Evan Wineland: Deploying is the Point. Affordable Robots that Work | Turn the Lens Ep51

Evan Wineland: Deploying is the Point. Affordable Robots that Work | Turn the Lens with Jeff Frick Ep51 

English Transcript 

© Copyright 2026 Menlo Creek Media, LLC, All Rights Reserved

Interview Introduction  - Sea Breeze Laundromat, San Francisco

Jeff Frick:
Hey, welcome back everybody. Jeff Frick here coming to you from Sea Breeze Laundromat on Castro Street in San Francisco. And we're really excited to release the next of the interviews that we did at Humanoids Summit. This one's with Evan Wineland and his company is called Weave Robotics. And this is one of the Weave Robotics robots right here. And it's working. And I'm not at their office. I'm not at a lab. I'm not at some fancy ERC or EBC. I'm actually here at the laundromat on Castro. And when large customers come in with kind of Door Dash situations where they send in a whole load of laundry and they run it through the washing machine and then they put the big pile here next to the robot and away it goes. So it's a couple interesting notes in this interview with Evan. I think the most important thing is his concept and their mission of 'Delivering is the point.' Get the thing out in the world. And I think they incorporated like late 2024. They had a POC by the middle of the year. And I think by late 2025, a little over a year, they were shipping product. So really interesting story, great conversation. Really excited to release this in collaboration with Humanoids Summit and ALM Ventures. Modar, Jesica, thanks a lot. I think you'll enjoy this one. Thanks for watching. Thanks for listening in. Take care. Bye bye.

Main Interview - Humanoids Summit SV 2025

Jeff Frick:
Hey welcome back everybody. Jeff Frick here coming to you from the Computer History Museum. We're back for Humanoids Summit. We did a summer show in London. They just announced they're going to be in Tokyo in summer 2026. So the momentum around the space is really crazy. I thought it was crazy a year ago, but it's amazing how far things come. One of the things that is really striking is how young a lot of the companies are that were presenting this year and our next guest they just started in 2024 which really wasn't that long ago. So we're excited to welcome in. He's Evan Wineland, the Co-Founder of Weave Robotics. Evan, great to see you.

Evan Wineland:
Great to see you Jeff.

Jeff Frick:
Absolutely. So I caught a little bit of your presentation earlier, but I think I caught the most important part, which was your mission, which is to be practical and capable. You don't necessarily hear those words all the time in the context of super advanced humanoid robotics. How did that become your founding principles there?

Evan Wineland:
It became our founding principle because my co-founder and I saw that nobody was actually building a robot that we would want to put in our own home that was actually going to ship anytime soon. And we realized that part of the problem was that people were not building things that could ship as soon as possible. We saw something that we could bring opinion to. We both come from Apple, he from a robotics R&D and me from on device intelligence and Apple intelligence. So we realized that deploying as quickly as possible is the point. It is the strategy. It is the value.

Jeff Frick:
I love that deploying as quick as possible is the point. And you in fact showed I did see you've got a robot working in public. Is that a demo robot? Is that working? Is it paid or is it...

Evan Wineland:
That's a paid deployment. It's a paid deployment.

Jeff Frick:
It's a paid deployment.

Evan Wineland:
Yeah, we have multiple across San Francisco now.

Jeff Frick:
So tell the folks that missed the video what was going on in that deployment.

Evan Wineland:
Totally. Yeah, so Weave is making robots that work for homes and businesses. And we're deployed in multiple laundromats across the city. We're shipping to more in the coming months. We're also talking to folks in hospitality as well as manufacturing, because we've been able to through well chosen vertical integration by building the product ourselves in-house, bring our costs down so much that our robots work for big businesses and mom and pops alike.

Jeff Frick:
And you talk a lot about vertical integration in your talk and the emphasis seemed to be in terms of, how you could really drive costs down. Seems to be a little bit counterintuitive to kind of a this modular approach and picking and pulling from different parts. So how did you decide to go. 

Evan Wineland:

Yeah, I didn't even mean that pun on picking. No.

Jeff Frick:
But how did you decide to go with kind of a vertical stack? And where were the big pieces that you get these step function type improvements over just buying stuff off the street?

Evan Wineland:
It's an incisive question because I actually think that pursuing vertical integration for its own sake is actually to be a bit misguided. Vertical integration should always be in service of either making the best thing or actually optimizing in the short or the long term for iteration speed. One of the things that vertical integration gets you is control over more and more of the product. So it means that you don't have to accept somebody else's T's and C's in order to make an important change. It means that you don't have to accept their privacy policy. This means that we're able to control that much more of the end user experience, and we only do it as much as is necessary to make the thing right.

