back to indexColin Angle: iRobot CEO | Lex Fridman Podcast #39
link |
The following is a conversation with Colin Angle.
link |
He's the CEO and co founder of iRobot,
link |
a robotics company that for 29 years
link |
has been creating robots that operate successfully
link |
in the real world.
link |
Not as a demo or on a scale of dozens,
link |
but on a scale of thousands and millions.
link |
As of this year, iRobot has sold more than
link |
25 million robots to consumers,
link |
including the Roomba vacuum cleaning robot,
link |
the Bravo floor mopping robot,
link |
and soon the Terra lawn mowing robot.
link |
29 million robots successfully operating autonomously
link |
in real people's homes,
link |
to me is an incredible accomplishment
link |
of science, engineering, logistics,
link |
and all kinds of general entrepreneurial innovation.
link |
Most robotics companies fail.
link |
iRobot has survived and succeeded for 29 years.
link |
I spent all day at iRobot,
link |
including a long tour and conversation with Colin
link |
about the history of iRobot,
link |
and then sat down for this podcast conversation
link |
that would have been much longer
link |
if I didn't spend all day learning about
link |
and playing with the various robots
link |
and the company's history.
link |
I'll release the video of the tour separately.
link |
Colin, iRobot, its founding team, its current team,
link |
and its mission has been and continues to be
link |
an inspiration to me and thousands of engineers
link |
who are working hard to create AI systems
link |
that help real people.
link |
This is the Artificial Intelligence Podcast.
link |
If you enjoy it, subscribe on YouTube,
link |
give it five stars on iTunes,
link |
support it on Patreon,
link |
or simply connect with me on Twitter
link |
at Lex Friedman, spelled F R I D M A N.
link |
And now, here's my conversation with Colin Angle.
link |
In his 1942 short story, Runaround,
link |
from his iRobot collection, Asimov proposed
link |
the three laws of robotics in order,
link |
don't harm humans, obey orders, protect yourself.
link |
First, does the Roomba follow these three laws?
link |
And also, more seriously,
link |
what role do you hope to see robots take
link |
in modern society and in the future world?
link |
So the three laws are very thought provoking
link |
and require such a profound understanding
link |
of the world a robot lives in,
link |
the ramifications of its action and its own sense of self
link |
that it's not a relevant bar,
link |
at least it won't be a relevant bar for decades to come.
link |
And so if Roomba follows the three laws,
link |
and I believe it does,
link |
it is designed to help humans, not hurt them,
link |
it's designed to be inherently safe,
link |
and we designed it to last a long time.
link |
It's not through any AI or intent on the robot's part.
link |
It's because following the three laws
link |
is aligned with being a good robot product.
link |
So I guess it does,
link |
but not by explicit design.
link |
So then the bigger picture,
link |
what role do you hope to see robotics, robots take
link |
in what's currently mostly a world of humans?
link |
We need robots to help us continue
link |
to improve our standard of living.
link |
We need robots because the average age
link |
of humanity is increasing very quickly,
link |
and simply the number of people young enough
link |
and spry enough to care
link |
for the elder growing demographic is inadequate.
link |
And so what is the role of robots?
link |
Today, the role is to make our lives a little easier,
link |
a little cleaner, maybe a little healthier.
link |
But in time, robots are going to be the difference
link |
between real gut wrenching declines
link |
in our ability to live independently
link |
and maintain our standard of living,
link |
and a future that is the bright one
link |
where we have more control over our lives,
link |
can spend more of our time focused
link |
on activities we choose.
link |
And I'm so honored and excited
link |
to be playing a role in that journey.
link |
So you've given me a tour.
link |
It showed me some of the long histories,
link |
now 29 years that iRobot has been at it,
link |
creating some incredible robots.
link |
You showed me Pacbot.
link |
You showed me a bunch of other stuff that led up to Roomba,
link |
that led to Braava and Terra.
link |
So let's skip that incredible history
link |
in the interest of time,
link |
cause we already talked about it.
link |
I'll show this incredible footage.
link |
You mentioned elderly and robotics in society.
link |
I think the home is a fascinating place for robots to be.
link |
So where do you see robots in the home?
link |
Currently, I would say, once again,
link |
probably most homes in the world don't have a robot.
link |
So how do you see that changing?
link |
What do you think is the big initial value add
link |
that robots can do?
