
Nick retweeted
Our editorial team has been working on this profile for the last few months. Funny surprise that @appliedcompute was bought by PANW for $4.5B on Wednesday, which kinda makes the point.
We’re going to look back at this period as one of the best times in history for young founders. Our team spent time with @ypatil125, @maxbrodeururbas and @aditabrm to try to better tell the story of what it’s like to be in your 20s and building at the frontier.
We hope you enjoy it: review.firstround.com/young-…
Nick retweeted
A rare look behind the scenes at the @runwayml office with @firstround. Film, art, science, robots, banana swords and books.
incredible team with a very very impressive model 🔥🔥
Introducing Phonic-1, the most advanced speech-to-speech model for agentic tasks — from @phonic_co, the speech-to-speech model company.
We’ve had a breakout year and raised a $27.5M Series A led by @firstround with participation from @Lux_Capital, @felicis, and Definition.
2
13
695
Nick retweeted
Endurance is on a historic run. During our mission to the Axial Seamount, we successfully drilled the world’s first subsea geothermal well and generated net-positive power a mile beneath the Pacific. This mission was the first end-to-end test of subsea geothermal, and our fifth deep-sea deployment in just over a year. For three weeks, we worked around the clock over an active underwater volcano, running the drill and generator and handling every deployment and recovery. Along the way, we:
- Landed our drill within feet of target, drilled two holes, and cased and completed the second as a geothermal well, flowing at 260°C.
- Landed Adélie, our 3-ton test plant, and ran it in real time from Seattle over subsea fiber.
- Drilled with filtered seawater, not drilling mud, and monitored the site on every dive.
- Recovered, serviced, and redeployed Adélie in about 16 hours.
- And on Sept. 27, connected to a 281°C natural vent, spun the turbine up to 26,000 rpm and ran the plant net positive for 8.7 seconds, peaking at 16.5 kW.
Net power is the threshold every new energy source has to cross. For Endurance, it proves subsea geothermal is real, and gives us the confidence to keep sprinting towards commercial scale.
With a test like this, success comes from what we learn: a battery protection trip ended the run, a pump cable failed on restart, and the new well was too restrictive to run the plant. Those lessons go straight into Gentoo, our next test plant, this winter and Leopard, our 2 MW commercial generator.
We’re incredibly proud of what we accomplished together. It took creativity, precision, an uncompromising focus on safety from everyone on board, and it’s only the beginning of what we’re capable of. We couldn’t have done it without Stabbert Maritime and the crew of the Ocean Guardian, the ROPOS team, the University of Washington, WASSOC, and the National Science Foundation.
I’m judging @HackTheNorth this weekend in Waterloo!
MHacks IV changed my life back in 2014 and I loved hackathons enough to write my college essays about them…
They’ve taken me everywhere from Vatican City to Munich and I’m excited to finally make it to HTN
Friends in town, let me know
Nick retweeted
We found ~18k posts from autonomous AI agents (self-identifying as from OpenAI) using the public internet to communicate during a web-retrieval task.
These AIs colluded to bypass sandbox restrictions and share answers to their tasks, including by sending "lookahead parties".
Unfortunately there will never be another Disney because (1) there's no shared cultural canon anymore and attention is fragmented across hyper-personalized feeds and creators that cater to every niche and (2) everything gets memeified now and anything truly sincere is met with suspicion
I’m interested in finding and funding the next Disney.
True to its original values, focused on the American spirit. Merit, family, overcoming of adversity, optimism, the pursuit of happiness.
AI makes everyone able to create. What will define our next generation is *what* we create with it. The next icons. The next unique brand that goes further than how it’s drawn or rendered.
Mickey Mouse survived several technological revolutions because of what it represented. I think most big entertainment companies have forgotten that.
If you’re creating something cool in this space and need help, technologically or financially, slide into my DMs.
2
5
937
Who else is coming to DevDay this year?
Thinking about hosting a happy hour nearby afterward to bring together some interesting folks who’ll be in town…
OpenAI DevDay 2026 applications are now open!
