Dave Limp: "The consumer electronics industry has rested on its laurels"

• Amazon Companies has increased its Spanish workforce by more than 60% during the pandemic

If Alexa, Amazon's virtual assistant, had to find a father for her,

would be a good candidate.

For 12 years, he has been responsible for the devices and services division of Amazon, an arm of the e-commerce giant that began operations with a focus on the Kindle e-book reader but is now at the forefront of so-called "environmental computing." "thanks to the Echo speakers and displays and the many home automation devices the company has launched over the last decade.

I'm going to start by asking what exactly you do within Amazon, because Devices and Services is a very broad description and I imagine that not all Amazon services are under your supervision, right?

I mean you are not the person I have to call if I have a problem with Amazon Video.No.

The group is called Devices and Services.

And I think that's an important distinction because we think of consumer electronics, which is most of what my organization does, not just as a product, but as a device that's deeply coupled with a service.

It comes to us by inheritance, from the Kindle.

It was a reading device but the most important thing is that it had a library associated with it and it was always connected.

We are responsible for Kindle, Fire tablets, Fire TV, Echo and Alexa,

Ring and the Ring security service, Blink and their cameras, the Eero routers... and then we have a couple of big bets that also fall within the organization like Kuiper, our network of low earth orbit satellites, or Zoox, which is our autonomous taxi initiative. This relationship is one that I have never fully understood because Amazon has its own products like Echo or Fire but also products like Eero, Blink, Ring, which are clearly independent.

How do you control all these companies that somehow maintain their identity despite being acquired by Amazon?

Why don't they become just another product, like an Echo speaker? I think the first way to manage anything with this kind of complexity and breadth is to think about teams and leaders.

Each of the products you just named have a person who does not have the title of CEO, but who for all practical purposes operates as the CEO of a company.

We have common goals, but as long as they meet them, they have great autonomy and can set their own roadmap and develop products with total freedom.

It is a type of organization that fits well with the philosophy of Amazon but is different from that of other companies.

For us, what is really important is to make sure that these products communicate with each other in a way that is easy for the customer.

And that's where I can help a little more.

What we're trying to do is figure out what's going to be the common thread between them.

If I have an Echo Show screen in my kitchen, when someone rings the Ring video doorbell,

Wouldn't it be wonderful if it appeared on the Echo Show screen?

What we do is create APIs for those kinds of things that are great for our internal teams.

We also often outsource those APIs so that someone else's doorbell other than Ring can also have the same functionality. Yes, but then what is the advantage that Ring has in terms of being an Amazon company?

Because it seems that they operate in the same way as any other. Yes, they are independent but we try to offer the best experience.

On a Samsung TV can you ask Alexa for things?

Yes, but Fire TV should be the best implementation of Alexa thanks to the Fire TV team.

And we think it is, voice search is the best on Fire TV.

And it's not because we're not outsourcing the right Alexa APIs. So,

when it comes time to make a new product, what is the deciding factor that indicates who makes it?

For example, the flying camera drone that you introduced last year.

That could just as easily have been an Amazon product instead of a Ring product.

How do you divide these things? For something that is evolutionary, it's usually easy.

Whoever did it before is in charge.

But it's also a question of who has the passion for it and who has the resources at the time.

For what is completely new we value several things.

The camera drone is focused on home security, and therefore falls under Ring's domain. But you also have a new robot, Astro, that has a camera and moves around the house, and that is from Amazon, not from Ring... Yes, there we did a deep integration with Ring,

but it was also very focused on Alexa.

So it made sense for that to come out of the Alexa organization because, first and foremost, it's Alexa on wheels.

In general, the most difficult are these cases, when it is a completely new product.

There you usually hire someone new who has experience.

Let's take Kuiper as an example.

We had no experience within Amazon on space issues.

We have never put anything into orbit.

So we went out and hired a lead for that task specifically. Well, in this case you have Jeff Bezos, who does have some space experience. Yeah, that helps (laughs).

I'm also very passionate about space, but that doesn't mean I'm going to get something into orbit, like Jeff. I want to ask a question about the current economic environment and then we'll get back to products.