link |
So iRobot has sort of, over the years,
link |
narrowed in on the home, the consumer's home,
link |
as the place where we want to innovate
link |
and deliver tools that will help a home
link |
be a more automatically maintained place,
link |
a healthier place, a safer place,
link |
and perhaps even a more efficient place to be.
link |
And today, we vacuum, we mop,
link |
soon we'll be mowing your lawn.
link |
But where things are going is,
link |
when do we get to the point where the home,
link |
not just the robots that live in your home,
link |
but the home itself becomes part of a system
link |
that maintains itself and plays an active role
link |
in caring for and helping the people live in that home.
link |
And I see everything that we're doing
link |
as steps along the path toward that future.
link |
So what are the steps?
link |
So if we can summarize some of the history of Roomba,
link |
you've mentioned, and maybe you can elaborate on it,
link |
but you mentioned that the early days
link |
were really taking a robot from something that works
link |
either in the lab or something that works in the field
link |
that helps soldiers do the difficult work they do
link |
to actually be in the hands of consumers
link |
and tens of thousands, hundreds of thousands of robots
link |
that don't break down over how much people love them
link |
over months of very extensive use.
link |
So that was the big first step.
link |
And then the second big step was the ability
link |
to sense the environment, to build a map, to localize,
link |
to be able to build a picture of the home
link |
that the human can then attach labels to
link |
in terms of giving some semantic knowledge
link |
to the robot about its environment.
link |
Okay, so that's like a huge, two big, huge steps.
link |
Maybe you can comment on them,
link |
but also what is the next step
link |
of making a robot part of the home?
link |
Sure, so the goal is to make a home
link |
that takes care of itself,
link |
takes care of the people in the home,
link |
and gives the user an experience of just living their life
link |
and the home is somehow doing the right thing,
link |
turning on and off lights when you leave,
link |
cleaning up the environment.
link |
And we went from robots that were great in the lab,
link |
but were both too expensive
link |
and not sufficiently capable to ever do an acceptable job
link |
of anything other than being a toy or a curio in your home
link |
to something that was both affordable
link |
and sufficiently effective to drive,
link |
be above threshold and drive purchase intent.
link |
Now we've disrupted the entire vacuuming industry.
link |
The number one selling vacuums, for example, in the US
link |
are Roombas, so not robot vacuums, but vacuums,
link |
and that's really crazy and weird.
link |
We need to pause that. I mean, that's incredible.
link |
That's incredible that a robot
link |
is the number one selling thing that does something.
link |
Yep. Something as essential as vacuuming.
link |
Yep. Congratulations.
link |
Thank you. It's still kind of fun to say,
link |
but just because this was a crazy idea
link |
that just started, you know, in a room here,
link |
we're like, do you think we can do this?
link |
So, hey, let's give it a try.
link |
But now the robots are starting to understand their environment.
link |
And if you think about the next step,
link |
there's two dimensions.
link |
I've been working so hard since the beginning of iRobot
link |
to make robots are autonomous,
link |
that, you know, they're smart enough
link |
and understand their task enough,
link |
they can just go do it without human involvement.
link |
Now what I'm really excited and working on
link |
is how do I make them less autonomous?
link |
Meaning that the robot is supposed to be your partner,
link |
not this automaton that just goes and does what a robot does.
link |
And so that if you tell it,
link |
hey, I just dropped some flour by the fridge in the kitchen,
link |
can you deal with it?
link |
Wouldn't it be awesome if the right thing just happened
link |
based on that utterance?
link |
And to some extent, that's less autonomous
link |
because it's actually listening to you,
link |
understanding the context and intent of the sentence,
link |
mapping it against its understanding
link |
of the home it lives in and knowing what to do.
link |
And so that's an area of research.
link |
It's an area where we're starting to roll out features.
link |
You can now tell your robot to clean up the kitchen
link |
and it knows what the kitchen is and can do that.
link |
And that's sort of 1.0 of where we're going.
link |
The other cool thing is that we're starting
link |
to know where stuff is.
link |
And why is that important?
link |
Well, robots are supposed to have arms, right?
link |
Data had an arm, Rosie had an arm, Robbie the robot had an arm.
link |
I mean, robots are, you know, they are physical things
link |
that move around in an environment
link |
and they're supposed to like do work.
