frontier tech events @autonomousevent // prev @GoldmanSachs @dartmouth
New York, NY
Joined October 2021
- Tweets169
- Following1K
- Followers132
- Likes261
The sold-out tag is great until 100+ people beyond your anticipation start lining up.
Crisis mode began 10 mins before the doors opened at AUTONOMOUS 2026.
We'd run out of badges
We'd hoped for 500 people but blew past 600
From WiFi emergencies to people not being able to move around freely
We handled everything that day.
What seemed detrimental to us
Is exactly what nobody cared about.
Because they came for the people in the room.
The right panels with the leaders who are shaping the future of Physical AI & Robotics.
Still getting that bigger venue I passed on before.
Autonomous 2027 is loading...
1
1
58
Rohan Savla retweeted
Egocentric video data is "booming"...but for how long?
The last piece yet to be solved is tactile data.
@JayLiStanford's vision is a future where the majority of mass data is still purely vision, but the higher quality, lower quantity data is only possible with touch included.
Watch the full podcast below 👇
What happens if the brains outperform the hands?
Hardware falls behind, failing to keep pace.
Proception builds to close this gap. Their ProHand, a tendon-driven hand with 22 degrees of freedom in a human form factor, and ProGlove, a sensor glove that captures the touch data egocentric video cannot, is how they do this. In this episode, we talk to Jay Li, Founder and CEO, about why he thinks hands are now the bottleneck for the whole robotics stack.
This year, they're ramping production from single digits a month to hundreds. We cover why tendons and five fingers, how glove data compares with teleoperation, a first model trained with zero teleoperation and zero robot data, and how Jay expects the humanoid market to split.
@JayLiStanford @proceptionAI
Full discussion 👇
01:44 The 2027 humanoid prototype and the hands business
06:51 Scaling to 100s of hands a month
11:41 Why tendons
13:22 Why five fingers, not three or four
16:18 Where tactile data fits alongside egocentric video
19:58 Why the glove carries 96 sensors
22:38 Closing the embodiment gap
25:12 A tactile model trained with zero teleoperation
27:15 Glove data versus teleoperation on cost and scale
31:18 Why hands are now the bottleneck
37:26 Will humanoids consolidate like LLMs
42:00 What has kept the team going
Rohan Savla retweeted
Kleiner Perkins just led a $30M round for Nidus, a startup teaching factory robots to build parts for America’s defence industry.
Nidus came out of stealth this week with over $1M in bookings from defence factories, all in about 4 months.
↳ Nidus starts with wire harnesses, the bundles of wires that carry power & signals through machines
↳ People control robot arms to prep the wires, & that data trains AI models that run on the robots factories already own
Founded in March by @jsealynyc, @harrisonnguyen, @DwivC & @CanePunma, who worked together on the AI hardware team at defence startup Vannevar.
@kleinerperkins led the seed & @costanoavc led the pre-seed.
🎥: @kleinerperkins
Myth: the winners in robotics will be the new AI labs with the biggest models.
Data: the companies actually deploying robots for paying customers started way before the hype.
@BuiltRobotics started in 2016
@burro_ai started in 2017
@DustyRobotics started in 2018
@FortRobotics’ founder @samueljreeves started his first robotics company in 2005
That’s 8 to 20 years of robots working out in the real world.
And the real world breaks things no AI model can predict.
Built’s robots kept glitching. The cause? A bird sitting on the GPS antenna.
Dusty’s robots spun in circles every October. One part didn’t like the cold.
You only find that stuff after years in the field.
Here’s the advice I took from being in a room with 100+ Robotics founders,
If you’re just getting started, get your robot in front of a real customer as early as you can.
Because years in the field are the one thing you can’t skip.
1
69
Rohan Savla retweeted
Google DeepMind's new robot brain is already sorting car parts on Mercedes-Benz production lines, with a 96% success rate.
Gemini Robotics 2.5 is its most capable robot model yet, tested on 200 tasks, from tying a bow on a gift box to pipetting liquid.
↳ 30% to 100% higher success than its previous models, while finishing tasks 20% to 30% faster
↳ Running on @Apptronik's Apollo 2 humanoid, it sorted parts into kitting trays at Mercedes-Benz plants in Hungary & Germany
Built by @GoogleDeepMind, with a training boost from Gemini 4 Argon, which turns a single camera view into thousands of simulated practice runs.
🎥 @GoogleDeepMind
I've spoken with 100+ robotics founders & engineers before the AUTONOMOUS Event.
Within 10 minutes, I could tell which category you’d fall under
You are either an early raiser or early revenue
Early Raisers are a cracked team of PhDs, huge vision, massive round.
While Early Revenue are the deployers who deal with install costs, maintenance, support, & training where customers can't find enough operators.
I am not against capitalising rounds, especially coming from Goldman Sachs.
That's definitely something you should strive for
But only after figuring out what moat you actually have and how sustainable it is.
Because funding dries up pretty quick without real revenue.
1
76
Shinkei are a super interesting startup.
Reed Ginsberg, their CTO, actually joined us on stage earlier this year to speak to us.
One thing he mentioned that I still can't get out of my head, 66% of all fish harvested never makes it to a plate.
Two-thirds of everything we pull out never gets eaten.
Massive industrial waste problem if that much product never gets used.
