I resent AI. Not AI itself–that’s just code, despite what tech guys with flashlights under their chins tell you. I resent the imposition, the idea that since LLMs exist, it follows that they should exist in every facet in my life. And that’s why, on principle, I really hate the rabbit r1.
The rabbit r1 is an AI gadget and voice assistant that uses Large Language Models and what rabbit calls a “Large Action Model” to make complex decisions. It’s kinda like an AI walkie talkie. If you ask an LLMs like GPT-4 to suggest a flight, it might do a reasonable job but won't book it for you. What the people at rabbit are pitching is a device that logs into all your services like Uber and Expedia and then lets you interface with them all at once through an LLM and agents called "rabbits," by way of a little device that does not contain those actual apps.
At first you may say to yourself, “Hey this thing looks kinda like a Playdate” and you wouldn't be far off. Teenage Engineering, the company behind a lot of nice-looking but comically overpriced audio gear (minus the pocket operator), did the design for both Panic’s Playdate and the rabbit r1 (rabbit’s CEO is on Teenage Engineering’s board of directors). But the difference is the Playdate is a fun, charming, intentionally simple little console that you play for five minutes whenever you remember you own it, and the rabbit is a cretinous imposition that I don’t ever want to hear about again after this blog.
If you are a masochist like me and watch this entire keynote from CEO Jesse Lyu, you will notice a few things immediately. First is the brazen, warmed-over Steve Jobs presentation, and while aping Apple is basically standard practice in the tech world, the degree to which it is done badly, coupled with the implication that this is somehow the second coming of Jobs, makes the entire thing embarrassing to watch. Another is that every single use case presented in the keynote is either generating useless LLM vomit, using an app, using several apps at once, doing stuff that Siri already does faster, a worse version of webhooks, or doing the laziest shit I can think of with computers. Here is a rough summary of what Jesse asks this demon square to do:
- What’s the nature of reality?
- What is the stock price of Coca-Cola?
- Who played the role of Oppenheimer?
- Play Kraftwerk’s “Pocket Calculator”
- Play another song from that album.
- Who wrote the lyrics?
- Which band sampled this song?
- What do you think about this song?
- Get me a ride from my office.
- Find me an Uber that can fit all of us.
- Order me a 12 inch pizza from Pizza Hut, the most ordered one on the app is fine.
- I want to take my family to London, it’s going to be two of us and a child of age 12. We like cheap nonstop flights, grouped seats, a cool SUV, and a nice hotel with wi-fi.
- Could you come up with a schedule of fun things to do while I’m over there?
- It seems like this is too intense. Could you plan an easier schedule for us?
- (Takes a picture of Rick Astley and has it rickroll him.)
- (Takes picture of the contents of his fridge.) This is what I have in the fridge, can you make me a nice dish that’s low in calories? (It then pulls up a recipe without citing its source.)
- (Takes a photo of an Excel spreadsheet on his screen, asks it to add a specific column. The AI then emails him the result, and then he replies to it in an email to have it make changes, which seems way more complicated than just typing on your laptop.)
- (Uses a “teach” mode to train it to be able to quickly go into a Discord and ask a bot to make a really ugly Midjourney image of a CG dog.)
You will notice that many of these tasks are just interacting with the lazy Silicon Valley dipshit treat ecosystem that already exists. This is funny because part of Jesse’s Steve Jobs pantomime involves the repeated insistence that the app-based ecosystem is old news, but the entire MO of Silicon Valley is reimagining an existing service in a more lazy and invasive way while claiming you are doing something new, so it’s sort of fitting. The idea that you would need to have an additional layer of abstraction for these services is funny enough, but the additional premise that you are now going to be carrying two devices with redundant purposes is even funnier.
Other tasks he asks it to do are so boring and unimaginative as to call into question the character of the person asking them. On some level I understand not wanting to book a trip, hotel and travel together, but that’s historically why travel agencies exist and why apps exist to undercut travel agencies. I would not trust a large language model to make those kinds of decisions for me, nor would I ask it to plan an itinerary in a new city, because I’m not a boring or unimaginative person who lets a cheap piece of plastic tell me to do the ten most common results for “stuff to do in London.”
But most importantly, I really don’t trust AI in its current form to make important decisions for me involving time or money, particularly since the device’s whole sales pitch is that it does not have apps on the device itself and doesn’t connect to any existing APIs (Application Programming Interfaces). I had just assumed when I first saw it that it was loading apps on there, but no, it’s doing everything through algorithms trained on workflows managed via a web portal (called the rabbit hole) where both rabbit and potentially you teach an AI how to click on websites. In all of their marketing, rabbit makes a huge deal about how privacy-focused they are, how they don’t store any data, there is no subscription fee, and how the camera and mic only work when manually turned on, but that’s not the worrying part of this scenario–it’s the idea of meaningful decisions about my life being done remotely by an OS and cloud service that I don’t really trust.
What’s particularly funny is the part where Lyu asks “Why would I need a new device if I already have a $1,000 dollar iPhone?” Great fuckin’ question, my guy! I sure don’t! In part because I know that every one of these companies is already injecting that stuff into my life non-consensually and they have more money than many medium-sized countries. Plus, rabbit is more complicated to understand than the concept of an app. I would say that all these features could theoretically be covered by a single app, functionally speaking, but I am not convinced that Apple or Google would let an app that functioned like this on the app stores because that sounds like a nightmare. Maybe that’s why they needed this weird ass little freak toy thing to exist.
What’s most annoying about all of this is the sheer repeated imposition of this horseshit. I’m sick of being forced to think about generative AI, large language models and large action models. I’m tired of these adult toddlers who need an AI to tie their shoes and make bad Pixar characters for them. Microsoft and Google keep shoving AI features into their software, and I absolutely should not have to worry about this garbage from Firefox of all places. Beyond the marginal utility of use cases like transcription and realtime voice cloning for goofy bullshit jokes, the majority of this stuff has made my experience on the internet (where I hang out a lot) measurably more unpleasant. This latest push for AI is making the world lazier, less curious, harder to navigate, ripping people off, and creating a topic somehow more tiring than that year these people wouldn’t shut the fuck up about NFTs and then never brought it up ever again when the market imploded.
So go away, leave me alone forever, and take that over designed plastic garbage with you.