link |
And if you think about it,
link |
if a robot doesn't know where anything is,
link |
why should it have an arm?
link |
But with this new dawn of home understanding
link |
that we're starting to go enjoy,
link |
I know where the kitchen is.
link |
I might in the future know where the refrigerator is.
link |
I might, if I had an arm, be able to find the handle,
link |
open it and even get myself a beer.
link |
Obviously, that's one of the true dreams of robotics
link |
is to have robots bringing us a beer
link |
while we watch television.
link |
But, you know, I think that that new category of tasks
link |
where physical manipulation, robot arms,
link |
is just a potpourri of new opportunity and excitement.
link |
And you see humans as a crucial part of that.
link |
So you kind of mentioned that.
link |
And I personally find that a really compelling idea.
link |
I think full autonomy can only take us so far,
link |
especially in the home.
link |
So you see humans as helping the robot understand
link |
or give deeper meaning to the spatial information.
link |
Right. It's a partnership.
link |
The robot is supposed to operate according to descriptors
link |
that you would use to describe your own home.
link |
The robot is supposed to, in lieu of better direction,
link |
kind of go about its routine,
link |
which ought to be basically right,
link |
and lead to a home maintained in a way
link |
that it's learned you like,
link |
but also be perpetually ready to take direction
link |
that would activate a different set of behaviors
link |
or actions to meet a current need
link |
to the extent it could actually perform that task.
link |
So I got to ask you, I think this is a fundamental
link |
and a fascinating question,
link |
because iRobot has been a successful company
link |
and a rare successful robotics company.
link |
So Anki, Jibo, Mayfield Robotics with their robot curry,
link |
SciFi Works, Rethink Robotics, these are robotics companies
link |
that were founded and run by brilliant people.
link |
But all, very unfortunately, at least for us roboticists,
link |
all went out of business recently.
link |
So why do you think they didn't last longer?
link |
Why do you think it is so hard to keep a robotics company alive?
link |
You know, I say this only partially in jest
link |
that back in the day before Roomba,
link |
you know, I was a high tech entrepreneur building robots.
link |
But it wasn't until I became a vacuum cleaner salesman
link |
that we had any success.
link |
So, I mean, the point is technology alone
link |
doesn't equal a successful business.
link |
We need to go and find the compelling need
link |
where the robot that we're creating
link |
can deliver clearly more value to the end user
link |
And this is not a marginal thing
link |
where you're looking at the scale and you're like,
link |
Maybe we can hold our breath and make it work.
link |
It's clearly more value than the cost of the robot
link |
to bring, you know, in the store.
link |
And I think that the challenge has been finding
link |
those businesses where that's true
link |
in a sustainable fashion.
link |
You know, when you get into entertainment style things,
link |
you could be the cat's meow one year,
link |
but 85% of toys, regardless of their merit,
link |
fail to make it to their second season.
link |
It's just super hard to do so.
link |
And so that's just a tough business.
link |
And there has been a lot of experimentation
link |
around what is the right type of social companion,
link |
what is the right robot in the home
link |
that is doing something other than tasks people do every week
link |
that they'd rather not do.
link |
And I'm not sure we've got it all figured out right.
link |
And so that you get brilliant roboticists
link |
with super interesting robots
link |
that ultimately don't quite have that magical user experience
link |
and thus that value benefit equation remains ambiguous.
link |
So you as somebody who dreams of robots changing the world,
link |
what's your estimate?
link |
How big is the space of applications
link |
that fit the criteria that you just described
link |
where you can really demonstrate an obvious significant value
link |
over the alternative non robotic solution?
link |
Well, I think that we're just about none of the way
link |
to achieving the potential of robotics at home.
link |
But we have to do it in a really eyes wide open,
link |
And so another way to put that is the potential is infinite
link |
because we did take a few steps,
link |
but you're saying those steps are just very initial steps.
link |
So the Roomba is a hugely successful product,
link |
but you're saying that's just the very, very beginning.
link |
That's just the very, very beginning.
link |
It's the foot in the door.
link |
And I think I was lucky that in the early days of robotics,
link |
people would ask me, when are you going to clean my floor?
link |
It was something that I grew up saying,
link |
I got all these really good ideas,
link |
but everyone seems to want their floor clean.
link |
And so maybe we should do that.
link |
Yeah, your good ideas.
