Audrow Nash Podcast

Episode 7 of the Audrow Nash Podcast.

Missing dies and multi-year tooling lead times can ground an aircraft for years and cost millions. Edward's bet is robotic sheet forming that turns some of that into days.

I talk with Edward Mehr, CEO and co-founder of Machina Labs, about high-mix manufacturing floors that switch parts and materials without rebuilding the factory, why physics FEA was too slow for their process so they lean on data-driven models, and why he thinks robotics startups need roughly 10× improvement to survive the failure margin.

You'll like this interview if you care about aerospace/defense manufacturing, process robotics, or how hardware founders pitch venture while carrying CapEx. The forming and modeling stretch (about 7:33–34:45) is the densest engineering bit; the 10× / market-choice stretch (47:16 on) is the densest founder bit.

EPISODE LINKS

- Machina Labs: https://machinalabs.ai/

- Machina Labs LinkedIn: https://www.linkedin.com/company/machina-labs/

- Edward Mehr LinkedIn: https://www.linkedin.com/in/edward-mehr/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Cold open

0:29 Episode intro

1:51 Introducing Edward and Machina Labs

7:33 AI in manufacturing

15:19 Data models vs physics-based models

18:51 Capturing data and modeling

34:45 About Machina Labs

39:15 Machina Labs funding

45:11 Being in LA

47:16 Why make a company that is 10x better

58:43 What if you weren't venture funded?

1:02:53 On growing quickly

1:10:26 Humanoids: challenges and Elon Musk

1:20:54 AI trends: transformers and hallucinations

1:35:32 Thoughts on manufacturing

1:42:15 Advice to get involved in manufacturing

Episode 6 of the Audrow Nash Podcast.

For years Open Robotics staff were the default maintainers. After Intrinsic acquired the for-profit side, Tully and Geoff argue that model can't be the long-term plan for ROS and the rest of the stack.

I talk with Tully Foote (community advisor to OSRF; open source lead at Intrinsic) and Geoffrey Biggs (CTO of the Open Source Robotics Foundation) about Willow Garage and ROS lore, why ROS 2 existed for product teams that wouldn't ship on ROS 1, what Intrinsic capital unlocked (including RMW Zenoh), and how OSRA is meant to replace a single-company "benevolent dictator" with paid members plus meritocratic contributors.

You'll like this interview if you ship on ROS, care about middleware and simulation infrastructure, or need a straight account of Open Robotics after the Intrinsic deal. OSRA motivation and membership (about 1:04:28–1:29:18) is the densest governance stretch; Zenoh and infra (47:38–53:27) is the densest engineering stretch.

EPISODE LINKS

- OSRA: https://osralliance.org/

- Open Robotics: https://www.openrobotics.org/

- Open Source Robotics Foundation LinkedIn: https://www.linkedin.com/company/open-source-robotics-foundation/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Episode intro

2:47 Willow Garage and Open Robotics lore

5:38 Player and ROS lore

9:06 Willow Garage and the birth of Open Robotics

13:17 Tully and Geoff's roles

22:13 Why make ROS 2

24:39 Player and ROS 2

25:42 Towards a robot revolution

29:16 Gazebo for robot simulation

34:39 OpenRMF project

38:22 Open Robotics infrastructure

44:42 Open Robotics before Intrinsic

47:38 Zenoh and bigger bets after Intrinsic

51:41 Rust for robotics

53:27 Intrinsic's acquisition of Open Robotics

58:31 OSRF since the acquisition

1:01:33 How former OSRC folks contribute now

1:04:28 Announcing OSRA

1:07:27 Motivation for OSRA

1:15:20 Merit-based contributors

1:18:06 OSRA consortium

1:20:23 Why companies join OSRA

1:23:55 OSRA membership types

1:29:18 Mentoring and getting involved

1:32:26 Mentorship process

1:34:25 Value of contributing to open source

1:39:19 OSRA badging for contributors

1:42:52 Educating more roboticists

1:47:26 Robotics in the next 5 years

1:52:10 Why open source?

1:55:36 Final thoughts

Episode 5 of the Audrow Nash Podcast.

Seed and Series A can bet on vision. Series B and C have to underwrite metrics. Sanjay walks the gap where a lot of robotics companies stall.

I talk with Sanjay Aggarwal, who leads robotics and automation investing at F-Prime Capital, about why they pulled back from unbounded AV bets toward industrial verticals, how F-Prime's State of Robotics report reads investment mix over recent years, and what he wants from founders (including when human-in-the-loop beats chasing 100% autonomy).

You'll like this interview if you're raising for a robotics company, allocating into the category, or trying to map VC incentives against hardware timelines. The B/C stretch (about 55:02–1:04:31) is the densest fundraising bit; reliability and human-in-the-loop (1:12:00 on) is the densest product-architecture bit.

