[jonshariat]

I think this is a great example that shows how our brains remember/recognize things. (Which may be help in machine learning)

This is how people have remembered a bike.

2 wheels, spaced apart Metal connecting Pedal in the middle Seat on top Handlebars connected to the front wheel.

The details of how it's connected, where, spacing, etc. Isn't really noted by our brains unless they see/use them often.

I wonder if this exercise could help use fine tune ML patterns.

[21]

I agree, but I also wonder how much motivation is a factor.

What would happen if the author said after the first drawing: "this bicycle is terrible and this bike would never work. Take this seriously, think about the mechanics of it, and I will pay you 1000 dollars if you get it correct"

As in, these pictures are just raw memory dumps, with minimal cognitive filtering. When stakes are higher, cognition should kick in and fix lower level errors.

[edgarvaldes]

I think it's funny how many comments give great importance to the object being a bicycle or the person being a bike rider to explain the output.

[jonshariat]

Agreed. It can really be anything. People don't look at things, they scan and store/recognize optimized patterns.

[pwg]

> I think this is a great example that shows how our brains remember/recognize things.

It may also have much to do with 'who' was asked. As an Engineer, I immediately noticed the missing frame member on the first rendering image at the very top of the web page. But, most people one would meet 'on the street' are not Engineers, and they very likely do not even look at a bike with a 'how does this work' eye. And if one does not even try to 'analyze' the bike mechanism, even a little, one surely will never remember the constituent parts.

[swaggyBoatswain]

Wouldn't surprise me but this is already being done by Google Captchas, signposts, and cars.