Vanguard Defense @Harvard - Prev. @northropgrumman @MITLL
LA
Joined January 2021
- Tweets12.4K
- Following4.8K
- Followers4.2K
- Likes19.9K
Johnny Ni retweeted
ever since I was a little girl I dreamed of joining a struggling second tier competitor
Johnny Ni retweeted
On Thursday we were happy to host Under Secretary of War and former HRC President Emil Michael.
Our discussion covered super intelligence, the threat of effective altruism, and the case for American optimism.
Johnny Ni retweeted
Warhound, the brain child of Ben Lamm and Palmer Luckey, emerges from stealth to genetically engineer dogs for military and law enforcement.
spr.ly/6011BHZNMF
Johnny Ni retweeted
We’re proud to welcome Robinhood Ventures Fund I (NYSE: RVI) as an investor, and glad to have their retail investors with us as we scale our Bank for America’s innovation economy.
Robinhood Ventures Fund I (RVI) has invested $30M in @ereborbank, a bank building for the next generation of American innovation.
Learn more: rbnhd.co/rvi-erebor
Johnny Ni retweeted
If you’re building a data vendor and not focused obsessively on scalability and durability you’re ngmi
Johnny Ni retweeted
Looking for 500MW (minimum 250, max 2GW) of powered land for a PortCo.
Company is completely price insensitive and has billions behind it. Ideally electrified shell is already built.
I understand everyone wants this, but figured I’d throw it out there!
Johnny Ni retweeted
say hi to kled v3, the data collection layer for agi.
any dataset a lab can describe, collected from the real world by 500,000 people in under 72 hours.
Johnny Ni retweeted
The @DiscipulusVent Cohort has arrived in Gundo (feat. @jared_western and 2 reactors)
Johnny Ni retweeted
If you're in LA for LA Tech Week in a few weeks (specifically Oct 14, 8pm), @zanemountcastle (@Picogrid) and I (@DiscipulusVent) invite you to join us for a Gundo bonfire.
This is a once-in-a-lifetime opportunity to see Gundo from the inside, so don't miss out.
Register on Partiful below 👇
Johnny Ni retweeted
At the beginning of last year, Ali Ansari’s company, Micro1, was a recruiting business with $7 million in ARR. It’s now generating over $500 million in ARR and two xAI cofounders invested its latest round.
Read more: forbes.com/sites/annatong/20…
📸: Micro1
Johnny Ni retweeted
The best eval for continual learning is: take a diary that contains a person's daily experiences. Update model weights on every experience, one at a time. Measure forgetting.
This is how humans learn. And the hardest eval for LLMs.
Johnny Ni retweeted
Poseidon Aerospace has expanded into our new 110,000 sqft HQ in Alameda, California!
Our new facility is at Alameda Naval Air Station, a former Navy hangar with an incredible history of building and supporting American aviation. We're bringing aerospace manufacturing back to it.
This facility will become the home of Poseidon’s production and manufacturing operations. Over the next few years, a lot of airplanes are going to come out of these doors.
Johnny Ni retweeted
jakob (@jakobdiepen) moved to el segundo before any other solo GP had set up shop. he quickly became the front door to the young talent solving america's hardest problems.
he now holds keys to the most exclusive demo days .. hard to get into for both investors and founders.
Johnny Ni retweeted
Blackbeard is going into production. The U.S. Navy has awarded Castelion a production order valued at up to $200 million for an initial lot of Blackbeard Hypersonic Weapons, following the 50 EOC rounds ordered in June. Engineered from day one for the production line. Grateful to PAE(A) for the trust and the urgency. castelion.com/news/blackbear…
Johnny Ni retweeted
If Dario wants to slow down the pace of AI, he should just move the company to Europe.
Johnny Ni retweeted
Longer truth nuke:
Hardware investing isn't a refuge from software commodification. There's no defensibility in most hardware. What's more, the way venture describes hardware as "cheaper, faster, at scale" is actually describing a race to the bottom that strips value from OEMs.
Sure, you can sell a lot of picks for a low cost and high demand and print. And you look up market and see most buyers are low profitability organizations, so you figure picks are the place to be. Strip back the layers and realize how much physical labor costs, and that every competitor pays the same for it. The second you build a mining robot you have a choice though: sell it and stay in the (fancy) pick business, which is a race to the bottom, or run it yourself and now you're the labor. The labor budget on any job is a multiple of the tool budget. And a pick gets priced against the next pick, while labor gets priced against the humans it replaces, with the cost of the robot doing it being the thing that's collapsing. It was only ever a bad business because it was made of people.
There's a brief moment in time right now where robotics has not yet been adopted at scale outside the factory, and broadly replacing physical labor is going to be a generational shift in how business gets done. That means there is a window in nearly every industry to eat up incumbents by doing the work they do today with robotics. It's ultimately a labor arbitrage. Incumbents have little way to combat it until they can buy a platform AND orchestrate it, and right now they can't because it doesn't exist in the market, so you have to build it yourself. And then figure out how to drive value with it.
But hardware commodifies, then orchestration commodifies, and ultimately the only thing that doesn't is owning outcomes: the contract, the liability, and the record of every job you've ever run on a customer's assets. A competitor can match your robots on day one but there's no way to buy ten years of history.
If you can replace full workflows with robotics and own that outcome, each additional workflow you can complete compounds your gravity. You start looking a lot like a labor force, which alongside energy and capital are the three fundamental forces of getting anything done. You essentially become infrastructure for the physical world.