How has the semiconductor crisis affected the Amazon split? It's hard to answer that question at a macro level because it's so much more complex than it first appears.

Basically, we are in a supply chain crisis, and the supply chain has many different dimensions.

Maritime transport is a problem because there are fewer ships, there are days when a factory can close due to an infectious outbreak of Covid, one week you may have supply problems with a component and the next you can get it without a problem.

Suddenly, an entire city or part of it is closed, as happened with Shanghai.

Imagine that you are trying to transport something from one end of Shanghai to the other.

Those two specific areas may not be closed, but the roads between them may be closed.

So it's not just a semiconductor problem.

You have to be aware of all these things. It's a constant headache, I imagine, but do you wake up every Monday thinking "my god, what's going to go wrong this week"?

Or after two years of Covid is it already something routine? The first, without a doubt.

I've spent more time studying the supply chain in the last two years than in the first 28 years of my career combined. , Amazon's attention has been diverted to these experimental ideas.

What is the objective? Is it pure experimentation or is there a commercial vision behind them? No, we are not, as they say, throwing a plate of pasta at the wall and seeing what sticks.

The way we work at Amazon is that we come up with ideas and write about them as if we were announcing them publicly tomorrow.

It is a document that has a press release and some FAQ pages.

We debated it vigorously and rigorously.

Many of those products never see the light of day.

But once we green light a project, we have conviction.

We believe that it will be a product that customers will like.

It is going to be a business that, if it works, if it scales in any dimension, it will be a good business for Amazon.

We are a for-profit company.

That said, I think we take more risks than others, and I think that's a good thing.

I would say that the consumer electronics industry in general is not taking enough chances.

It has fallen asleep a little on its laurels,

But technology is still advancing and it is advancing very fast.

If you think about the revolution that AI has brought about, how sensors have improved... I feel like it's part of our mission not to experiment with customers, but to push technology forward and take risks.

This leads you to great successes.

Who would have thought that millions and millions of people would have video doorbells or smart speakers at home.

But those risks paid off.

Sometimes you also have failures, of course.

But that's fine too, because if everything works 100% every time, then you're just iterating. I was going to ask you about failures because I feel like the Fire Phone fiasco, for example, was instrumental in getting Amazon to focus on speakers and AI.

Perhaps if the telephone had been successful, you would never have explored it, or not with the same passion. Yes,

It is difficult to know what would have happened if we had been successful, but what you can see is that it has led us to not think of the telephone as the center of the technological world.

If you sell hundreds of millions of phones, you want it to be there all the time.

We want to decentralize smart objects throughout the house so that they communicate with each other, it is this idea that we call "ambient computing".

But without a doubt failures teach us a lot.

They tend to inspire you.

You don't just want to get away from them.

You want to learn from them and wonder why it didn't work.

If you're really self-critical, when you look at these failures, I'm sure they'll inspire you for other successes. But the phone is still very central to our lives, and I'm really surprised that you've never tried it again.

I feel like it's something that would fit into your current catalogue.

Have you ever been tempted to re-release one? It's a huge segment of the market.

I think we would build a phone tomorrow if we could come up with an idea that was differentiated.

The thing about phones, though, is that they're all the same.

It is an incredibly well served market.

There are good companies and products, and they have reasonably affordable app stores where we can offer services.

There's no reason to put more effort into a place where customers already have plenty of options. Is Amazon's strategy then to focus exclusively on the home? It started in the home.

That's certainly where we got the initial push and we still have a lot of presence.

But I think ambient computing is not tied to a particular location.

It's a new interface for new use cases.

When you're writing an article, you probably still type it on a keyboard because it's the most efficient way to write a long document.

The phone is good for a lot of different things but so is ambient computing and I think even though it started in the home it will evolve to other places that make sense.

The next logical step is in the car.

Autonomous cars are still a long way off.

We could argue about how far away they are, but it won't happen tomorrow or next year.

If you're driving, you don't want to be distracted, you want to drive, you want to focus on the road.

A voice interface is exactly what you need in a car.

And that's why we have more than a dozen car manufacturers that have integrated Alexa.

I think there are many more possibilities.

We now have an agreement with Disney World and its hotels.

