Re: Re: Re: Re: Robot automation will take 800 million jobs by 2030
Originally posted by Putinbot1
You dont understand the third world and how much cheaper humans are here than machines.
Economies of scale.
If it costs $7,400,000 to develop the software that can make food, it costs $200, annually to maintain, has a MTBF of 90 days, and makes 228 meals a day (service operation time is 5AM to midnight, every day of the year, at 5 minutes of meal prep and serve time)...
That's 20,520 meals made between each failure. And assume one day for failure (4 production-disrupting failures per year). 365.25-4 = 82,365 meals per day (this number not accidental).
If your manufacturing costs for this device are 24,000 per device (which is actually less but the R&D portion for the hardware-software combo increases the average cost per device and the End of Life and End of Service cycle for this particular device will hit a stopping point and you will repeat it with a Mark II version) and you make 1,000 of them, your hardware costs for the first year are $31,600,000.
Assuming you reach 3 sigma in quality for the device (meaning, your critical failure rate for the devices is 6.7%) per year, you will have to replace 67 devices so add an additional $1,608,000 for the first year.
Now you're looking at a first year cost of $33,208,000.
Second year costs will now be: $1,808,000 (200,000 in maintenance costs and 1,608,000 in hardware replacement costs for the critical failures).
2 year costs: $35,016,000.
That's for development, hardware, and maintenance. Doesn't include a pre-prod testing period (a pilot) or marketing, or regulatory and compliance costs.
And the cost per meal over that 2 year period is going to be:
$35,016,000./164,730,000 meals = $0.21 per meal + costs of food per meal (transportation, certification, regulatory and compliance, food cost itself from distributor, storage (cold storage and dry storage), and import taxes).
21 cents per meal doesn't sound bad but all that food will likely have to be imported and stored. Marketing can be very cheap in poorer places. If food costs can be less than a dollar (extremely unlikely), this is feasible in many places.
Also, I did not account for maintenance techs travel in that cost.
I did all that to show that I agree with you. The "economies of scale" has to be in the tens of thousands before this is feasible. That's why a place like McDonald's might be a better choice before a poor country is a good choice. McDonald's already has the acquisition, manufacturing, and storage problem solved for food. They just need the food making bots. And their bots will cost more than $100k for each device, require more maintenance, and have a higher failure rate. 3 sigma is not that great of a quality standard.
Re: Re: Re: Re: Re: Robot automation will take 800 million jobs by 2030
Originally posted by dadudemonto use most African countries as an example the average wage in many is 40$ a month or 480$ a year, that's for a house chef, gate man or driver. Some it's far less.
Economies of scale.If it costs $7,400,000 to develop the software that can make food, it costs $200, annually to maintain, has a MTBF of 90 days, and makes 228 meals a day (service operation time is 5AM to midnight, every day of the year, at 5 minutes of meal prep and serve time)...
That's 20,520 meals made between each failure. And assume one day for failure (4 production-disrupting failures per year). 365.25-4 = 82,365 meals per day (this number not accidental).
If your manufacturing costs for this device are 24,000 per device (which is actually less but the R&D portion for the hardware-software combo increases the average cost per device and the End of Life and End of Service cycle for this particular device will hit a stopping point and you will repeat it with a Mark II version) and you make 1,000 of them, your hardware costs for the first year are $31,600,000.
Assuming you reach 3 sigma in quality for the device (meaning, your critical failure rate for the devices is 6.7%) per year, you will have to replace 67 devices so add an additional $1,608,000 for the first year.
Now you're looking at a first year cost of $33,208,000.
Second year costs will now be: $1,808,000 (200,000 in maintenance costs and 1,608,000 in hardware replacement costs for the critical failures).
2 year costs: $35,016,000.
That's for development, hardware, and maintenance. Doesn't include a pre-prod testing period (a pilot) or marketing, or regulatory and compliance costs.
And the cost per meal over that 2 year period is going to be:
$35,016,000./164,730,000 meals = $0.21 per meal + costs of food per meal (transportation, certification, regulatory and compliance, food cost itself from distributor, storage (cold storage and dry storage), and import taxes).
21 cents per meal doesn't sound bad but all that food will likely have to be imported and stored. Marketing can be very cheap in poorer places. If food costs can be less than a dollar (extremely unlikely), this is feasible in many places.
Also, I did not account for maintenance techs travel in that cost.
I did all that to show that I agree with you. The "economies of scale" has to be in the tens of thousands before this is feasible. That's why a place like McDonald's might be a better choice before a poor country is a good choice. McDonald's already has the acquisition, manufacturing, and storage problem solved for food. They just need the food making bots. And their bots will cost more than $100k for each device, require more maintenance, and have a higher failure rate. 3 sigma is not that great of a quality standard.
This topic made me remember something I learned awhile ago.
You know map makers for Doom? Turns out there's automated programs, that do such a good job, they beat the professionals. People have gotten awards for just running a random map generator.
And that's just the tip of the iceberg:
Originally posted by cdtm
This topic made me remember something I learned awhile ago.You know map makers for Doom? Turns out there's automated programs, that do such a good job, they beat the professionals. People have gotten awards for just running a random map generator.
And that's just the tip of the iceberg:
Procedurally generated levels/maps has been a thing for a long-ass time.
Deep Blue with chess was in the 90s, to put things into perspective.
We need much more automation in farming before we can work down the path of automated food production.
Hope the birth of true AI comes soon. They can make these improvements faster than we can.
Originally posted by dadudemon
Procedurally generated levels/maps has been a thing for a long-ass time.Deep Blue with chess was in the 90s, to put things into perspective.
We need much more automation in farming before we can work down the path of automated food production.
Hope the birth of true AI comes soon. They can make these improvements faster than we can.
Oh yeah, I'm way late to the party on random map generators.
That one that's been running for decades non stop, producing a map a day every day. I'd like to know how often the computers been replaced, if ever, and if it's just been sitting in a basement somewhere, long forgotten.
The 6 Worst Jobs Ever (Were Done by Children)
Article is nearly 9 years old and still funny. This is back when Cracked was good.