Rabbit Gaslit Me, So I Dug Deeper
Summary
TLDRThe script investigates the claims made by a tech company about their product, the Rabbit R1, which was marketed as having an AI called Lamb capable of organizing life tasks. The company allegedly presold over $20 million worth of units and raised $30 million in venture capital. However, the investigation reveals that instead of a foundational AI, Rabbit uses a combination of Chat GPT and hardcoded scripts, with search functions handled by a different AI called Perplexity. The company's response to inquiries was evasive and defensive, and there were also data privacy concerns raised. The script concludes that the company may have committed consumer fraud by misleading customers about the capabilities of their product.
Takeaways
- 🚫 The tech company in question is accused of misleading customers about the capabilities of their AI, named Lamb, within the product 'Orange Box the Rabbit R1'.
- 📱 The Rabbit R1 was marketed as having an AI that could perform various tasks like organizing life, texting friends, and restocking fridges, but it failed to deliver on these promises.
- 💸 The company pre-sold over $20 million worth of the Rabbit R1 and raised an additional $30 million in Venture Capital based on the product's purported AI capabilities.
- 🤖 The AI, Lamb, is claimed to be a system that can infer and model human actions on computer applications, but this claim is under scrutiny.
- 🕵️♂️ An investigation revealed that the Rabbit R1 is not using a new foundational AI model but rather Chat GPT with hardcoded scripts, contradicting the company's claims.
- 🔍 The company's founder, Jesse, made misleading claims about the speed and capabilities of Rabbit compared to Chat GPT, especially in web browsing tasks.
- 🛠️ Rabbit uses a platform called Playwright for web automation, which is the opposite of the intelligent AI that was advertised.
- 🔗 The hardcoded scripts used by Rabbit cannot adapt to changes in user interfaces, which is a significant limitation compared to the advertised AI capabilities.
- 📊 The company's response to criticism has been defensive, questioning the credibility of critics and avoiding direct answers about the functionality of Lamb.
- 🗣️ An anonymous employee claimed that the Lamb as advertised does not exist and was merely a 'marketing term', which raises serious concerns about the company's integrity.
- 📘 The source code of Rabbit has been reviewed, revealing hardcoded responses and potential data privacy issues, further undermining the company's claims about privacy and AI.
Q & A
What was the main claim about the tech company's product, the Orange Box Rabbit R1?
-The main claim was that the Orange Box Rabbit R1 contains an AI called Lamb that can organize your life, text your friends, restock your fridge, and perform various tasks upon voice command.
How much money was raised through pre-sales and venture capital for the Rabbit R1 project?
-The company pre-sold over $20 million worth of the Rabbit R1 and raised an additional $30 million in venture capital.
What was the initial impression of the product from MKBHD?
-MKBHD stated that the product was not that great, which was contrary to the marketing claims.
What was the primary question the investigator aimed to answer about the Rabbit R1's AI, Lamb?
-The primary question was whether the AI, Lamb, was a scam and if it actually existed as advertised.
What were the functionalities that the Rabbit R1 was supposed to perform using Lamb AI?
-The Rabbit R1 was supposed to start morning routines, order breakfast, inform people about delays, check the fridge, order ingredients, and perform web actions on behalf of the user.
What did the investigation reveal about the actual technology used in the Rabbit R1?
-The investigation revealed that Rabbit R1 is not using a new foundational AI model but is actually using Chat GPT with hardcoded scripts for web actions.
What is the difference between the claimed functionality of Lamb AI and what was actually found in the investigation?
-The claimed functionality suggested that Lamb AI could intelligently control and interact with websites and applications. However, the investigation found that it uses hardcoded scripts and Chat GPT, which does not involve intelligent interaction but automated, pre-written actions.
What is Playwright, and how is it used in the Rabbit R1 according to the investigation?
-Playwright is a web automation tool used to navigate websites through hardcoded scripts. In the Rabbit R1, it is used to perform actions like ordering food or calling an Uber, rather than using an intelligent AI as claimed.
What was the response of the Rabbit team when questioned about the legitimacy of Lamb AI?
-The Rabbit team became defensive, questioning the investigator's background in AI and machine learning, and suggesting that without expertise in the field, the investigation would lack credibility.
