Hacker News Reader

Comment by shmatt | Hacker News Reader

It’s a well known industry “secret” that just walk out shopping (like Amazon Go) is just cheap labor checking you out remotely

Yes, AI powered software can help them associate items to a person, but not even close to the 100% success rate companies like Amazon and Trigo claim

View on HN

> It’s a well known industry “secret” that just walk out shopping (like Amazon Go) is just cheap labor checking you out remotely

Interesting, because this wasn't a well-known "secret" within Amazon.

Recode has learned that Amazon has staff on call behind the scenes to assist the computer vision system that is supposed to detect which items a shopper pulls off a shelf and carries out of the store.

An Amazon spokesperson confirmed the setup and said that Amazon staff is asked to help out when the system used in the new Amazon Go store can’t make a determination

https://www.vox.com/platform/amp/2017/1/6/14189880/amazon-go...

I work in the industry and the 100% success rate touted by these companies just doesn’t exist in live in person real time inferencing. We will get there one day, but in many cases “magical” computer vision AI has humans verifying and changing AI outputs

Imagine Amazon was flexing the stores being fully automated before “Track Everything Everywhere All At Once” was written

This seems fine though? The postal office has workers who decipher the vanishingly small percent of mail automated systems can’t figure out.

It would be an issue if they were lying about how often humans were needed, but as far as I can tell they’ve never claimed anything like the “100% success rate” stated in the OP’s post.

I do wonder why people automatically assume "AI" is "100% AI"?

Maybe we need to acclimatize. We all know when a company claims their soda is the tastiest or that their cars will make you cool, it's marketing fluff. Or that packaged food has to have some amount of preservatives. We should just assume that "AI" means "AI system", one that involves humans for software maintenance, taking care of edge-cases, and providing training inputs.

That would be a very weird assumption for someone who works in the industry, as the OP claims.

The "vanishingly small percent", isn't when it comes to "AI" systems. It's more like 5-10%

How did you come up with 5-10%?

Can you expand on this? Google just returns advertisements posing as news articles parroting the joys of Amazon Go.

It's BS. I have friends on that team, and the tech works (it's very expensive and doesn't scale too well, but it works a hell of a lot better than having some Indian dude squinting at security camera footage)

I’m sure your friend is under many NDAs where they understandably might or might not be comfortable discussing implementation. But as of 2017 Amazon admitted they are using humans in the loop for go, you can ask your friend at what year they stopped

You basically implied that the whole thing is a scam and doesn't work at all, now you're walking back your initial egregious slander to a more defensible position

>it's very expensive and doesn't scale too well,

If the system was AI and image recognition based, this would not be the case, as it would scale very very well.

The fact that it's expensive and doesn't scale well is in fact a great red flag that it relies on human labor.

Where does Amazon claim a “100% success rate” or that humans are never used for review behind the scenes?

Is there a source on this?

That sounds extremely expensive and error prone.

It’s error prone when no humans are involved. humans in the loop as the industry calls it, is what makes things fool proof

I posted the recode article about Amazon admitting to it above

It's a lot less expensive when you hire exclusively third world laborers for a dollar a day.

Reply on news.ycombinator.com