You come back from spending a day in the park with your access bracelet and you can see the photos of your experience on the hotel screen in an almost magical way.

That's a good example of another ambient experience that's really taking the magic of Disney to the next level. The buzzword in the tech world right now is metaverse, and I seem to recall you've been critical of the concept in the past.

Don't you see options there for ambient intelligence? Couldn't Alexa be a virtual assistant in the metaverse? I think voice assistance will have a role in whatever the final version of the metaverse is but it's one of those concepts that means different things for different people.

If you talk to someone from Minecraft,

he has a different view of what the metaverse is than someone in Meta.

Not that I'm opposed per se.

I have nothing against the metaverse concept.

What I don't want is the dystopian version.

I don't want to live in a society where we are all locked in a virtual world and we lose sight of this wonderful thing that we have around us that is planet Earth, with our friends and our families.

I think we should be building technology that allows us to participate in the world we know and love, that makes us pick up our heads from the phone from time to time and listen to music together, watch a movie together, enjoy working out together... that kind of things. I don't have a lot of time left and I wanted to ask you specifically about Alexa and privacy.

Many of Amazon's products have microphones and cameras, and asking someone to put them in your home means first building a relationship of trust.

What is the process when designing a product in terms of privacy, and how do you make Alexa smarter without conflicting with privacy? The philosophy of how we design products hasn't changed in my twelve years at Amazon.

Safety comes first in everything we do.

The first thing we consider in the Astro robot, for example, is that it cannot hurt anyone, neither a pet nor a person.

But just as important as security is privacy.

The only thing we have as a company is the trust of our customers, if we lose it, there is nothing left.

And you can lose it very quickly.

It is lopsided because it takes a long time to earn it.

From the first Echo, for example,

we electrically disconnect the microphones when you press the mute button.

We decided from day one, too, that we weren't going to always be listening, only when the word Alexa is detected.

It was a conscious choice.

We could have sent everything to the cloud to process and be much more accurate in the first generations, but we didn't because we felt it was a privacy issue.

What has happened over the years is that technology, both hardware and software, have evolved so that we can now do so much more without relying on the cloud.

It doesn't mean that the cloud isn't important yet.

It is, but we can do much more without it.

We can now do automatic speech recognition with our built-in inference engine.

We never send the raw voice recording anymore.

And the same in face recognition.

He's charming when you walk into the house and Astro says "Hi Dave" in his little robotic voice.

All of that is done locally.

None of that facial recognition goes to the cloud. But how much smarter can Alexa get?

Are we near the limit of what we can do now with AI?

Because the experience is still very basic: ask a question and get an answer. Yes, I think that's a good description of the Alexa experience, which is now about eight years old.

We're getting a little better now, but seven of those eight years has been a transactional UI where you ask Alexa and Alexa answers.

What has evolved dramatically in that early stage is the depth and breadth of what Alexa can respond to.

When we launched Alexa,

he could do only 13 things and none of them particularly complex.

Now Alexa can do hundreds of things with great breadth.

I can ask much more about any topic.

We had only one language when we started, now we have more than a dozen and we can detect multiple languages ​​at the same time.

All this is what I call the first phase.

And then, since a couple of years ago, we've started what I would call the second phase, which is where Alexa is proactively doing more things.

If you turn off the garden light at night for 30 days and forget on the 31st, Alex will say, "hey, I think you normally turn off that light. Do you want me to do it automatically for you in the future?"

And if you agree automatically create a routine.

That gives you an idea of ​​the type of interaction in the next phase, which is much more proactive.

And then there would be the third phase, where we are already taking the first steps, which is to make the experience more like what we are doing here, a conversation.

I haven't used your name once, but you know when I'm talking to you.

It's not just voice.

We can use gestures too.

That will be the next level. And do you think that will help people discover the usefulness of assistants?

Because I feel like when someone puts an Echo speaker in the house, they basically use it to listen to music. The data tells us that music is a very popular use, but the number of things that people do with Alexa is already very large.

They check the weather or follow sports results, for example.

But I agree that we can constantly improve on feature discovery.

Conforms to The Trust Project criteria

Know more

• amazon

• Amazon Prime Video