One of the standout speakers on stage just because the problem they solve is so niche, no one had heard about it. With robotics, that's kind of the point.
we have a winner for the most absurd AI startup ever
this Peter Thiel backed company uses AI robots to stab fish in the brain so they taste better and last up to 3x longer
it's called Shinkei and it modernizes a centuries-old Japanese technique
dozens of Michelin-star restaurants are already lining up to buy the fish
here's how it works:
> the problem is when a fish gets caught, it usually just suffocates on the boat
> that stress floods the meat with hormones and lactic acid, which hurts the flavor and texture
> the old Japanese fix for this was spiking the fish through the brain right after the catch, which stops all that
> but doing it by hand is impractical for most American fishing crews
> so Shinkei built a fridge-sized robot that sits right on the boat's deck
> the crew feeds in a fish and AI identifies the exact species, so it knows exactly where the brain and gills are
> the robot then drives the spike through the brain and cuts near the gills in about 7 seconds
> the fish goes into an ice slurry so the blood drains out
> so the fish doesn't sit there stressing out, and that tasty flavor and texture is preserved
example number 5,392,417 of how AI can be applied to literally anything
ALT A gloved fisherman on a boat feeds a live salmon head-first into the steel intake of a white Shinkei Systems robot, with the ocean behind.
1
4
243
More money is going to fewer rounds in robotics. The capital is concentrating as the industry grows.
Publishing your work on X, in public, for everyone to see is the best way to grow your project
v4 is flying!
ground-up redesign, 6 kg lighter, new propulsion system, stiffer structure, and rewritten firmware modules.
hover stability is solid and control authority is significantly better.
team’s been cooking
1
58
The robots I thought we were getting vs what we got
Why are we still dragging 100-pound trash bins through freezing rain, snow and 6 AM dark like it's the 19th century?
Meet Curbie, an autonomous robot built by Revere Robotics that is finally killing off this miserable weekly chore. It automatically hauls your waste bin to the curb on collection day and brings it right back, completely unsupervised.
The genius is that it works with what you already have. Instead of forcing you to buy expensive proprietary gear, it couples directly onto the existing steel axle of your current municipal bin.
Packed with high-torque brushless motors and heavy-duty winter autonomy, it is specifically engineered to muscle heavy loads through deep snowdrifts and icy public sidewalks.
Rohan Savla retweeted
Hugging Face just put the world’s top robot AI models on 100+ real robots.
Even after 100 training demos, not one model finished a single task on Unitree’s G1 humanoid.
Before, every lab tested robots its own way. 10 labs gave 10 different answers.
RoboColosseum now tests every model on the same robots, in real homes & restaurants:
↳ 140+ in-the-wild scenes, starting in Singapore
↳ 5 robot models incl. @NVIDIA GR00T & @physical_int π0.5, up against GPT-6, Opus 5.5 & Grok
↳ The general-purpose AI models held their own against specialised robot models
↳ Every run, failures included, goes public on Hugging Face
Built by @huggingface, @BitRobotNetwork, @eaic_sg & NUS professor @DJiafei, with 10,000 free testing hours for academic researchers.
🎥 @RoboColosseum
Recommend bookmarking this for your weekend
incredible perspective on things as they are
The benefits of abundant intelligence will accrue far outside of San Fransisco.
Spent the last year writing about capital, sovereignty, and the coming reordering of economic power.
THE COMMANDING HEIGHTS
Commanding-Heights.com
1
90
Rohan Savla retweeted
Robotics & industrial machinery are now a national security issue for the US.
A Presidential Proclamation signed on 28 Oct, the outcome of the Section 232 investigation, found that the US relies on imports for most robots, industrial machines & their parts.
The talks cover imports from any country: Canada, Germany, Italy & Japan supply half of them, and China 8%. There are no new tariffs yet. The Department of Commerce & the US Trade Representative will negotiate with trading partners & report back within 180 days.
Safety standards might be the only thing holding humanoids back
Humanoid safety standards are still being written — and Agility is helping write them.
Our Sr. Director of Robotics Safety Kevin Reese is Project Leader of ISO 25785-1, the first safety standard for dynamically stable industrial mobile robots 👇
youtube.com/watch?v=WpGs7DKe…
#Agility #Safety #Digit
1
1
200
The largest battle consumer robots (mostly humanoids) will have to win is the one of human acceptance and trust.
Cheap hardware and zero shot models can only fix so much
64
Rohan Savla retweeted
Manufacturers aren't ready for what's to come.
Physical AGI is revolutionising the factory floor, a bolt, a simulation and an edge case at a time. What we are seeing unfold on production lines is a technical revolution comparable only to that of the Industrial Revolution.
In this episode, we talk to Lukas Pankau & Gilwoo Lee of @IndustrialNxtAI, who are reshaping the manufacturing landscape with R2S2R, their new 10,000 sq ft facility and more.
IndustrialNext has had a stellar year. From acquiring Zordi, a BMW Group Technology Office partnership and much more. We cover their technology, when physical AGI will be truly achieved, and the steps to get there.
Full discussion 👇
06:28 When physical AI actually reaches the factory floor
09:40 How learned robots differ from traditional automation
11:31 China, hardware costs and speed to market
14:55 Reshoring, shifting regulation and the first target industries
19:13 Where learned control ends and deterministic control begins
23:20 Why zero shot has no place in a factory
25:20 Fitting robots into legacy factories
29:51 What transferred from strawberry picking
34:13 Inside the Real2Sim2Real pipeline and its gaps
46:10 The new 10,000 sq ft facility and its 50+ robots
50:35 What has kept them going on their journeys