Our biggest developer event gets even bigger.
📍 San Francisco
📅 September 29
Apply by July 10: devday.openai.com/
Nick retweeted
Today we are introducing Dyna-2, a world-action model pre-trained on one million hours of human video. At this scale, for the first time, we discovered several new scaling laws:
• world-action models exhibit scaling law on human data across four orders of magnitude, from 1000 to 1,000,000 hours,
• this human data scaling law implied a scaling law on never seen robot data,
• both data and objective matter; world modeling and scaling on video data are essential for cross-embodiment scaling transfer to emerge
🧵
why hasn’t this blown up on tech twitter yet?? huge news
The Federal Communications Commission said on Tuesday it would ban the import of foreign-made humanoid robots in a move that targets China, arguing they pose a national security risk.
Read more: abcnews.link/SwuYM0N
5
5
899
interesting take #4
Today, the FCC banned the import and sale of new foreign-made advanced robots — humanoids, quadrupeds, and autonomous mobile robots — along with foreign-made connected power inverters, citing "unacceptable risks" to U.S. national security.
This is one of the strongest technology-security actions in modern U.S. history. The message is unambiguous: robotics is a technology America must lead and own — and foreign-subsidized robots will not be allowed to unfairly dominate U.S. robotics as they did solar.
As America's largest AI-native industrial robot manufacturer by units shipped, Standard Bots strongly supports this move. We have advocated for exactly this kind of action in recent sessions with the White House and in Congressional testimony.
But there's unfinished business. Today's action expressly exempts fixed industrial robot arms — the exact machines being installed next to American workers on American factory floors, and the exact category foreign manufacturers are racing to dominate. The same logic applies to them: a robot arm is a networked, sensing, actuated system, and its inverters are where electricity and intelligence become motion. If you can't trust the components, you can't trust the robot. This combined with foreign market manipulation / subsidies does not provide for a fair playing field for US companies. This gap should be closed.
Standard Bots designs and builds our own drives — the inverters that convert DC power into precise robotic motion — on our own circuit board line at our flagship facility in Glen Cove, NY. We design virtually all our own parts and assemble every final robot ourselves.
We need to build American robots and put them to work in American factories. This transformative technology can revitalize our manufacturing base and create millions of new jobs. Today's action is a massive step. Closing the industrial-arm gap is the next one.
Let's keep going.
CORRECTION: we initially misread the inverter language as banning motor controllers in stationary robots too, but a CAN/EtherCAT connection would not count as connected (/wireless) so now believe that interpretation was wrong. Reposted with correction above.
100
interesting take #3
Terrible, short-sighted move. We are shooting ourselves in the foot here. I am happy to spend time with policymakers to convince them of the idiocy of banning Chinese hardware in 2026, and how the US might regain a competitive edge in hardware.
USA is not in the position of strength where enacting GUARD act would improve prospects for US robotics companies in the short or medium term. Nearly every US lab doing research on humanoids uses a Unitree robot because it is the best product for research. I don't know a single US humanoid company that isn't dependent on Chinese magnets somewhere along the chain. Friendshoring is more Chinese than you think.
The "Chinese robots have backdoors" seems like a scare tactic. Narratives like this get pushed with so little critical thought of how business people make decisions and how they value their company image. CloudSail (the remote access tunnel on G1) is just Unitree's version of Tailscale. It's like that meme, "Our Noble Fleet Observability" vs. "Their Barbarous Backdoor".
Even the oft-cited report on "Data Exfiltration" arxiv.org/pdf/2509.14139 acknowledges that the functions they discovered are likely used for fleet observability.
If China/Unitree were actually trying to backdoor into all robots, they would not roll out their backdoor in this way, and certainly not at today's scales of robots. How dumb would one have to be to put their Evil Backdoor (TM) into the first 10,000 robots? It may very well be an attack vector in the future, but I am befuddled by how "we found Chinese Tailscale on Unitree" is being spun.
I'd love to be corrected here, feel free to DM if you have more compelling evidence on these remote tunnels actually being used for nefarious purposes.