link |
Earn the right to do the next thing after that.
link |
So the good ideas have to match with the desire of the people
link |
and then the actual cost has to like the business,
link |
the financial aspect has to all match together.
link |
Yeah, during our partnership back a number of years ago
link |
with Johnson Wax, they would explain to me
link |
that they would go into homes and just watch how people lived
link |
and try to figure out what were they doing
link |
that they really didn't really like to do,
link |
but they had to do it frequently enough
link |
that it was top of mind and understood as a burden.
link |
Hey, let's make a product or come up with a solution
link |
to make that pain point less challenging.
link |
And sometimes we do certain burdens so often as a society
link |
that we actually don't even realize,
link |
like it's actually hard to see that that burden
link |
is something that could be removed.
link |
So it does require just going into the home and staring at,
link |
wait, how do I actually live life?
link |
What are the pain points?
link |
Yeah, and getting those insights is a lot harder
link |
than it would seem it should be in retrospect.
link |
So how hard on that point?
link |
I mean, one of the big challenges of robotics
link |
is driving the cost down to something
link |
that consumers, people would afford.
link |
So people would be less likely to buy a Roomba
link |
if it cost $500,000, which is probably
link |
sort of what a Roomba would cost several decades ago.
link |
So how do you drive, which I mentioned is very difficult,
link |
how do you drive the cost of a Roomba or a robot down
link |
such that people would want to buy it?
link |
When I started building robots, the cost of the robot
link |
had a lot to do with the amount of time it took to build it.
link |
And so that we build our robots out of aluminum,
link |
I would go spend my time in the machine shop
link |
on the milling machine, cutting out the parts and so forth.
link |
And then when we got into the toy industry,
link |
I realized that if we were building at scale,
link |
I could determine the cost of the Roomba
link |
instead of adding up all the hours to mill out the parts,
link |
but by weighing it.
link |
And that's liberating.
link |
You can say, wow, the world has just
link |
changed as I think about construction
link |
in a different way.
link |
The 3D CAD tools that are available to us today,
link |
the operating at scale where I can do tooling and injection
link |
mold, an arbitrarily complicated part,
link |
and the cost is going to be basically
link |
the weight of the plastic in that part,
link |
is incredibly exciting and liberating
link |
and opens up all sorts of opportunities.
link |
And for the sensing part of it, where we are today is instead
link |
of trying to build skin, which is really hard.
link |
For a long time, I spent creating strategies and ideas
link |
around how could we duplicate the skin on the human body
link |
because it's such an amazing sensor.
link |
Instead of going down that path, why don't we focus on vision?
link |
And how many of the problems that
link |
face a robot trying to do real work
link |
could be solved with a cheap camera and a big ass computer?
link |
Moore's law continues to work.
link |
The cell phone industry, the mobile industry
link |
is giving us better and better tools that can run
link |
on these embedded computers.
link |
And I think we passed an important moment maybe
link |
two years ago where you could put machine vision
link |
capable processors on robots at consumer price points.
link |
And I was waiting for it to happen.
link |
We avoided putting lasers on our robots to do navigation
link |
and instead spent years researching
link |
how to do vision based navigation
link |
because you could just see where these technology
link |
trends were going.
link |
And between injection molded plastic and a camera
link |
with a computer capable of running machine learning
link |
and visual object recognition, I could
link |
build an incredibly affordable, incredibly capable robot.
link |
And that's going to be the future.
link |
So on that point with a small tangent,
link |
but I think an important one, another industry in which I
link |
would say the only other industry in which there
link |
is automation actually touching people's lives today
link |
is autonomous vehicles.
link |
What the vision you just described
link |
of using computer vision and using cheap camera sensors,
link |
there's a debate on that of LIDAR versus computer vision.
link |
And the Elon Musk famously said that LIDAR
link |
is a crutch that really in the long term,
link |
camera only is the right solution, which echoes some
link |
of the ideas you're expressing.
link |
Of course, the domain in terms of its safety criticality
link |
But what do you think about that approach
link |
in the autonomous vehicle space?
link |
And in general, do you see a connection
link |
between the incredible real world challenges
link |
you have to solve in the home with Roomba?
link |
And I saw a demonstration of some of them, corner cases
link |
literally, and autonomous vehicles.