Jeff Frick:
That's great. So I mentioned you got started in early 2024. When did you ship the robot? Give us a little bit of the I mean that is a crazy fast from starting the company to having something out in the field. Give us a little bit of the timeline and how you were able to accelerate that so fast.

Evan Wineland:
Absolutely. So we raised funding last fall. We got into our office in late October, November. We hired most of the team of experts that we wanted to have multiple PhDs, roboticists.

Jeff Frick:
And how do you find that many? How do you find your team that fast?

Evan Wineland: Yeah.

Jeff Frick:
Not easy to do in this environment.

Evan Wineland:
It is not easy to do. And we're very, very fortunate to have the team of experts that we have. I will actually say hiring was not as hard for us as it has been for some other companies. A lot of people were genuinely excited to work on something as audacious as what we were trying to build, and indeed have built and shipped. So that actually made it more straightforward. People want to work on hard things.

Jeff Frick:
‍Yeah. 

‍Evan Wineland:
‍And so back to the previous question about our timeline. So we actually got our first fully integrated design iteration that gave line of sight to the thing we want to ship this spring. And then we actually shipped the first version 

Jeff Frick:
A year or so? 

Evan Wineland:
This spring. Yeah, so like actually

Jeff Frick:
Spring 2025 right. 

Evan Wineland:
Summer 2024 was when we actually like incorporated.

Jeff Frick:
Okay. 

Evan Wineland
Yeah.

Jeff Frick:
Wow, so like a year.

Evan Wineland
Yeah. And then we actually shipped to businesses, for the first time in September of this year.

Jeff Frick:
And did they have extended POCs or did, were you able to work it out to just get it in and get it working?

Evan Wineland:
This is indefinite. Yeah, we're keeping it going, and we're actually expanding with those partners to more laundromats, too.

Jeff Frick: Okay. 

Evan Wineland:
Yeah. 

Jeff Frick:
So life is, has more variety than we could think of. What are some of the interesting things that have happened from some of those deployments that you that really validate your strategy to deploy first and then iterate based on the feedback. What are some of the funny things that happened?

Evan Wineland:
Yeah. Well, so let's see on the more prosaic but like technical product side, we are operating these things across the city and we just encounter the kinds of things that we were in a sense looking for by actually deploying out into the real world. How well does this thing work when it's deployed? Not in our office but miles away? What are the components that most often need to be replaced and how can we improve or just entirely remove them? So I would say that's on the sort of prosaic side. But then on the fun side, people are delighted by watching this thing. One of our deployments is at a laundromat in Noe Valley, owned by this lovely couple, Justin and Kay. And people congregate outside of this laundromat that has this totally picturesque decaling that says 'Coin Operated' on the outside. There was one day where I was filming something for our website, and there were passers by and one of the people who began to help form a little crowd was this little girl who got up on a chair and was cheering on our robot. She was so excited to watch the thing do its job and in an environment where people are wondering what the future is going to look like, it's actually quite clear that for well-chosen problems with people building for the right kinds of things the right kinds of tasks there is excitement about a more abundant future. And that's what we're trying to do by going after the most mundane things that people don't necessarily want to have to do like fold their laundry. 

Jeff Frick:
So did, do people let the machine fold their laundry so it's done with the dryer and they just take it over to the robot and that's part of the service available there.

Evan Wineland:
So here's how this works. We partner with a on demand laundry service that essentially works like DoorDash for laundry. They're called tumble laundry. And anybody can go to their website today and then pay to have somebody come and pick up a bag of their clothes, take it to a laundromat. 

Jeff Frick:
Old school laundry service. 

Evan Wineland:
And then take it to one of our partner laundromats. And then when it comes time to fold, sometimes like 40 pounds of laundry for multiple weeks for an entire family, our robot gets to work, and then we hand it off to a DoorDasher or who sends it back to your front desk to your front door.

Jeff Frick:
Okay. That's great. 

Evan Wineland:
Yeah. 

Jeff Frick:
Safety. What are some of the safety challenges? What were some of the customer safety concerns that maybe were legit or not legit based on you know, exposure to this new technology? And then from an experience point of view, what were your kind of top, kiind of  primary safety concerns that you had to account for?