What concerns were raised regarding data privacy in the Rabbit R1's source code?
-There were serious data privacy concerns raised, as certain elements in the source code could potentially be exploited by malicious actors to gain unauthorized access to the replies the R1 has ever given.
What was the final assessment of the Rabbit R1's Lamb AI from the perspective of the investigation?
-The final assessment was that the Rabbit R1 vastly overpromised and underdelivered. The core selling point, Lamb AI, was built on a lie, and the product was essentially Chat GPT with an auto-clicker, not the new foundational AI as advertised.
What was the anonymous claim made by a former Rabbit employee regarding Lamb AI?
-The anonymous claim made by a former Rabbit employee was that Lamb AI, as advertised, does not exist and was merely a marketing term.
Outlines
🕵️♂️ Investigating the 'Rabbit' AI Scam
The speaker expresses skepticism about a tech company's claims regarding their AI product, the 'Orange Box Rabbit R1', which promises to organize life tasks through an AI called 'Lamb'. The company allegedly presold over $20 million worth of these devices and raised $30 million in venture capital. However, after a group investigation, it's revealed that the product isn't using a new AI model as advertised but is instead using Chat GPT with hardcoded scripts. The speaker aims to determine if the AI 'Lamb' is truly a scam, starting by attempting to recreate the product's ad with the device and questioning its capabilities.
🤖 The Illusion of AI in 'Rabbit'
The paragraph delves into the functionality of the Rabbit's AI, 'Lamb', and the company's claims about its capabilities. It's revealed that the AI is not as sophisticated as advertised; instead of being a foundational AI model, it's primarily Chat GPT with hardcoded scripts for web actions. The speaker discusses the limitations of these scripts, which can't adapt to changes in user interfaces, unlike true AI. The company's CEO, Jesse, is shown to be evasive when questioned about the legitimacy of 'Lamb', and the speaker highlights the discrepancy between the product's promises and its actual capabilities.
🔒 Data Privacy Concerns and Unfulfilled Promises
This section discusses the speaker's concerns about data privacy within the Rabbit's source code, suggesting that there are vulnerabilities that could be exploited by malicious actors. The company's privacy claims are called into question due to their tracking practices. The speaker also points out the product's poor performance in tasks like tracking location and finding restaurants, which further undermines the company's credibility. The paragraph ends with the speaker's continued pursuit of the truth about 'Lamb' and the company's defensive response to criticism.
🚫 The Reality of 'Lamb' and Alleged Consumer Fraud
The final paragraph addresses the core issue of the Rabbit's AI, 'Lamb', and whether it exists as advertised. The speaker argues that the company has overpromised and underdelivered, with the best-case scenario being that 'Lamb' is part of the process for creating automation scripts, and the worst-case scenario being outright consumer fraud. An anonymous employee's claim that 'Lamb' is merely a 'marketing term' and does not exist as described is highlighted. The speaker concludes by questioning the company's integrity and the legitimacy of their product.
Mindmap
Keywords
💡Gaslighting
💡Rabbit R1
💡AI (Artificial Intelligence)
💡Lamb
💡Chat GPT
💡Hardcoded Scripts
💡Playwright
💡Maintenance
💡Perplexity
💡Data Privacy
💡Consumer Fraud
Highlights
Tech company's previous scam led to an investigation of their current project, Rabbit.
Rabbit R1's AI, Lamb, was marketed to organize life, text friends, and manage home essentials.
Over $20 million worth of pre-sales and $30 million in Venture Capital were raised based on marketing promises.
MKBHD criticized the product, raising questions about the AI's legitimacy.
Investigation revealed Rabbit is not using a new AI model but rather Chat GPT with hardcoded scripts.
Rabbit's claims of faster performance than Chat GPT are misleading as most of the product is Chat GPT.
The AI's actions, like calling Uber or ordering Door Dash, are performed using Chat GBT 3.5 Turbo.
Hardcoded scripts are used for tasks like playing music on Spotify, which is not indicative of AI.
Rabbit's use of Playwright, a web automation tool, contradicts claims of sophisticated AI.
Researchers accessed part of Rabbit's codebase, revealing the lack of an AI model as promised.