link |
So there's absolutely a tremendous overlap
link |
between both the problems a robot vacuum
link |
and an autonomous vehicle are trying to solve
link |
and the tools and the types of sensors
link |
that are being applied in the pursuit of the solutions.
link |
In my world, my environment is actually
link |
much harder than the environment an automobile travels.
link |
We don't have roads.
link |
We have a near infinite number of patterns and colors
link |
and surface textures on the floor.
link |
Especially from a visual perspective.
link |
So the way the world looks is an infinitely variable.
link |
On the other hand, safety is way easier on the inside.
link |
My robots, they're not very heavy.
link |
They're not very fast.
link |
If they bump into your foot, you think it's funny.
link |
And autonomous vehicles kind of have the inverse problem.
link |
And so that for me saying vision is the future,
link |
I can say that without reservation.
link |
For autonomous vehicles, I think I
link |
believe what Elon's saying about the future
link |
is ultimately going to be vision.
link |
Maybe if we put a cheap lighter on there as a backup sensor,
link |
it might not be the worst idea in the world.
link |
So the stakes are much higher.
link |
The stakes are much higher.
link |
You have to be much more careful thinking through how far away
link |
But I think that the primary environmental understanding
link |
sensor is going to be a visual system.
link |
So on that point, well, let me ask,
link |
do you hope there's an iRobot robot in every home
link |
in the world one day?
link |
I expect there to be at least one iRobot robot in every home.
link |
We've sold 25 million robots.
link |
So we're in about 10% of US homes, which is a great start.
link |
But I think that when we think about the numbers of things
link |
that robots can do, today I can vacuum your floor,
link |
mop your floor, cut your lawn, or soon
link |
we'll be able to cut your lawn.
link |
But there are more things that we could do in the home.
link |
And I hope that we continue using the techniques I described
link |
around exploiting computer vision and low cost
link |
manufacturing that we'll be able to create these solutions
link |
at affordable price points.
link |
So let me ask on that point of a robot in every home,
link |
that's my dream as well.
link |
I'd love to see that.
link |
I think the possibilities there are indeed
link |
infinite positive possibilities.
link |
But in our current culture, no thanks to science fiction
link |
and so on, there's a serious kind of hesitation, anxiety,
link |
concern about robots, and also a concern about privacy.
link |
And it's a fascinating question to me
link |
why that concern is amongst a certain group of people
link |
is as intense as it is.
link |
So you have to think about it because it's a serious concern.
link |
But I wonder how you address it best.
link |
So from a perspective of vision sensors,
link |
so robots that move about the home and sense the world,
link |
how do you alleviate people's privacy concerns?
link |
How do you make sure that they can
link |
trust iRobot and the robots that they share their home with?
link |
I think that's a great question.
link |
And we've really leaned way forward on this
link |
because given our vision as to the role the company intends
link |
to play in the home, really for us,
link |
make or break is can our approach
link |
be trusted to protecting the data
link |
and the privacy of the people who have our robots?
link |
And so we've gone out publicly with a privacy
link |
manifesto stating we'll never sell your data.
link |
We've adopted GDPR not just where GDPR is required,
link |
We have ensured that images don't leave the robot.
link |
So processing data from the visual sensors
link |
happens locally on the robot.
link |
And only semantic knowledge of the home with the consumer's
link |
consent is sent up.
link |
We show you what we know and are trying
link |
to go use data as an enabler for the performance of the robots
link |
with the informed consent and understanding of the people who
link |
We take it very seriously.
link |
And ultimately, we think that by showing a customer that
link |
if you let us build a semantic map of your home
link |
and know where the rooms are, well, then
link |
you can say clean the kitchen.
link |
If you don't want the robot to do that, don't make the map.
link |
It'll do its best job cleaning your home.
link |
But it won't be able to do that.
link |
And if you ever want us to forget that we know that it's
link |
your kitchen, you can have confidence
link |
that we will do that for you.
link |
So we're trying to go and be a data 2.0 perspective company
link |
where we treat the data that the robots have
link |
of the consumer's home as if it were the consumer's data
link |
and that they have rights to it.
link |
So we think by being the good guys on this front,
link |
we can build the trust and thus be entrusted
link |
to enable robots to do more things that are thoughtful.
link |
You think people's worries will diminish over time?
link |
As a society, broadly speaking, do you
link |
think you can win over trust not just for the company,
link |
but just the comfort that people have with AI in their home
link |
enriching their lives in some way?
