I Witnessed the Future of AI, and It's a Broken Toy

https://www.theatlantic.com/technology/archive/2024/04/rabbit-r1-impressions/678226/

by mikestew • 2 years ago

39 23 2 years ago

theatlantic.com

mikestew• 2 years ago

https://archive.ph/agpzd

krimuthu• 2 years ago

Thanks for this

RevEng• 2 years ago

They showed it off to a large crowd without even testing it on site. How typical of "move fast and break things" development. Don't worry, they'll fix it eventually, but until then they look stupid and your fancy new toy is an expensive paperweight.

SushiHippie• 2 years ago

Related on the front page, a review by MKBHD:

Rabbit R1: Barely Reviewable [video] | https://news.ycombinator.com/item?id=40206063

tim333• 2 years ago

I just watched his review and I don't really get why the device exists. As in it's a box with kind of a bad camera, bad screen, bad speaker, bad battery and some connectivity that links to an AI that you can ask questions of.

Why is that not just an app on your smartphone which already has superior hardware and everyone already carries?

As one of the comments said it might have made sense in an alternative universe in which smartphones were not invented.

janalsncm• 2 years ago

MKBHD touched on this, but it seems a lot of products are coming out whose value is partially justified by the promise of future features. Tesla does this too, trying to take my 2024 dollars for a feature which may or may not arrive sometime in the future.

I call this a “tech tease”. When it arrives, I’ll pay for it. Until then, pony up for the R&D or take out a real loan or raise from VCs.

m4632• 2 years ago

Tesla goes even further. It doesn't even ship hardware anymore, shifting to software.

For example, there is no shifter stalk - software will guess if you want to go forward or reverse.

Automatic lights/high beams and automatic wipers are really annoying. yes, it is helpful sometimes, but manual control can be key for good driving. Sometimes a good safe driver wants to flash their high beams, turn off their lights, or control the interval wiper speed. Nope.

Kirby642• 2 years ago

I’d agree there’s plenty of future promises for Tesla software especially around FSD, but the examples you all listed are just not true on a Tesla today.

There’s a slider on the screen to let you pick forward or backward. No need to let software guess if you don’t want to. High beam flashing is a button on the steering wheel. And wiper control can be mapped to one of the scroll wheels, if you want manual control.

There’s no real reason to turn off headlights in a quick manner, though. I don’t think that’s possible from the steering wheel… but using the screen to disable it while you’re parked seems not particularly onerous.

rsynnott• 2 years ago

I don't think it's even that; their value is that they are somewhere for VCs to put their money that has 'AI' in the title. That's it. That's what they're for.

hysan• 2 years ago

Why take a loan when you can use customers as an interest free loan without any obligation to pay it back if you fail? \s

That’s pretty much what this ongoing trend towards releasing unfinished products is. It’s a way for companies to exploit customers even more and squeeze every last penny out of people. It’s probably worse than that too because it’s slowly shifting the norm and can result in every hype idea becoming a money grab.

throwaway5959• 2 years ago

I hope they’re successful. We need more fun gadgets. Phones, tablets and computers are all so boring and functional now.

pona-a• 2 years ago

So instead we need something that is fun and dysfunctional? Let's be honest, even with all the LLM tech, voice interfaces could never be more than a stop-gap before the user does reach for a screen. That's why smart speakers may have actually been at least very very very slightly successful, since at least they have hands-free timers/music/smart-home controls going for them.

The real problem is that tech produced by corporations, whose explicit goal is to minimize input labor/time/salaries and maximize revenue, does tend to slowly degenerate into repetitions of itself.

When software is treated as a handmade thing, it transcends its inherent value as a tool, it becomes an ornament, a sentimental object, a type of self-expression.

When you design software as a product, intent on copying your competitors exactly to satisfy the investors, even if that market niche is taken, software becomes boring.

You can build a culture out of trash, but it will only be a trash culture.

karmajunkie• 2 years ago

i mean, who doesn’t enjoy winning a lottery? now you can play every time you use your device!

(in all seriousness, voice devices are a crucial affordance or adaptation for some disabled users. i ordered one of these (somewhat impulsively i admit) to evaluate it for a family member with a neurodegenerative disease who uses alexa to a degree that seems almost insane to me but it allows him to navigate his daily activities.)

gfourfour• 2 years ago

We need more fun software

dhab• 2 years ago

This article is relating a bad experience with a push-the-envelope device on the bleeding edge, extrapolated to "Future of AI".

fnands• 2 years ago

Yeah, it just sounds like the beginning of the "trough of disillusionment" is here.

vouaobrasil2• 2 years ago

This isn't the future of AI, it's just another example of a product gone wrong.

Or perhaps "right" for us. Because everything that strengthens the connection with AI erodes the potential for human connection. Every time you ask an AI a question, you are less likely to embark upon the journey of self-discovery. If you get help from an AI, you are less likely to get help from a fellow human being, and you are less likely to HELP another human being.

How is that different from using a search enigne? Actually, not much, except that AI provides a few more pieces of the puzzle of a substitute for human connection, and brings that to the next level. It is an order of magnitude better that passive reading, and thus an order of magitude better in separating people.

When something like this succeeds, it will be another failure for humanity.

blamestross2• 2 years ago

It's almost like GenAI is the ultimate comodification of human connection.

Literally scraping the largest accessible corpus of human communication and attempting to emulate it and placing it as a MITM attack between us.

vouaobrasil2• 2 years ago

Indeed. That's an interesting viewpoint. It is a commodification by abstracting it into the mode of a computer, driven by global capitalism to transform ordinary human expression into the ultimate commodity to brace late capitalism up for one last breath of life. Or at least, I hope it is the last.

abrichr• 2 years ago

Rabbit has said the device will be able to learn any app, if you teach it.

We're building this over at https://github.com/OpenAdaptAI/OpenAdapt. OpenAdapt learns to automate tasks in desktop apps by observing human demonstrations.

Early demo: https://twitter.com/abrichr/status/1784307190062342237 (more coming soon!)

The demo is overly simplistic to keep it short -- it also works with arbitrary applications and operations.

Also, we're open source. Contributions and feedback are welcome and encouraged :)

I Witnessed the Future of AI, and It's a Broken Toy

Related Stories