this post was submitted on 02 Apr 2024
871 points (100.0% liked)

196

16416 readers
1787 users here now

Be sure to follow the rule before you head out.

Rule: You must post before you leave.

^other^ ^rules^

founded 1 year ago
MODERATORS
 
top 50 comments
sorted by: hot top controversial new old
[–] [email protected] 14 points 7 months ago (1 children)
[–] [email protected] 2 points 7 months ago

For cruise, yes

[–] [email protected] 3 points 7 months ago
[–] [email protected] 12 points 7 months ago (1 children)

The self-driving cars secretly being driven by somebody is literally the plot to Captain Laserhawk

[–] [email protected] 11 points 7 months ago

It's also already happening. Waymo admitted to needing wayyyy more manual interventions than they let on

[–] [email protected] 11 points 7 months ago (1 children)

I mean ask Apple or Google how many people listen to their voice systems to manually improve them for accuracy...you have to train the AI somehow

[–] [email protected] 7 points 7 months ago (2 children)

Yeah except with there being only 27 stores in the US that use this tech if you have 1000 people reviewing the purchases is it really a machine learning system or are you just outsourcing the process to people in another country.

[–] [email protected] 1 points 7 months ago

You still have to keep training the Model. These stores were in large busy markets, and having people watch and critique the AI is how they continually train the model. It took Apple over 8 years to 'announce' they're doing on device voice recognition(they probably aren't), and that was just voice recognition and LLM training vs image recognition which is hard on its own. Let alone tracking a person THROUGH a store, recognizing that someone picked something up and took it vs put it back or left it on another row.

The real reason this probably happened is because those 1000 people training the model reported metrics of failures on top of the stores showing losses due to error. The margin of error was probably greater than they wanted. Or add in the biometric data they had integrated into it adding more layers of cost and privacy protection...it probably just doesn't return the money they wanted and they'll try again in a few years probably utilizing more RFID on top of the image recognition and people tracking.

[–] [email protected] 6 points 7 months ago

Recommendations for rectification:

Stop being a Chargers fan. Jesus christ look at what the Spanos have done to you.

[–] [email protected] 5 points 7 months ago
[–] [email protected] 13 points 7 months ago (1 children)

Waymo and Cruise definitely have humans waiting to take over if the autonomous mode gets stuck... so that's already true!

[–] [email protected] 6 points 7 months ago

Its not about automation; its about making labor fungible, kinda like the garmet industry-local workers rights happen, call center instantly relocates.

This is the globalization you get while jacking off with the monkey paw.

[–] [email protected] 8 points 7 months ago* (last edited 7 months ago)

If it's true that's just plain dystopian sci-fi fking wrong.

Forget mad max and those movies reality is worse

[–] [email protected] 27 points 7 months ago
[–] [email protected] 36 points 7 months ago

Tfw I accidentally alt-tab from GTA to Drive From Home application and launch passenger's taxi off the ramp for stunt bonus.

[–] [email protected] 10 points 7 months ago

You can't just steer people into bridge abutments

So thaaaaaat's what happened with the Francis Scott Key Bridge.

[–] [email protected] 13 points 7 months ago (2 children)

How long is it gonna take for them to use data created by those indians to train their AI model and replace them?

[–] [email protected] 17 points 7 months ago

They've surely started working on it already. Current "AI" (LLMs) aren't perfect. They require constant human adjustments.

I'm an auditor for a "machine learning" algorithm's work, and it develops new incorrect processes faster than it corrects them. This is because corrections require intervention, which involves a whole chain of humans, whereas learning new mistakes can happen seemingly spontaneously. The premise of machine learning is that it changes over time, but it has no idea which changes were good until it gets feedback.

So, to answer your question, I'm sure they're throwing a ton of money at that. But when will it be viable, if ever?

[–] [email protected] 5 points 7 months ago

Long enough for people to organize or nah?

[–] [email protected] 46 points 7 months ago (4 children)

I'm all for casheirs and taxi drivers working from home if this translates to that.

[–] [email protected] 3 points 7 months ago

Everyone knows that in the near future all taxi driving will be relegated to Robert Picardo.

[–] [email protected] 8 points 7 months ago

Cashiers, sure. Drivers? No thank you. If we get into a situation, I want to make sure they feel the urgency with me.