Evan Wineland:
Totally. Safety is one of those things that you have to build for from the beginning and you have to build for it at every single part of the product. And that literally means hardware to software. You have to control everything. And this is one of the reasons why it's actually really important to be vertically integrated. We write our own software from what you could call the OS to our controllers, to our drivers that actually move the motorized joints of the robot. So we took precautions there. We took precautions in the components that we selected. The actual motors and their properties is extremely important. We picked a wheeled base because it passively balances. We really tried to think about this from every single angle, and then going back to the overarching strategy here the ultimate thing to do to make sure that your priors are right about how safe your product is is to put it out in the real world in a well chosen, careful, guarded way to make sure that those hypotheses are tested and confirmed. So that's why deploying is the point.

Jeff Frick:
I love that. So what are some of the what are the, hallucinations? How do the hallucinations manifest themselves before everything is super buttoned down.

Evan Wineland:
Yeah, hallucinations are a little bit less of a problem for our specific application. So we have what is called a vision language action model. [VLA] That is the model that powers the autonomous movements of the robot. And hallucinations are more a property of models where the output is text. In our case, what we ultimately care about is how well is it actually performing these motions?

Jeff Frick:
Would there be a case or does that not qualify as a hallucination where it either just goes to where it's not supposed to go at all, or versus just, you know, fine tuning the actual like

Evan Wineland:
I suppose you could call that a hallucination of a kind. We ultimately just think about it as a performance issue. We want to make sure that we are constantly folding these things really precisely because we had our origins as a consumer product. We actually care about making sure that things are aligned that they're as crisp as possible and then once we have that foundation of what I would call accuracy we would also care about how quickly we're doing it too.

Jeff Frick:
Right, right. So what's the thing when you talk to your Aunts. The holidays' coming up who are like, oh my gosh, I what is going on with these robots? Are people scared? Are they excited, what you know, what do you find out there. And then what do you tell them as to give them some comfort as to what's going to happen in the very short future?

Evan Wineland:
Yeah. Well, I think that, you know, some of this apprehension about what the future is going to hold is justified. There is lots that is changing very rapidly about the world. I think that this is ultimately one of the bigger responsibilities of the builders themselves. What problems are you choosing to solve? Are they the most mundane ones? 

There's this really lovely quote by a woman named Joanna Maciejewska I might be butchering the last name, but what she said resonated. She said I want AI to fold my clothes and do my dishes so that I can focus on my art and writing, not do my writing. Make my art for me, right? I actually did get a text from my Aunt Julie the other day who is a teacher in Maryland, and she said that she had told her class of students about the robot that we are building and that they wanted to tell they wanted to buy it for their families so that they could bring it back  and then alleviate some of the work that people have to do. I think that this boils down to are you choosing the right things to help, alleviate mundanity for people? And then you have to choose those problems well.

Jeff Frick:
Yeah. Great. Well Evan, a super, super good story. I'm going to take your 'deployment is the whole purpose' and use a lot. I promise I'll give you a proper credit.

Evan Wineland:
Heck yeah. It's been a lovely talking, Jeff.

Jeff Frick:
All right. He's Evan, I'm Jeff we're at the Humanoids Summit in Mountain View, California at the Computer History Museum. Thanks for watching. We'll catch you next time.

Bonus Interview - Sea Breeze Laundromat, San Francisco

Jeff Frick:
Hey. Welcome back everybody. Jeff Frick here. Coming to you from San Francisco at the Sea Breeze laundromat on Castro Street. I came up here to do a short introduction video to follow up on some of the interviews from last month at Humanoids Summit. And who do I run into but the founder of Weave Robotics he's Evan Wineland. came by to see so he could show off this amazing machine. Evan, great to see you again.

Evan Wineland:
Great to see you again, Jeff.

Jeff Frick:
So we had a great conversation in December. And the thing that really struck me was your whole conversation about 'Delivering Is the Point'. We're here at one of your customers, you got a commercial deployment in less than 36 months, whatever it was, in a slightly over a year. What were some of the real details about trying to get something actually in a real customer not just talking about commercial deployments or POCs?

Evan Wineland:
I think that one of the first things that we had to do is we had to focus on how we were going to deliver value for people. We had to pick a task. One of the first things that we wanted to do when we bring our home robots to customers this year is to fold their laundry. This is one of the most ubiquitous tasks in the home but it's one of the most loathed. And we thought how can we tackle this and then earn the right to be in people's homes as well? And we realized there's actually a product here that can start serving businesses today. And we realized that one of the best ways to validate that we'll be able to do this task in other environments is to just start doing it as quickly as possible. These robots here are of a slightly different form factor, but they directly borrow from the design of our home robot. We were able to put it out and do the exact task we're going to be doing at homes and short circuit the time to getting validation in the real world. The reason we were able to achieve this was focus, plain and simple.

