AI can’t remove/replace wheels, or fix a transfer case!
If you are going to make an argument against AI at least use a sensible one.
1 Like
So what? A mechanic can’t figure out Protein structures. For decades it took scientists YEARS and YEARS to figure out a protein structure. Some people got their PHD thesis on creating just 1 protein structure. AI comes along and has been able to do it minutes.
1 Like
Never say never.
Ever seen how many robots are involved in car assembly now?
It won’t be long before they are ubiquitous even in sub-assembly operations.
It has to be. The biggest costs in manufacturing are labor costs. The company that has the lowest labor costs will win.
Computer vision systems are already dominating the circuit board assembly process. They verify the proper component has been placed and is properly seated. Automated systems do all the work.
Ten years ago, I witnessed the power of Hololens in assisting humans in complex electro-mechanical assembly. Our company developed assembly processes using the mixed reality glasses that blend computer generated images overlaying realtime video of what the operator is viewing. It can point them to the next fastener to torque and is tied into the computer reading the torque wrench output so it knows if you achieved the correct torque. It knows if you forgot one and won’t go to the next step in the work instructions until you complete all tasks. How much more effort do you think it will be to replace a human with a dedicated robot?
We recently developed an automated assembler for one of our products. A COTS robot costing less than $8k has 4 articulating joints and can achieve placement accuracy better than 0.0015".
I think I mentioned that the car dealer I worked with on my Rav4 had an AI receptionist answering calls. It took me a few interactions to realize it was a computer program.
What I look forward to is integration with the act of driving. When some buffoon insists on crossing the dotted lines into my lane on the rotary and acts like it is my error by honking at me. When they attempt to turn into me- “I’m sorry Dave, I’m afraid I can’t do that now” 
1 Like
Sadly, in our “Brave New World” it is de rigueur for the guilty party to cast blame on the person or persons who caught them in the act, and to insult those people.
2 Likes
One of the main reasons manufacturing is coming back to the US is because of robotics. It’s now cheaper using robotics to manufacturer then it is using cheap labor in places like Thailand.
1 Like
+1
And, when you consider that AI is already used to assist surgeons, it is only a matter of time (probably only a few years) until AI is used to perform surgery without human intervention.
If AI can successfully perform Gall Bladder surgery (which it has done), I’m pretty confident that AI can remove and replace wheels, and then torque the bolts correctly.
That’s actually being done already in many automotive manufacturing plants.
1 Like
Yes, now I recall seeing video of that.
That seems to make…
… seem rather pointless.
AI has many amazing applications, my concerns are
-
Huge use of energy and water
-
Favorite tool of criminals and fact destroyers
-
Use in education by students to avoid learning
1 Like
+1 For those who are concerned that their electric utility bills have soared over the past year or so, you can look to the electrical consumption of AI centers as one of the major causes:
In NJ, it is possible that AI centers will have to pay a tariff on their energy usage, in order to protect consumers from more rate hikes:
https://newjerseymonitor.com/briefs/nj-bill-data-centers-electric-costs/
1 Like
It’s not just energy, the explosion in data center applications is also disrupting incumbent supply chains for more traditional or legacy products. Their demand is huge and so suppliers are abandoning relatively lower volume business in order to meet the burgeoning demand of data center applications.
1 Like
These applications are driving the development of next generation gallium oxide semiconductors that will help to significantly reduce power consumption along with other benefits. As with any evolving new or innovative technology, there are developments that will provide far reaching benefits to not just the driving technology but to other areas as well.
Unfortunately, that’s way in the future. Annual electric usage in 1 AI server farms uses more electricity than thousands EVs. And there are about 5,000 AI hubs and growing. Add into that mix the Crypto farms which even use MORE electricity then a typical AI hub.
1 Like
If you look at computer power consumption per MIPs (Million Instructions Per second) from early SS computers to now - power consumption daily usage has dropped from 20KW to 20w and computer power has increased by well over 1,000,000,000.
Not counting the very first computers built with Tubes. Those computers really skew the equation because they had so little compute power and were extreme power hogs. With an early tube computer system to now - MIPS have increased by well over 100,000,000,000 and power consumption has dropped from the Eniac using 3,600 KW/day to 20w.
Even if the new technology is available today it’ll take decades before there’s a noticeable impact.
1 Like
We’ll disagree on that point. Switching power supplies are currently around 85 to 90% efficient in these applications. And there are multiple power conversions that occur from the grid to the chip. GaO has the potential to not only be fundamentally more efficient, but eliminate numerous power conversion steps (e.g. number of power supplies needed in the step down process). So the amount of wasted power reduction is significant and will have immediate effect. Especially when you look at the volume of computational systems in use and projected to increase exponentially…
People/companies are not usually that proactive.
Yep, we have one therapist that is about 300 miles away, all sessions are done with the camera in the phone. Why this therapist? Only one that accepts insurance plan!