link |
I think we're in an interesting place today
link |
where it's less about winning them over
link |
and more about finding a way to talk about privacy in a way
link |
that more people can understand.
link |
I would tell you that today, when there's a privacy breach,
link |
people get very upset and then go to the store
link |
and buy the cheapest thing, paying no attention
link |
to whether or not the products that they're buying
link |
honor privacy standards or not.
link |
In fact, if I put on the package of my Roomba,
link |
the privacy commitments that we have,
link |
I would sell less than I would if I did nothing at all.
link |
And that needs to change.
link |
So it's not a question about earning trust.
link |
I think that's necessary but not sufficient.
link |
We need to figure out how to have
link |
a comfortable set of what is the grade A meat
link |
standard applied to privacy that customers can trust
link |
and understand and then use in their buying decisions.
link |
That will reward companies for good behavior
link |
and that will ultimately be how this moves forward.
link |
And maybe be part of the conversation
link |
between regular people about what it means,
link |
what privacy means.
link |
If you have some standards, you can say,
link |
you can start talking about who's following them,
link |
who does not have more.
link |
Because most people are actually quite clueless
link |
about all aspects of artificial intelligence,
link |
the data collection, and so on.
link |
It would be nice to change that for people
link |
to understand the good that AI can do.
link |
And it's not some system that's trying to steal
link |
all the most sensitive data.
link |
Do you think, do you dream of a Roomba
link |
with human level intelligence one day?
link |
So you've mentioned a very successful localization
link |
and mapping of the environment, being
link |
able to do some basic communication to say,
link |
go clean the kitchen.
link |
Do you see in your maybe more bored moments,
link |
once you get the beer, to sit back with that beer
link |
and have a chat on a Friday night with a Roomba
link |
about how your day went?
link |
So to your latter question, absolutely.
link |
To your former question as to whether a Roomba
link |
can have human level intelligence, not in my lifetime.
link |
I think you can have a great conversation,
link |
a meaningful conversation with a Roomba
link |
without it having anything that resembles
link |
human level intelligence.
link |
And I think that as long as you realize that conversation
link |
is not about the robot and making the robot feel good.
link |
That conversation is about you learning interesting things
link |
that make you feel like the conversation that you
link |
had with the robot is a pretty awesome way
link |
of learning something.
link |
And it could be about what kind of day your pet had.
link |
It could be about how can I make my home more energy efficient.
link |
It could be about if I'm thinking about climbing
link |
Mount Everest, what should I know?
link |
And that's a very doable thing.
link |
But if I think that that conversation
link |
I'm going to have with the robot is
link |
I'm going to be rewarded by making the robot happy,
link |
well, I could just put a button on the robot
link |
that you could push and the robot would smile.
link |
And that sort of thing.
link |
So I think you need to think about the question
link |
And robots can be awesomely effective at helping people
link |
feel less isolated, learn more about the home
link |
that they live in, and fill some of those lonely gaps
link |
that we wish we were engaged learning
link |
cool stuff about our world.
link |
If you could hang out for a day with a robot
link |
from science fiction, movies, books,
link |
and safely pick its brain for that day, who would you pick?
link |
I think that A, data is really smart.
link |
Data has been through a lot trying
link |
to go and save the galaxy.
link |
And I'm really interested actually in emotion
link |
And I think you'd have a lot to say about that.
link |
Because I believe actually that emotion
link |
plays an incredibly useful role in doing reasonable things
link |
in situations where we have imperfect understanding of
link |
In social situations when there's imperfect information.
link |
In social situations, also in competitive or dangerous
link |
situations that we have emotion for a reason.
link |
And so that ultimately, my theory
link |
is that as robots get smarter and smarter,
link |
they're actually going to get more emotional.
link |
Because you can't actually survive on pure logic.
link |
Because only a very tiny fraction of the situations
link |
we find ourselves in can be resolved reasonably with logic.
link |
And so I think Data would have a lot to say about that.
link |
And so I could find out whether he agrees.
link |
If you could ask Data one question,
link |
you would get a deep, honest answer to what would you ask.
link |
What's Captain Picard really like?
link |
OK, I think that's the perfect way to end it.
link |
Colin, thank you so much for talking today.
link |
I really appreciate it.