Australia markets closed
  • ALL ORDS

    7,381.10
    +19.20 (+0.26%)
     
  • AUD/USD

    0.6977
    -0.0049 (-0.69%)
     
  • ASX 200

    7,127.70
    +22.30 (+0.31%)
     
  • OIL

    86.77
    +0.24 (+0.28%)
     
  • GOLD

    1,789.60
    -0.10 (-0.01%)
     
  • BTC-AUD

    34,248.92
    -388.54 (-1.12%)
     
  • CMC Crypto 200

    570.08
    -1.84 (-0.32%)
     

Robotics and AI are going from cage to stage

·6-min read

A lot of promising companies come out of work by researchers at universities, or even grad students who have struck on some new innovation. But the transition from tech-focused research group to product-focused startup isn't easy to make; fortunately three experts in the matter joined us at TC Sessions: Robotics to discuss a few ways to get through it successfully.

Milo Werner is a new general partner at MIT's The Engine, an accelerator and fund focused on "tough tech." Joyce Sidopoulos is a co-founder of MassRobotics, a community and advocacy group for the sector's startup ecosystem. And Pieter Abbeel is a professor at UC Berkeley and the co-founder of Covariant, which is designing a new generation of warehouse robots (he also just won the ACM Prize — belated congratulations, Pieter).

[youtube https://www.youtube.com/watch?v=PWfC5mhmynE?version=3&rel=1&showsearch=0&showinfo=1&iv_load_policy=1&fs=1&hl=en-US&autohide=2&wmode=transparent&w=640&h=360]

Our panel started out with some of the most obvious technical considerations founders need to keep in mind when shifting from a research to a mass production process. (Quotes have been lightly edited for clarity and continuity.)

"When the technologists are designing the product itself, they just want it to work, right? But when you actually go to manufacturing, the manufacturer will say, we can't put that board on top of that, we can't assemble it that way. So you really, from the beginning, should be thinking about manufacturing, and design for manufacturing," said Sidopoulos.

Werner pointed out that the tolerances, precision and monitoring will never be as good as your own lab, so be ready to accept that — as well as compromises for cost. "The reality is you go through a series of iterations once you enter manufacturing, for cost down, and managing the ramp up," she said.

"In a research lab, you're just trying to get a prototype working. And often you can write a paper that gets a lot of visibility and excitement for showing something for the first time — it can work!" said Abbeel. "Then you go to something in production... and all of a sudden, it's not the fact that you can make it work once that matters — it's the consistency, having it essentially always work. So, chasing the kind of high nines of performance and reliability, I think is probably the biggest difference."

As demands on the company change, the company itself must change to accommodate them. Werner pointed out that many founders, having come off four to eight years of work in the area, have a passion and familiarity with the material that's difficult to match — but that can be a barrier to building a team.

"The part that they often face is that they need to build a big team, and that team needs to be just as strong or stronger than them," she said. "Coming from the academic, institutional space, it can be a very individual contributor ideology. And moving to the business space, it's completely team oriented. And you are only as strong as your team."

"I would totally agree with that," Sidopoulos said. "We find that a lot of time, the co-founders are passionate about what they're doing. But then they have to bring in, you really have to bring in a strong business-focused person, because how do you sell this thing? How do you put it into a pitch deck? That's not what you're taught when you're in engineering school. Finding the right people and the right combination. It's like getting married, right? You're marrying someone who you're going to be spending a lot of time with. And so you have to make sure that you have the same goals, you have the same mission, you have the same work ethic."

Werner pointed out it's always wise to hire ahead of the need — and joining a community of like-minded people can help a founder build their network and get a sense of where others are in the process.

Abbeel told a story about a serendipitous relationship that Covariant built:

Our first lead investor, Amplify, had a partner who they made available five hours a week, who had a ton of operational experience in sales, on the business development side, many previous startups, where he had worked full time, and now he was a venture capitalist," he said. "And he brought all that operational experience into what we were doing. And we quickly found that five hours wasn't [enough]... we wanted more! Over time, we got him so excited to join us as our COO! It's a great example of somebody who most likely would not have been in our natural network, but came our way in a way that actually it was possible to get to know each other very well, for several months before we really started working full time together. With a completely different set of experiences and skill set — so valuable."

If only all founders were so lucky! It's not exactly advice, but it does show that it pays to strike while the iron is hot.

Finding a product market fit was another topic we touched on, and each panelist had variations on the idea of focusing fast.

"We get a lot of startups that have an awesome technology that can really solve a lot of problems... if you're not focused on one solution, for one industry, you're spreading yourself too thin," Sidopoulos said. "You have to really know your customer, what their challenge is, what they want solved, and what they're willing to pay for. Get that focus on one, get it accomplished — then you have a story, then you have someone who's using your technology. Then pivot to another case. It really, really helps with investment as well."

Abbeel pointed out that this can be extended to the timeline as well. Sure, you might be able to solve problem Y in six months and problem Z in 18 months — but what's problem X you can solve tomorrow? Someone out there wants to pay you for that, even if you plan on having something way better later.

"To us, that was very eye opening," he said. "Because many others were also very excited about robots, but it was excitement for, in some sense, of the future of the company. It wasn't excitement for the today of the company. To us, that was really the big factor in our decision making, who's excited for robots today?"

Werner actually recommended partnering with a large strategic in the sector, as their deep knowledge (and pockets) can help you build that first use case. Even companies like Tesla (where she worked before The Engine) started out partnering with bigger ones to test and prove out their product.

"You're kind of using your partner to de-risk your technology," she said. "Once you've de-risked that technology, build on that and go farther."

But, she warned, be careful not to fall into the trap of over-specialization or over-reliance on the partner's resources, or you'll find yourself dependent and tied to their roadmap instead of your own.

That's just a handful of the topics we covered today — you can watch the rest of the panel for free right here.

Our goal is to create a safe and engaging place for users to connect over interests and passions. In order to improve our community experience, we are temporarily suspending article commenting