EPISODE LINKS

- F-Prime Capital (robotics): https://fprimecapital.com/robotics

- F-Prime Capital: https://fprimecapital.com/

- Sanjay Aggarwal LinkedIn: https://www.linkedin.com/in/sanjaykaggarwal/

- F-Prime Capital LinkedIn: https://www.linkedin.com/company/f-prime-capital-partners/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Episode intro

2:13 Introducing Sanjay and F-Prime

4:40 Sanjay's background

11:31 Investing in AVs then industrial robots

15:29 F-Prime's robotics portfolio

26:55 A new type of robotics founder

28:21 Humanoids

35:04 F-Prime's State of Robotics report

39:29 Macro forces for declining investment cash

46:32 Why are companies not IPO'ing lately?

48:51 High interest rates and the VC model

55:02 Challenges raising B and C rounds

1:04:31 Why we're in the early days in robotics

1:09:23 Will labor shortages affect robotics investments?

1:12:00 Reliability and human-in-the-loop

1:17:20 Common pitfalls for robotics companies

1:20:27 How to look attractive to investors

1:23:36 Biggest challenges in robotics

1:25:44 Opportunities for new robotics companies

1:27:53 How would you start a new robotics company?

1:32:12 How to pick a Series A investor

1:34:39 Trends in the robotics industry

1:40:28 Future of robotics

Episode 4 of the Audrow Nash Podcast.

Bluewhite's bet is not a new tractor. It's a kit dealers install so a grower's existing machine is manual in the morning and autonomous by afternoon, aimed at permanent crops where labor is scarce and GPS often isn't reliable under canopy.

I talk with Ben Alfi, CEO and co-founder of Bluewhite Robotics, about bolting actuation and sensing onto mechanical tractors, redundant localization (GPS/RTK, LiDAR, vision, odometry) that re-prioritizes itself, selling through John Deere and New Holland dealerships, and blending into farm culture instead of trying to replace it.

You'll like this interview if you care about ag robotics, go-to-market through equipment dealers, or autonomy in GPS-denied orchards and vineyards. The kit and sensing stretch (about 3:53–18:41) and the labor / operator-role bit (48:14 on) are the densest arguments.

EPISODE LINKS

- Bluewhite: https://www.bluewhite.co/

- Bluewhite on LinkedIn: https://www.linkedin.com/company/bluewhite/

- Ben Alfi on LinkedIn: https://www.linkedin.com/in/ben-alfi-aaa6b1146/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Episode intro

1:09 Introducing Ben Alfi and Bluewhite Robotics

3:53 Outfitting a tractor with Bluewhite's tech

8:52 What types of jobs are they automating?

11:12 How do you do localization? What sensors do you use?

18:41 Adding sensors to the tractor's tools + levels of support

23:36 Being a full solution for agriculture companies

25:14 Sending data to the cloud

31:43 Configuring their robot's behavior + onboarding

34:54 Working with farm equipment dealerships

38:01 Spiral development

40:32 Scaling their operation

44:46 How large is the tractor market?

48:14 Labor shortages and people not wanting to be farmers

56:43 Future of unmanned tractors

1:07:14 Data as a Service + Fellowship

1:13:21 Blending into the agriculture culture

1:17:58 Making systems safe

1:21:48 Recent Series C investment round

1:23:48 Working with the community

1:29:26 Episode outro

Episode 3 of the Audrow Nash Podcast.

Most robotics startups still rebuild their own mobility stack before they can work on the thing that makes them different. Polymath productizes point-to-point autonomy for off-highway machines so customers can buy that layer instead of raising millions to reinvent it.

I talk with Stefan Seltz-Axmacher (CEO) and Ilia Baranov (CTO), co-founders of Polymath Robotics, about a modular stack from sensor drivers to throttle and steering, why big expensive vehicles can be easier to automate than home robots, and how they keep machine learning above a deterministic safety core they want to certify.

You'll like this interview if you build industrial mobile robots, sell into mining or ag, or care about safety certification for autonomy. The reinventing-the-wheel stretch (about 4:33–23:11) and the big-vehicles / safety bits (1:01:38 and 1:15:52 on) are the densest arguments.

EPISODE LINKS

- Polymath Robotics: https://www.polymathrobotics.com

- Stefan Seltz-Axmacher on LinkedIn: https://www.linkedin.com/in/stefanseltzaxmacher/

- Ilia Baranov on LinkedIn: https://www.linkedin.com/in/iliabaranov/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Start

2:51 Introducing Stefan, Ilia, and Polymath Robotics

4:33 Let's stop reinventing the wheel

23:11 Is your code open-source?

26:42 Why focus on mobile robots?

29:27 Stefan's experience starting Starsky Robotics

33:45 A good task to automate and not get lit on fire

36:55 Integrating with sensors

40:39 Approaching the problem from another side than most robotics startups

43:20 Ambitions for Polymath Robotics

48:13 Running a different kind of business with fewer hats

50:22 Stefan's bankrun t-shirt campaign

54:40 How Polymath plans to grow

57:14 How Polymath uses machine learning

1:01:38 Big and expensive vehicles are easier to automate?

1:08:18 Thoughts on AI and machine learning

1:15:52 Making safety certified systems

1:27:12 Live demo!

1:32:46 Web tools and robotics

1:39:11 Automate It podcast

1:50:20 Advice for getting involved in robotics

1:54:27 Stefan's former startup experience and path to Polymath

2:01:18 Thoughts on side projects

2:04:56 What Stefan and Ilia are excited about in robotics

Episode 2 of the Audrow Nash Podcast.

Warehouse aisles were built for people. Melonee walks through why a dynamically balancing biped can keep a small footprint, reach high, and still carry a competitive payload when wheeled bases start tipping or eating payload with accessories.

I talk with Melonee Wise, CTO of Agility Robotics, about Digit doing tote manipulation for machine-assisted ops, how they compare to people on throughput rather than burst speed, the ethics and labor-shortage frame, and why she thinks humanoids stay industrial for years before homes.

You'll like this interview if you care about humanoids in logistics, US manufacturing supply chains for robots, or a sober take on LLM hype versus hardware and controls timelines. The why-humanoids stretch (about 12:07–19:01) and the timeline bit (1:02:28 on) are the densest arguments.

EPISODE LINKS

- Agility Robotics: https://agilityrobotics.com/

- Agility Robotics on LinkedIn: https://www.linkedin.com/company/agility-robotics

- Melonee Wise on LinkedIn: https://www.linkedin.com/in/meloneewise/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Start

0:33 Episode introduction

1:45 Introducing Melonee + Agility Robotics

3:56 Introducing Digit

5:24 How Melonee joined Agility

9:31 Pivoting from curb to doorstep delivery

12:07 Why humanoids?

16:18 How does Digit compare to people?

19:01 Ethics of humanoid robots

25:53 LLMs, automation, and white collar jobs

37:36 Robot factory + suppliers in North America

42:56 Reestablishing a trade class in the US

53:45 Digit's hands

56:43 Thoughts on humanoids

1:02:28 Timeline for humanoids

1:06:04 What if you had your own ChatGPT?

1:09:15 Best time to be in robotics or worst time?

1:10:11 Advice for new roboticists

1:13:32 Amazon trials

1:15:30 Agility Robotics in the next 2-5 years

1:19:16 Humanoids and spreadsheets? Is this the tipping point?

1:21:34 What are you excited about in robotics now?

1:23:47 Wrapping up

Episode 1 of the Audrow Nash Podcast.

Electric Sheep doesn't lead with robot-as-a-service. They buy profitable landscaping companies, put small electric mowers on those jobs, and use the fleet as a training ground for an outdoor world model they call ES1.

I talk with Nag Murty, CEO and co-founder of Electric Sheep Robotics, and Michael Laskey, VP of Autonomy, about learning robots without classical maps, why they shrank the mower to cut accident severity, and why buying the service business beats selling subscriptions while the model is still maturing.

You'll like this interview if you care about outdoor autonomy, embodied AI that has to hit six nines instead of SaaS-quality answers, or hardtech business models that optimize for deployed robots and data. The safety-by-form stretch (about 42:12–54:29) and the RaaS critique (1:11:40 on) are the densest arguments.

EPISODE LINKS

- Electric Sheep: https://sheeprobotics.ai/

- Electric Sheep on X: https://twitter.com/sheeprobotics

- Electric Sheep on LinkedIn: https://www.linkedin.com/company/electric-sheep-robotics/

- Nag Murty on LinkedIn: https://www.linkedin.com/in/nag-murty-b003383/

- Michael Laskey on LinkedIn: https://www.linkedin.com/in/michael-laskey-4b087ba2/

PODCAST LINKS

- Podcast Website: https://audrownashpodcast.com/

- YouTube: https://www.youtube.com/@audrow

- Apple Podcasts: https://podcasts.apple.com/us/podcast/audrow-nash-podcast/id1716486786

- Spotify: https://open.spotify.com/show/74jWpWiLwsasY2QHtDcl8I?si=6c92796bf9554162

TIMECODES

0:00 Start

1:37 Introducing Nag, Mike, Electric Sheep

3:09 Large world models

6:24 Learning physics

10:09 Robot sensors

17:57 Introducing ES1 as ChatGPT

20:17 ES1's input / output

22:22 End-to-end learning for robots

27:11 Getting data

27:47 Electric Sheep's background

30:50 Business philosophy at Electric Sheep

33:47 Sim-to-real + long tail events

36:38 ChatGPT

42:12 Building for safety

49:54 Safety with AI

54:29 Apple's surprise new features

58:31 Generalizing learnings for other tasks, like snowplowing

1:04:12 AI architecture

1:07:41 Training to mow straight lines

1:11:40 Why not Robot-as-a-Service?

1:17:29 Boring businesses + robots + AI

1:25:02 How do we build the data model?

1:29:22 Getting into dexterous tasks

1:30:29 Building a monopoly

1:31:23 Advice for our new world of AI

1:40:47 Links and contact info