Jeff Frick:
Right, right. It seems a little counterintuitive because the other point of emphasis that you made was vertical integration and doing as much vertical integration as you can so that you can really control the customer experience and you're not with beholden to some other person or some other policy. But speed would beg that you're going to use more off the shelf components and use as much kind of things that are out there or open source or, or, regular components as possible. How do you how did you go fast and also have vertical integration?

Evan Wineland:
That's a fantastic point. And I think that a lot of people do get confused about why to vertically integrate. It's not for its own sake. It's not for reasons of purity. You vertically integrate to build the best product and to learn and iterate as quickly as possible. You do pay an upfront cost. It would be in a sense, faster if there was the perfect off the shelf component which is why not all parts of our design we have built internally. But most of it we have. And we did that because we wanted to have control and the ability to iterate meaningfully from when we first deployed this robot to even what you see today. That's the payoff. So it allows you, after you pay that upfront cost to iterate more quickly, which is essential to being a successful robotics company deploying in the real world.

Jeff Frick:
Even then, how did you accelerate the beginning of that thing? Because that would beg the question that you'd have a lot more long term advanced development for all these all these proprietary parts. Again, you were fast, so you just the component pieces are there for you to build the solutions? Again, you're developing a lot of stuff from scratch in a really short period of time.

Evan Wineland:
Yeah. What I'll say about that is that central to how we're approaching general purpose robots that Weave is distilling these form factors into their simplest possible. We thought, how can we make a general purpose manipulator but truly remove anything un-essential to doing the job. That comes in, in how we designed our arms. That comes in, in the gripper that we chose and designed in-house. The parts that we removed from the base of these more stationary workhorse units that we've deployed in laundromats. It's removing anything that's not essential wherever it's not necessary.

Jeff Frick:
Right. Okay. Well, let's talk about safety. This thing is not behind any type of guard or shield. There's a couple kids have walked in since we've been here. How safe is it? How did you deal with safety? How does the customer here at Sea Breeze worry about safety? And, you know, their business insurance to have this thing in here?

Evan Wineland:
Absolutely. So we care about design like safety in all aspects. This is something you have to think about in your design. Even like your motor selection is where you have to be thinking about this at this level. We care about it in the software level. So when people get in the workspace of the robot, which we of course discourage we pause so that there aren't concerns. And then plain and simple, one of the best ways to ensure safety is actually to be really, really specific in how you deploy these things. By reducing the number of tasks that we expect of this robot we can ensure a safer workspace around it which allows you to make sure that you're doing this task really well and really safely before you graduate to more complex models of interaction. And that is the reason why we're able to safely operate in here day after day. So I would be remiss too if I didn't say that at some point learning how to make sure that things are safe in the real world comes back to deploying.

Jeff Frick:
Right, right. 

Evan Wineland:
You have to make sure that you actually get out there so that you can see the things that you couldn't have predicted in advance as well.

Jeff Frick:
Right. So you've been in deployment for a while. You guys are moving very quickly. What are the priorities for 2026? Are you just kind of doubling down on this path, are there some tangential places that you want to go. What are the priorities in the near term?

Evan Wineland:
Yeah. We have a recipe that is already scaling for business deployments. So that'll be one of the first things that we're pushing on. And then we're also going to be shipping to home customers as well in 2026.

Jeff Frick:
‍Wow wow. And again affordable enough I won't put your feet in the fire for pricing, but affordable enough to be here at Sea Breeze Laundry and again, it's a paid deployment, right? This is a this thing is working for a living.

Evan Wineland:
It is. Yeah. It has to earn its keep. And that's one of the reasons why we care that the task really mattered to people. When we chose laundry as a first beachhead task for ourselves.

Jeff Frick:
Yeah. Very cool Evan. Well, thanks for inviting me up. It's fun to come up and actually, get out of the lab and see it in action.

Evan Wineland:
Yeah. Great to see you, Jeff.

Jeff Frick:
All right. He's Evan. I'm Jeff. You're watching 'Turn the Lens' with Jeff Frick. See you next time. Thanks for watching. Thanks for listening on the podcast. See you next time. Take care. Bye bye.

COLD CLOSE - OFF MIC

Jeff Frick:
Cool

Evan Wineland:
Awesome. Super Fun

—---------–

Evan Wineland: Deploying is the Point. Affordable Robots that Work | Turn the Lens with Jeff Frick Ep51 

English Transcript 

© Copyright 2026 Menlo Creek Media, LLC, All Rights Reserved