Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Inexact change.

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


       

Please log in.
Before you can vote, you need to register. Please log in or create an account.

Amazon/Costco/Walmart stuff database prompts higher quality

Automatically "airport scan" a percentage of the things that travel through Amazon/Costco/Walmart/AliExpress; have software process the images and prompt manufacturers to improve the bottom 10%. That bottom 10% benefits from their new ability to compete on a better product
 
(+2, -2)
  [vote for,
against]

Think if 1/10th of 1% of all the packages going through Amazon went through an airport scanner. Or better than an airport scanner, a sequence of scanners: cheap-version permanent magnet MRI, THz scanner, Quantum Camera*, and an ultrasound scanner. A big database of tens or hundreds of millions even a billion instances of products as data objects (mostly images) would be formed. It is possible for the software to then lookup at its database say the number of product seams/joins an image processor counts, and how closely (well) they meet on millions of products, or look at the image in the databse at internals like batteries or power supplies that are built-in.

Then Amazon, Walmart, Costco, or aliexpress start improving the product. They send email to the manufacturers of the 50,000 products with AAA batteries, saying "Engineers frequently replace AAA batteries with pancake button lithium cells, which cost less and are cheaper to source, and are cheaper to build battery-cases for (SMD)".

The software can also prompt improvement on the big one, what percentile a thing is. The software looks at both the number of seams/joins on consumer products, and then looks at its database to find the intraproduct variability in Seams of that specific product form. It then emails manufacturers, "Your product has more seams and joins than 90% of the products in the same category. It is at the 10th percentile or below. Consumers often prefer seamless smooth products; you have the opportunity to improve your product with reengineering". Another big one is quality seam/join variability, "Your product has more between sample seam and join gap variability than 9 out of ten similar products; in some products that is perceived as shoddy, and may contribute to breakage; you have the opportunity to improve your product".

Anyone with an environmental urge might like automated email to manufacturers if their packaging is higher mass or made of more toxic stuff than 95% of all the products at the database. "Your use of CFC-blown styrofoam cups to package action figures...etc". SImilarly, batteries were mentioned but a scanner-view of a power supply let's software estimate energy effciency easily. How much wire in the transformer? How many capacitors? does the power supply have a power transistor? Those would contribute to a higher energy efficiency estimation than wire transformer alone.

Amazon is described as having 12 million nonbook, nonfood, nonreferral products. (Several hundred million including referral products.) Just addressing the bottom 10% of anything, from seams and joins, to lousy power supplies, to stitching quality on garments finds millions of automatically prompted improvements, and alerts manufacturers to the opportunity to more successfully compete with the higher functioning 90%.

Benefit: People's stuff gets better, more rapidly. The stuff they get lasts longer, uses less energy, and wears out more slowly.

Another benefit: The big scanner database of things goes on the cloud, and AI is able to learn, "know" at some depth what things are (like the fitness watch example). This seems like it would benefit people a lot as what's called the built environment makes up most of people's nonsocial environment and AI could improve it.

Many beneficial programs on the database function without AI: Find the 5,000 people who write most like me, with my ratio and types of variously "thought" and "feeling" words; find the 9-300 things these similar person rated the very most highly on Amazon/Costco/Walmart/AliExpress. Those are items, some of which I have never heard of, likely to improve my life. Or, if I already use those items switching to the personality-writing-cosimilar's highest rated (and noting this halfbakery item, best manufactured) product is also beneficial.

beanangel, Jan 12 2021

[link]






       // Find the 5,000 people who write most like me //   

       There's a challenge for anyone who's interested ...   

       Number of humans on this planet - about 7.6E9.   

       Number of humans on this planet who write like beany ... one.   

       Number of other humans who also write like that ... unknown, but very small. At most, a single-digit integer.   

       Chances of finding them ? From the outset, 1 in 7.6E9 ... probably much worse.   

       That's the Class Assignment for February (January is of course the Spot-The-Führer challenge).
8th of 7, Jan 12 2021
  

       So ultimately it is going to come down to collaboration. An individual company's R&D AI is not going to have the analytical parameters of an AI with your collective, shared results, airport scanner data throughput. Or for that matter, an AI with a countries collective manufacturing data.   

       Extrapolating, the globe is the limit to this AI manufacturing data set. Humanity has to share, though.
wjt, Jan 14 2021
  

       This poses both a security risk and opportunity for beanie. If he writes similarly on any site, he could be identified by his writing style alone, so anonymity could be difficult.   

       It could also be a highly-secure captcha for him. "Write a beanie-style idea sentence..."
RayfordSteele, Jan 14 2021
  
      
[annotate]
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle