Reflecting on Covariant’s journey as a first-generation AI robotics company, it’s clear that while today’s technology stack is much advanced, many of the hardest lessons around deploying robotics and delivering real-world value are more relevant than ever.
I’m grateful for the opportunity to share those lessons and spend time with the next generation of builders—there’s never been a better moment to be building.
Thanks to the team @BessemerVP for bringing us together.
Building a robotics business today looks nothing like it did two years ago.
We talked to Ury Zhilinsky (Mind Robotics), Sandy Hefftz (Bellboy Robotics), and @TAStinson (former CEO, @CovariantAI)—three leaders who've built and deployed robots at @amazon, @Waymo, and Covariant, now running them in hotels, warehouses, and auto plants.
Here are 7 lessons from the experts 🧵
@tlehuup@adcock_brett We've deployed single arm robots for this use case. These processes run less than 24 hours, and task repurposing is a huge ROI advantage, as long as you can get humanoid performance up near manual rates.
Watching for a bit I was impressed w performance - speed and reliability.
@ericjang11 Agreed - super impressive strides forward! and yet both bounds are still demos, not examples of actual use in the real world.
How far away are we from the point where the threshold for newsworthiness is it working in the wild?
@YorkYang5050 post is quite insightful and very relevant to my Covariant experience.
I see a future world where all three paths lead to durable companies - embodied intelligence first, hardware first, integration... however the last path has had the least focus + capital
The idea I'd add is the diversity of real world is real ;).
What people see as the same "job" - eg: packing a box in a warehouse - has enough variance from company to company, site to site, station to station, that from a deployment perspective it can be near infinite variation
Yes, 2024 is shaping up a big year for robotics!
Introducing @CovariantAI's RFM-1, which just like Sora can generate video, but RFM-1 does it for robotic interaction with the world.
But there is so much more it can do. RFM-1 is a multimodal any-to-any sequence model. RFM-1 tokenizes 5 modalities: video, keyframes, text, sensory readings, robot actions. Any-to-any means RFM-1 can handle any subset on the input and any subset on the output.
We're excited to announce Perplexity Pro:
1. Practically unlimited Copilot searches (300 + a day)
2. GPT-4 for all search queries
3. GPT-4 for text generation as a focused use case
4. Access to new upgrades which will be rolled out continuously over the coming weeks
5. Discord server with fast support from the team
All at 20$/mo (or 200$/yr)!
At this price point, you get everything you need from a Generative AI subscription service for text-based knowledge-work (research, writing, coding, travel, shopping, trivia) and more to come!
Thank you to the team members, customers and partners who have been part of our journey - this investment is a reflection of your passion and excitement for our vision!
Onward we go! 🚀🚀🚀
We are excited to announce our funding of $75M to meet growing demand for increased warehouse efficiency and resilience through AI-powered robotic automation.
https://t.co/yxmd9VWLsy
Working with @dhaliwas first as a colleague & now as a lead investor this is every bit deserved... a player's coach who has helped me see around the corner, dust off after a tough break, and delivered the sometimes needed kick in the pants. plus he's occasionally funny.
"as AI-generated insights and predictions become the status quo, the less flashy work of data maintenance can get overlooked. It’s like building a house on sand" @pabbeel in @VentureBeat
https://t.co/mB9fyYxUY5