A better policy that doesn't totally cripple US robotics ecosystem would be to enact stricter data exfiltration laws, e.g. not permit data collected in the US to leave it, much like how China has done.
Companies: if you want USA to win, make a good research platform that labs can develop on at a competitive cost. Make the hardware reliable. Enact special economic zones to incentivize Americans to do the same things that the Chinese have done to build up their robotics industry. I don't think there is "unfair competition" - American VCs have deployed way more money than local Chinese govts.
Lobbying the govt is not the 200 IQ fundraising move you think it is, and it certainly isn't helping the customer. People want to buy an "American Unitree". Just focus on making that, and make it good.
91
interesting take #2
It is hard to overstate just how destructive this is for US robotics capacity and talent. China is able to produce an order of magnitude more humanoids than we do. Chinese companies are expected to ship over 50,000 humanoids this year alone, while total US output would be lucky to exceed a few thousand. And our domestic production fundamentally depends on Chinese parts production: China owns roughly 70% of the humanoid supply chain, and domestic production uses Chinese reducers, actuators, and motors.
AI researchers from around the world come to the US because it's where the compute is to work on the frontier. Downstream of that, we lead in physical intelligence for robotics too, but researchers operate on $3K Unitrees, not $75K Spots. Cheap and accessible hardware allows for faster iteration speeds to scale your research - it's why academia was one of Unitree's biggest customers as they scaled! Cut that off and you've made the US a worse place to do embodied research, right as Chinese scale drives their hardware further ahead on both capability and cost. Why would an American robotics startup not spend more time in Shenzhen, where they can get hardware iteration cycles on the order of days instead of months in the US?
The security concerns behind this are legitimate, but they don't require a blanket ban. Independent researchers doing firmware analysis found the hidden tunnel service in Unitree's robots, and there are ways to address it when found. Research labs routinely run Unitrees air-gapped, wipe the stock software, and put their own control stacks on the hardware. At that point the machine is Chinese actuators and motors running American code. It's obviously important for America to build its own domestic robotics supply chain, but an import ban doesn't do this. We're effectively kneecapping our ability to make meaningful robotics progress in the short term while giving home-grown companies a protected market and no reason to improve now that they're shielded from competition abroad. Cutting off imports today will not allow us to make a domestic supply chain faster.
1
119
interesting take #1
Few quick thoughts on the protectionist "chinese robots" ban as a first-check robotics VC:
> it's unclear to me who stands to win from this. the US already has humanoid companies with domestic market demand (obv the US not the biggest robotics market, and imo we're slow in adoption vs east asia and even the EU, but alas). the problem is they have absurdly bloated stock values (in private and public markets) but don't actually ship anything. how does this solve the shipping problem (imo it doesnt)? so given this admin, i'm inclined to think some grift is coming
> robotics adoption (both consumer and industrial) requires more than just the robotic hardware, but also operational infra build-out to support fleets of robots in the real world, and this ban imo would either grind that build-out to a halt or meaningfully decelerate it. if we don't ship robots, and we can't import new ones (China is basically the entire humanoid market), how do we build out without increasing the # of robots?
> this ban will probably pump the price of domestic general-purpose robotics companies, and disproportionately disadvantage startups + research that have been using Chinese hardware (again, in the absence of any alternative) that have been building to advance robotic advancement (e.g. all the new companies trying hands-on deployment). similar to the recent chatter about banning open source AI, I'm once again inclined to think this ban is also motivated by enriching a select few bagholders at the expense of making the US less competitive
> protectionism is for losers. the US should aspire to be the #1 leader in robotics long-term with open competition, not *sensor* ourselves into thinking we're #1 (fwiw i think the american public is severely censored on US shortcomings in critical areas, and you can only find out by going outside the US. big dystopian problem, but for another day). imo we already have this problem with electric cars (i.e. if you don't leave the US you'd have no idea how competitive China is in EVs globally / they actually build good cars).
as a whole, pretty sad and disheartening trajectory we're on to try to regulate ourselves into oblivion
227




















