> You will notice that many of these tasks are just interacting with the lazy Silicon Valley dipshit treat ecosystem that already exists.
Honestly I'm hesitant about AI devices as well but dismissing it entirely is very Luddite-like. Will these smart AI devices succeed? Who knows... It's a risk but I can absolutely see areas where these devices will succeed - like bringing tech to the tech illiterate (or even just illiterate).
And Rabbit's price-point puts it in a very competitive position to be the iPhone of AI devices. Worth watching to see how these things continue to develop.
Luddites aren’t anti-tech but pro-worker, so being skeptical of AI is a fine thing to Ludd.
Think of the ophthalmologist who wears glasses but sells you LASIK surgery. (He's skeptical of it due to an intimate understanding of the risks.) And now they're complaining that a type of LASIK surgery is being inflicted on everyone, as all consumer devices shoot beams at your eyes to auto-correct your vision.
I mean, I get it, someone thinks they're going to be the AI iPhone and become the next Apple. What I think they're missing is many people enjoy using their phones to watch cat videos on Tik Tok, and very few people want to use their communicator badge to sell bitcoin.
This is a startup and tech community. What's the point of Dropbox when "you can already build such a system yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem"?
Just because there's alternatives doesn't mean this product doesn't have a space in the market.
He uses a dedicated GPS device in his car and it speaks directions very slowly. Seems to work for him.
Basically, if you want to make a medical assistive device, you kind of have to be all in on it, instead of trying to convince everyone that your device is the future of human-computer interaction (because its probably not and you'll just forget and ignore those communities)
I often hear on HN how Apple's lockdown of iOS is justified because grandma/grandpa need it that way. The only grandpa I've ever seen featured in an Apple product presentation is Tim Cook.
And where it doesn’t, an app can add these exact features to devices people already own, since the LLM isn’t running on device on the rabbit (it’s far too small)
Also this product definitely isn’t being positioned as an accessibility device.
I'll stand by the argument that the device is useless, because it does not appear to bring anything new to the table, "think of the children^Wdisabled" arguments aside.
It's not a popular view, but I hope Teenage Engineering is acquired by a very large company soon so that they can be relegated to functioning somewhere as a harmless novelty devision.
This way they will get a big pay day and be treated like a prize, while their influence to create proprietary devices for the landfill is effectively quashed.
I don’t think that’s a common outcome of being “acquired by a very large company”.
Very large companies tend to only run large projects, even if they run them as hobbies. They simply cannot have small divisions.
As a simple example, if BigCo made this thing, they’d want to translate it in a gazillion languages, vet it for not stepping on some country’s or minority’s toes, set up support pages, have their IP lawyers go over it, etc.
Another positive-ish thing to say would be is that this device is likely going to increase in value. Not because its popular but because it is going to be rare.
I admire your dedication to finding positive things, but if this particular cloud-dependent chunk of plastic appreciates in value I’ll eat my Audrey.
It does beg the question of why it's implemented in hardware though, and not just an app on your phone.
It sounds great on paper. If I were selling widgets, I might want an API so people have yet another avenue to buy widgets and put money in my pocket. But at that point I’ve lost control of the widget-buying experience and I can’t collect as much data on the audience of widget buyers.
Public APIs disappeared as quickly as they showed up and companies focused on partnerships. You can’t ask your voice assistant to order you a pizza from the local joint, but if you say “I want a Pizza Hut pizza” we’ve got you covered.
I expect this product to go the same way. People will find a way to block their browser automation and the universe of things it can do will shrink until it’s a glorified timer and to-do list. But if you want a medium stuffed crust and a side of wings, we’ve got you covered.
With the limited bandwidth of voice, your interaction with the phone will be tiktokified: some algos will decide what’s spam, what’s legit, what you should see right now. Unless you can afford a flagship phone, your phone will be as worthless as google search.
Like, when you let the morons in, they become a Security Hazard.
