs to be involved as we try to cope with technology of the newer types.
AI isn't new. It dates back to the 1950s in the current mode. But, people have grappled with the related ideas for a very long time.
Simulation goes back further and was part of the earliest computing.
---
Note: Control can be improved with sensor/activator coupling with good simulation software. But, even then, the theory of control and of the domains involved have to be the prime focus with careful attention to the details of experience and experiment. How far can simulation be pushed? Don't expect that to be a closed issue anytime soon; if anything, we will get wiser about what's involved with the limits that lurk.
Remarks: Modified: 10/11/2025
10/11/2025 -
No comments:
Post a Comment