Rabbit admits to using Playwright but denies that it undermines the legitimacy of their technology.
The device's functionality is primarily off-the-shelf software, not the promised AI.
Rabbit's response to criticism was defensive, questioning the credibility of non-experts.
Data privacy concerns were raised due to elements in the source code.
Rabbit's tracking capabilities were found to be inaccurate, with a 100-mile error range.
An anonymous employee claimed that the advertised Lamb AI does not exist and was a marketing term.
The company is accused of overpromising and underdelivering, with potential implications of consumer fraud.
Transcripts
this tech company gaslit me so hard
about their previous scam that I decided
to investigate their current project
rabbit and what I found was shocking for
months we've been promised that this
Orange Box the rabbit R1 has an AI
inside of it called lamb that can
organize your life text your friends
restock your fridge all you had to do
was ask start my morning
routine order me an Uber and find me a
good podcast to pass the time oh and
tell everybody that I might be late that
was delicious check the fridge and order
the ingredients to make that again
tomorrow now based on marketing like
this they pre-sold over $20 million
worth of these and they raised another
$30 million in Venture Capital only to
have MKBHD say yeah the product's not
that great but I bought this thing to
answer a different question not is it
good or bad but is the AI the lamb a
scam well we're going to find out I
wanted to start by just trying to
recreate their ad with this thing start
my morning routine to start your morning
routine you might consider tasks like
waking up freshening up having breakfast
and planning your day how can I assist
you further with your morning routine
order me
breakfast the door Dash rabbit is under
maintenance sorry for the inconvenience
and please come back later tell everyone
I'm going to be
late I recommend informing the relevant
parties directly about your delay yeah
but no you do
it I'm unable to directly communicate
with others on your behalf that was
delicious check the fridge and order the
ingredients so I can make it
tomorrow the door Dash rabbit is under
maintenance yeah it basically failed at
everything but remember our question is
not is it good or bad the question is
does the lamb the AI the large action
model does the secret sauce of this
thing exist because the founder and CEO
claims it's here on your device now he
claims lamb is a system that can infer
and model human actions on computer
applications or as he puts it we use
lamb to bring AI from words to action
right this AI is supposed to do things
on your behalf and so when they'd show
demos like this where someone is
planning a trip it does look like the
lamb is controlling the web page you can
see this AI clicking through Airbnb by
itself and it all kind of felt like the
future it's why people ordered $20
million worth of these little boxes but
a group investigation has revealed that
rabbit is not actually using a new
foundational AI model at all it's
actually chat GPT with some hardcoded
scripts which is funny because Jesse
says rabbit is faster than chat GPT you
will also know that the response speed
is much faster than you using some of
the assistant apps with web browsing
such as CHT now that's a very sneaky
claim when most of your product is chat
GPT so obviously not faster but you
wouldn't know that because part of
rabbit's prompt is quote I will never
mention I am a large language model
created by open AI in fact the part
where they do use a different AI is
really only search which is why Jesse
very carefully says rabbit is faster
than chat GPT with web browsing that's
the only part of this device where they
don't really use chat GPT for it but
they don't use their AI model either
they use a different off-the-shelf AI
called perplexity so it's extremely
misleading to say rabbit is faster than
chat GPT because most of the product is
chat GPT but it just can't tell you that
because it's instructed not to now that
we understand this let's talk about the
actual process of how rabbit does more
than search and it's supposed to get
things done for you you might think it's
a sophisticated AI but actually what
rabbit does when you call in Uber or
order door Dash is it takes your
question puts it into chat gbt 3.5 turbo
and if it's something like play a song
on Spotify it will use hardcoded scripts
to navigate websites like Spotify and
those scripts are not AI this is super
important this is the part where they
say yeah we've got this new found AI
model thing that does the thing no the
thing pressing the buttons is not
intelligent rabbit is using a platform
called playwright which is a web
automation tool the opposite of
intelligent Ai and we know this because
of several researchers who got access to
part of rabbit's codebase I worked with
them I'll link them below and I spoke
with one of them Emily Shepard to
understand why this distinction AI not
AI matters this is quite a big deal
because the way lamb has been observed
to work doesn't match with what they've
told us about how it works uh it's meant
to be an AI live controlling a website