But if this device, can keep tech illiterates, from using and interfacing with the tech we need to secure, All the better.
https://www.youtube.com/watch?v=S8s9uzPIqQ4
Yuki did nothing wrong.
I wonder if it's possible to homebrew this odd little square.
What this based guy doesn't realize is that recognizing that complex neural networks are "conscious" is precisely the way to stop companies from abusing them for everything.
Also, is it just code? Can you show me the code that lets an LLM reason in the way it does? You'd have just as much luck pointing to the neurons that let a human do that. If it WAS just code, it'd be GOFAI (Good Old Fashioned AI) -- a literal algorithm. That's code. And that'd be better since it would be something we could understand and troubleshoot, not a magic black box of brute forced insane computing power that megacorps and nation states can monopolize.
This kind of take, which is common with AI, always assumes this odd state of blind trust. As if I can't read and reason about the output like I would anything (or anyone) else's output. Are there people I trust more than AI? of course. Do I trust AI more than some people? also true. Would I blindly trust a person or AI without doing some of my own reasoning at some point in the process? no.
Some of this will probably shake out to be real long term but a lot less than people think right now.
I think you're making a lot of assumptions about how all this stuff works out, tbh. I'm kind of assuming it'll be like the last three AI bubbles (that time in the late 90s when voice was going to be your primary interface to Windows vNext, the self-driving car/CV hype early last decade, and the brief chatbot mania of 2016 that was killed off with extreme prejudice by Microsoft Tay); much sound and fury, but probably rather limited lasting effects once the dust settles.
Like, it seems rather unlikely to me that that main way people book holidays in a decade will be by talking to a machine and having it select hotels and 'cool SUVs' (trick question, no such thing) for them. If nothing else, it is unclear how this would be meaningfully different from just choosing the first hotel from Expedia/whatever matching the search filter.
For me to trust this it would need some kind of Lockheed Martin style engineering and 4 decades of development with mathematical proofs that it works and can't not work.
But in the end the device would cost $50,000 and be 35 years out of date.
To your point: https://en.wikipedia.org/wiki/The_Feeling_of_Power
Good to see some form of innovation. Maybe they will create a new category of device, as dozens of Chinese imitators come up with clones? Either way this is all positive.
A device like this with the battery capacity and processing power for running local AI would be very good for privacy.
I could see some people buying it, if someone advertised it as totally private, with no logs or sensitive data stored on even the device itself (i.e. in RAM and gone on power off). So you can ask it anything, it will answer, and there is no trace at all once it's switched off.
I would be astounded, astonished, and amazed if Apple doesn't have Siri++ in development to do exactly this, but with more guardrails.
So when A.I. gets hacked, it can start shit-posting on all your accounts? Sounds as useful as lastpass...
That sounds like... something that would make my life slightly better, in a non-essential but desirable way.
Can my phone do this already? When will it be able to do it?
There are websites that have databases of recipes you can sort by ingredient, which is a much better approach to this problem.
But I don't see a world where you can feasibly support the required infra of providing LLMs for your users with only hardware sales as a revenue stream. Their website specifically calls out that there is no subscription fee, so either they'll introduce one later or something else will need to be the other source of revenue.
Examples: ordering an uber, playing a song on spotify, ordering a pizza.
As others have suggested, this is novel, and ML in everything will be coming to products soon, but this isn't it.
And I like that it's "push to talk". It listens on my command, not on the whims of a voice query in mid-air.
I'll give this a crack because I'm part of the weird population that likes "stand-alone" appliances for things. It's got a certain charm to it that the workflow just would work better for me and my brain that an app alone just can't touch.
Every new thing has these sorts of articles written about it, they’re best to just ignore.
I do love how people complain about clickbait yet ragebait appears in the front page regularly. There you go - that’s why clickbait is a thing. Techies are not immune
Obviously people will sometimes express their opinions on societal trends and products, especially in their blogs.