[–] [email protected] 7 points 7 months ago (2 children)

It's going to translate into a trained AI that takes all of those jobs.

[–] [email protected] 1 points 7 months ago

You mean the thing they were advertising in the first place? That's some false advertising inception. Somehow, that only makes it worse.

[–] [email protected] 7 points 7 months ago

Will knowledge workers e-riot hard enough for UBI?

Tell me Y Combi will email their buddies again… they better!

[–] [email protected] 26 points 7 months ago (1 children)

You know, as someone who's been a cashier and a taxi driver, I can see both of those being vaguely plausible.

Probably moreso if businesses actually embraced remote work.

[–] [email protected] 26 points 7 months ago (4 children)

I swear I remember hearing about a science fiction story where the "self-driving" cars actually just had people hidden in them.

[–] [email protected] 1 points 7 months ago

Wrong Way by Joanne McNeil

[–] [email protected] 5 points 7 months ago

Slow march to 40k servitors not looking far fetched. Why develop an AI/algorithm when you have no moral conscience and can just make a human do it.

[–] [email protected] 7 points 7 months ago (1 children)

scifi dystopia future where a race of gnomes are created to sit inside things and play mechanical turk

[–] [email protected] 5 points 7 months ago* (last edited 7 months ago)

I watched this sketch comedy show in Iceland as a kid, where every week they had a section called "the men behind the curtains". It was just people hidden away inside ATMs, vending machines etc. pretending it was a machine doing the work.

[–] [email protected] 3 points 7 months ago* (last edited 7 months ago) (1 children)
[–] [email protected] 2 points 7 months ago (1 children)

No, there are no cars in that, although I get the connection.

This was a short story or novel I think, but I haven't read it just heard about it.

[–] [email protected] 1 points 7 months ago

Makes sense, I was just joking :)

[–] [email protected] 13 points 7 months ago

I remember trying MTurk out back in the day to try and make money on the side. It's such a mind numbing activity. Doesn't surprise me that this is still the model for smart "automated" systems like this.

[–] [email protected] 18 points 7 months ago (2 children)

If it were Amazon, you know that at least 150 out of those 1000 workers had already been threatened with PIP before being put on a plan. Look up Focus and Pivot, Amazon's policy that puts around ~5-15% of corporate workers below director level a year on forced attrition.

[–] [email protected] 3 points 7 months ago* (last edited 7 months ago)

Focus and Pivot

This is also known as "stack ranking" and "rank and yank".

It's a super-gross way to run a business. I can see how you might want to "cut the fat" when starting out or growing. But keeping a policy like that for the long haul means selecting for employees that are good at that surviving. And that may not require one to even be all that productive, just good at working the system.

Anecdotes: https://news.ycombinator.com/item?id=4195136

It's also a recipe for a toxic work environment:

https://www.cultureamp.com/blog/what-is-stack-ranking

[–] [email protected] 18 points 7 months ago (1 children)

Nah, this was definitely outsourced to a company in India; you can abuse contract/vendor employees with way less effort than it takes to abuse full-time employees.

[–] [email protected] 4 points 7 months ago (1 children)

Perhaps, although many labelling teams at Amazon for other orgs are in-house or are part-time hours. Amazon likes keeping things in-house because they don't particularly give a fuck about abusing staff.

[–] [email protected] 2 points 7 months ago (2 children)

No, I can assure you even if they work directly in the building there is no way there isn't a middle man contract holder that allows the immediate firing of employees without effort. I work in this kind of tech. The goal is to have the least amount of actual employees as possible for these companies because it gives the least liability and the fastest route to letting them go.

Amazon does this with their delivery drivers what makes you think they aren't making sure there is a middle contract holder here too.

[–] [email protected] 2 points 7 months ago (1 children)

Previously worked for Amazon data labeling. The priority was largely having everything done in-house due to privacy concerns. It's a lot easier to act on privacy leaks coming from within. That said, the long term strategy is using crowdsourced labeling for anything not having to do with customers or customer data. So looks like you're both right :)

[–] [email protected] 1 points 7 months ago
[–] [email protected] 1 points 7 months ago

Mostly because I work for Amazon, and can both see the org structure and know how it works in other orgs here.

load more comments
view more: next ›