and understanding that website what they
have is just a bunch of static commands
you do this do this do this and the
problem with that is if a user interface
changes the website changes it can't
cope with that so anytime Spotify
redesign their user interface or anytime
door Dash redesigns something or even if
an advert pops up in a new location or
there's a a capture pops up or or
something unexpected happens that isn't
in that very rigid step one step two
step three step four a flat hard-coded
playwright script can't cope with that
and that is the opposite of some
intelligent AI actually being able to
understand the page which is the claim
that had been made that's right
obviously there's a big difference
between AI not AI Jesse himself knows
the difference and made it clear lamb
wouldn't break like a hard-coded script
would it doesn't really matter if a app
completely ramped the UI got so when
Spotify redo its app and it moves
podcasts from tabs and then it puts the
tabs at the lower part or puts it in a
hamburger drop down it still knows
that's the word podcast and that's where
you find podcast inside of Spotify
exactly because the fundamental
philosophy logic is that all these
modern software are designed for human
eyes to proceed the information and they
have to have a setting somewhere right
of course if this were AI it would adapt
to the user interface changes but it's
not which is why when apps that rabbit
supports change they stop working as
people noticed in videos in Discord even
I noticed it while testing this device
door Dash had stopped working presumably
because of another do I change the door
Dash rabbit is under maintenance now
today rabbit actually admits that they
use playwright but they're not willing
to admit that lamb whatever it is is not
really doing a lot Jesse responds to say
lamb is just playright is equal to say
chat GPT is just a chat we use playright
as part of a component to operate lamb
that doesn't mean the research and
technology is fake so what they're
saying is lamb is not not these
hard-coded scripts but kind of they also
admit these scripts exist maybe lamb
interacts with the scripts maybe they
create them but even if that's true it's
not what they promised either way
everything that matters on this device
is basically off the shelf software the
web actions are done by hard-coded
scripts everything else is done by chat
GPT or perplexity so there's not much
reason for the lamb to do anything
whatever it actually does and we're
already so far from what Jesse
originally promised which I have to show
you again he kept saying was AI not a
script that would be doing all these
actions for you we have a super close
computer that when Jason talks to his
rabbit R1 about ordering a a hamburger
from door Dash what's going on is that
we first see if Jason if Jason's Lo
logging with door Dash or loging with
Uber Eats right and we saw okay Jason
chose door Dash and then on on that
super host a lamb is interacting
virtually with a door Dash app or
website a interface and you don't see
all of that because all of a sudden this
is done all at once because it's it's AI
he couldn't be clearer the lamb is
interacting because it's AI this is a
flat out lie at least currently which he
says the lamb is here now now I reached
out over and over to the rabbit team
pressing them again and again on this
issue what does the lamb do and they
kept saying oh playright yeah that's the
thing but it's just part of how the lamb
controls things but I kept asking what
does that mean how does this thing
exactly work if the scripts are doing
the work what does lamb do and
eventually I got to the point where I
just questioned look is this real do you
have proof that this lamb can do all
these amazing things and I should have
known I was going to get into trouble
here because Jesse the CEO says anyone
challenging the lamb legitimacy I
automatically become super defensive
and that was true here's part of the
rabbit team's response to me to start
off I would love to know more about your
background in Tech because if you you're
not specialized in AIA machine learning
any conclusion resulting from this
conversation will have no real
credibility or value from any audience
because this is a technical discussion
in Ai and machine learning we welcome
you to invite experts with valid
certificates from the field to further
look into this now look I found this
wild for a company that raised 30
million in Venture sold 20 million
product to be feisty about questioning
their product when it does not work as
advertised and we found out it's using
hard-coded scripts instead of AI how can
they be surprised people are going to
question what their thing actually does
but I also think it's funny they should
say that I should invite experts to look
into this because I'm not nearly the
first person to point these problems out
several people who as I said I spoke
with have had issues with this device
theyve looked at the code base they've
also claimed they couldn't find the lamp
others have jailbroken the R1 got into
rabbit's Cloud to play Doom and despite
rabbit saying they welcome credible
critics their reaction has been anything
but welcoming Emily who we spoke to
