This is the truly scary part for students using LLMs. We’re going to end up with civil engineers who don’t know how to calculate weight distribution. Doctors who don’t know how to read an EKG. Lawyers who lose cases they should have won, because they didn’t follow basic procedures.
I work in live events. I do things like hanging like array speakers for concerts. Just last month, I was on a show that was hanging an LED wall behind the stage. A LED wall is made up of individual panels, which lock together and form a solid screen in the size and shape you need. There have been several high-profile LED wall collapses, because they’re huge, heavy, and easy to fuck up. And when they collapse, people get seriously injured.
Typically, you build them by hanging truss from chain motors, and then mounting the screen to the truss. The screen gets built in rows, with one row getting snapped together across the truss, then the truss is raised slightly, then you snap the next row on. Repeat until the screen is at the desired size. So this means the screen gets heavier and heavier as you build it. Part of the reason why collapses can be so dangerous is because it often happens while the screen is actively being worked on. You’ll have crew underneath the truss, hanging panels to build the next row. And then suddenly the chain motors start to slip because they quietly blew right past their weight limits as the crew added additional weight with each row.
While working that gig, I overheard the lead rigger (the one in charge of calculating weight distribution on the hanging motors, designing the truss system to hold everything, deciding exactly where the motors should be mounted, etc) utter the words “yeah, ChatGPT says these panels are only 25 pounds each. That means we’re right at our weight limit. We should be good.”
In stagehand work, we use 25 pound sandbags all the time. I know what 25 pounds feels like. The same way a farmer would be able to feel if a hay bale is too light. I had been snapping these panels together all morning, and I knew without a doubt that they didn’t feel like 25 pounds. I looked up the actual tech specs directly from the LED panel’s manufacturer. Each panel was 35 pounds, not 25.
ChatGPT just hallucinated the 25 pound weight, and the dumbass rigger didn’t bother to double check any of it. We were going to be ~40% over our chain motors’ posted weight ratings, and needed like four extra motors to help carry the load. The screen was already halfway built on the truss, so hanging new motors was a giant pain in the ass. It required a scissor lift and climbing riggers to go install new pick points. It easily added an extra hour to the install, while we just sat and waited for the riggers to work.
That rigger is still working in the industry. I hope that it was a learning experience for him.
I work in health tech. Maybe new physicians will have issues, but most physicians I see don’t even trust other physicians unless it’s a different specialty. Most don’t trust AI at all.
I do production rigging and want to assure any would be concert goers that this is still a minority horror story as far as ive seen. The video walls i see fall most often are ground based because people will either not weight the bases of them or they will remove weight too early.
Shit happens but the typical standard is to rate motors for 3-5x less than what they can atually take.
Either way, the danger is usually to the crew, not the patrons. That said, there are literally tons of equipment suspended overhead most concerts. Do with that info as you will.
Yeah, standard safety factor in my area is 7:1. So if you expect 1 ton of load, you actually rate your rig for 7 tons. But that safety factor is mostly to account for things like shock loads, where gear “weighs” more when it bounces. So like if all of your motors stop moving at the same time and the rig bounces slightly, it will temporarily put more weight on the motors than the static load normally would. So if you hang 1 ton on a motor that is rated to fail at 1.5 tons, you can easily cause a failure when the load bounces.
The safety factor also helps add a buffer for things like one motor being slightly more loaded than the rest. Even a small discrepancy can cause huge weight differences where one motor is holding a lot more weight than the rest. The 7:1 factor helps buffer that, where the motor won’t fail just because it’s slightly higher than the rest.
This is the truly scary part for students using LLMs. We’re going to end up with civil engineers who don’t know how to calculate weight distribution. Doctors who don’t know how to read an EKG. Lawyers who lose cases they should have won, because they didn’t follow basic procedures.
I work in live events. I do things like hanging like array speakers for concerts. Just last month, I was on a show that was hanging an LED wall behind the stage. A LED wall is made up of individual panels, which lock together and form a solid screen in the size and shape you need. There have been several high-profile LED wall collapses, because they’re huge, heavy, and easy to fuck up. And when they collapse, people get seriously injured.
Typically, you build them by hanging truss from chain motors, and then mounting the screen to the truss. The screen gets built in rows, with one row getting snapped together across the truss, then the truss is raised slightly, then you snap the next row on. Repeat until the screen is at the desired size. So this means the screen gets heavier and heavier as you build it. Part of the reason why collapses can be so dangerous is because it often happens while the screen is actively being worked on. You’ll have crew underneath the truss, hanging panels to build the next row. And then suddenly the chain motors start to slip because they quietly blew right past their weight limits as the crew added additional weight with each row.
While working that gig, I overheard the lead rigger (the one in charge of calculating weight distribution on the hanging motors, designing the truss system to hold everything, deciding exactly where the motors should be mounted, etc) utter the words “yeah, ChatGPT says these panels are only 25 pounds each. That means we’re right at our weight limit. We should be good.”
In stagehand work, we use 25 pound sandbags all the time. I know what 25 pounds feels like. The same way a farmer would be able to feel if a hay bale is too light. I had been snapping these panels together all morning, and I knew without a doubt that they didn’t feel like 25 pounds. I looked up the actual tech specs directly from the LED panel’s manufacturer. Each panel was 35 pounds, not 25.
ChatGPT just hallucinated the 25 pound weight, and the dumbass rigger didn’t bother to double check any of it. We were going to be ~40% over our chain motors’ posted weight ratings, and needed like four extra motors to help carry the load. The screen was already halfway built on the truss, so hanging new motors was a giant pain in the ass. It required a scissor lift and climbing riggers to go install new pick points. It easily added an extra hour to the install, while we just sat and waited for the riggers to work.
That rigger is still working in the industry. I hope that it was a learning experience for him.
As an engineer, what the fuck‽ Who doesn’t double check the weights then add a safety margin (20% is traditional for us)‽
I work in health tech. Maybe new physicians will have issues, but most physicians I see don’t even trust other physicians unless it’s a different specialty. Most don’t trust AI at all.
They will have to work with AI or the C-suite will decide it’s cheaper to have AI provide shitty health care than to pay doctors for good health care.
Oh yeah docs complaining about what others is as commons electricians complaining about other electricians or programmers about their past selves
I do production rigging and want to assure any would be concert goers that this is still a minority horror story as far as ive seen. The video walls i see fall most often are ground based because people will either not weight the bases of them or they will remove weight too early.
Shit happens but the typical standard is to rate motors for 3-5x less than what they can atually take.
Either way, the danger is usually to the crew, not the patrons. That said, there are literally tons of equipment suspended overhead most concerts. Do with that info as you will.
Yeah, standard safety factor in my area is 7:1. So if you expect 1 ton of load, you actually rate your rig for 7 tons. But that safety factor is mostly to account for things like shock loads, where gear “weighs” more when it bounces. So like if all of your motors stop moving at the same time and the rig bounces slightly, it will temporarily put more weight on the motors than the static load normally would. So if you hang 1 ton on a motor that is rated to fail at 1.5 tons, you can easily cause a failure when the load bounces.
The safety factor also helps add a buffer for things like one motor being slightly more loaded than the rest. Even a small discrepancy can cause huge weight differences where one motor is holding a lot more weight than the rest. The 7:1 factor helps buffer that, where the motor won’t fail just because it’s slightly higher than the rest.
If youre pushing that close to your limit with a bounce you really need to reevaluate your rig lmao