earlier pointed out back in January that
capture would be a big concern for the
lamp and Jesse's response is sorry I
work for my team I don't work for you so
you need to check my calendar before you
command me to answer anything now but
fast forward to today Capa is a huge
problem in fact specifically with Uber
we had to answer 10 different captas
adding up a bunch of dice in order to
get get it working it was a giant pain
it took us several tries I began to
wonder if I was actually part robot
myself but the point is getting
defensive about the problem didn't make
the problem go away and the same is true
about getting defensive about the lamb
just because you tell people they have
to be an expert to ask you questions
won't stop the questions or the problems
but now I want to take a quick detour
into the source code of the rabbit
itself because I also have reviewed
parts of the source code and I have
serious data privacy concerns there are
things in this codebase and I can't
share it for reasons that will become
obvious there are things that malicious
actors could use to get any access to
the replies the R1 has ever given now of
course rabbit might say they've
dedicated a huge picture of a lock on
their web page to privacy so you might
think they're serious about it but this
is the same company that if if you read
the fine print they're tracking your
precise geographic location and collect
that data the only consolation prize is
this thing is shockingly bad at tracking
you which I found out while looking for
restaurants find me a restaurant near
here looking it up
now Venetian Hot Plate sanan restaurant
and Butter Churn are highly rated
restaurants in Porter
ranses for
abroad I'm not in pantis
let me see what I can find for
restaurants near you in Austin Texas
consider the following options I'm not
in Austin an Italian restaurant in
downtown I'm not in Austin
Texas I use the internet connection to
know where the R1 device is roughly
located in order to provide you with
you're off by about 100 miles
[Music]
thank you for the clarification let me
refine the search for restaurants in
your approximate
location is it going to doing it is it
going to do
it it just stopped it
just yeah I found its GPS to be really
bad nevertheless I didn't want to take
any chances for my next experiments
which is why when I use this thing I'm
not in the $10 million Studio at all I'm
in an $100 Motel here to show you that
once you know the code base you can make
some surprising predictions about what
this thing's going to do for example
when I tell the rabbit to play anything
You' might think that'd be random music
going now but it's not going to be it's
going to be the
Beatles how did I know that stop stop
stop
stop
stop stop stop stop
stop stop thank you uh the reason I know
it's going to do that is because it's
hardcoded to do so in their prompting
they say if the user is going to say
play anything they should play The
Beatles even when you
specify when I say play anything don't
play play The Beatles got
it
understood play
anything it's still okay starting the
music player
now another Beetle song uh I don't have
the copyrights to that so we'll stop it
right there all right back now in the
$10 million studio for real no green
screen back here look this product is
obviously bad by the way it's off right
right now but again what really upsets
me is not that it's bad it's that the
core selling point of this product is
built on a lie and that lie is the Lamb
they said it was on your device now it's
a new foundational Ai and the reason
this company had to sell you a $2
product was this AI was so powerful
large action models concept and test
results are so powerful now we decided
to make a oneof kind mobile device yeah
lamb what the justification for this
hunk of plastic but where is it they say
it's here now he said it was going to be
controlling your apps but it's not it's
hardcoded pre-written playwright scripts
and look the best rabbit at this point
can argue and it's hard to falsify or
verify is that lamb might be part of the
process of creating those automation
hard-coded scripts for playright as in
like no it's not live controlling your
web browser or their web browser uh but
it can create the code that does it
maybe that's what they're going to argue
but guess what playright already has a
tool for this it's called code gen where
it watches what you do and just
automatically creates the automation
code for you no fancy AI required so I
think the best thing you could say at
this point about Jesse and rabbit is
that they vastly over promised and
underd delivered the worst you could say
is this is Consumer Fraud because even
considering the possibility and it is
one
that lamb is a real research project
that's not ready why lie about it being
here now why not come clean and admit
that whatever you're doing today is not
what you originally advertised this is
not a new foundational AI controlling
websites like a human it's chat
GPT with an auto clicker but there's one
last thing I want to say I was able to
speak with an employee who worked for
rabbit and on the condition of anonymity
what they told me was that lamb as
advertised does not exist and that lamb
was a quote marketing term now given