Anthropic has more cash on hand than every single company except for Google, Berkshire, Goldman, and Morgan Stanley. Basically capitalized like a bank
Anthropic:
$134b of cash today
$141b of total capital raised
<$10b of total burn to get to $16.8b in quarterly revenue (Q3 figure that was widely reported, Q3 not in S-1 yet)
1
1
360
First time we got a zero-shot task completion, I gasped. It's a good model. Great work @jang_yoel and team!
Congrats to the Dream Labs team!
Have been testing their model on our robot and we are really impressed with the quality for the price. Making inference and teleoperation better + cheaper is key to making robots abundant 🫡
1
3
30
992
Kyle Noble retweeted
text-to-cad plugin extension is now live on codex!
- generate 3D models as STEP, STL, 3MF, or GLB
- DFM for printing, sheet metal, CNC, injection moulding
- connect to fab services like bambu and sendcutsend
100% open source and free, runs locally inside the codex desktop app
58
174
1,682
139,363
Really excellent post on the economics of robots by Anthropic. Cost is the critical input for adoption. (We also wrote about this in depth here: robo.inc/blog/robotics-model…)
anthropic.com/research/what-…
2
21
162
18,222
Kyle Noble retweeted
robots are going to be part of everyday life
we think they should have a friendly face!
9
4
66
6,735
Had a great time talking robots with Brian. You should sign up for Trove. It’s a great newsletter
Trove 004 is out! 🤖
Huge thanks to the @roborobots founders @KyleNoble and @gpudad for the conversation.
We broke down:
- The team's backstory
- The data flywheel that will be the core of the business
- The bottlenecks facing all robotics companies
substack.com/@brianderario/n…
2
12
778
Billions of robots
We have a huge news to share today!
Today we are unveiling the first truly accessible RL robot - welcome Microduck
A 25 cm tiny open-source biped with 15 actuators and packed with sensors (camera, speaker, LiDAR, NFC, bluetooth, wifi, etc) that you train yourself with reinforcement learning.
It's also playable out of the box with more than half a dozen fun and playful pre-trained policies to have it walk, sit, crouch, roller-skate, pick up objects with its articulated beak, and recover on its own.
And all for less than $400.
See all the details, play with the simulator and order it at: pollen-robotics.com/microduc…
(video with sound on 🔊)
6
324
Flywheel central. Cheap robots beget cheaper robots.
We believe robots are going to reshape the world, making work and businesses that were previously impossible a reality.
Over the last three months, we’ve been working on a robot useful enough to do real work and affordable enough to deploy at scale.
1
2
25
2,112
We believe robots are going to reshape the world, making work and businesses that were previously impossible a reality.
Over the last three months, we’ve been working on a robot useful enough to do real work and affordable enough to deploy at scale.
17
19
103
76,209
Also available on our blog with interactive graphs and some more incredible design by @gpudad:
robo.inc/blog/robotics-model…
16
666
a note from tokens per watt ball knowers: models as smart as the leading open weight models, priced at the market clearing inference price, are more economic for oai than for any other provider.
major price cuts today:
*80% drop for GPT-5.6 Luna, now $0.20 per million input tokens and $1.20 per million output
*20% drop for GPT-5.6 Terra, to $2/$12
*GPT-5.6 Sol gets Fast mode in the API, up to 2.5x the speed for 2x the price, same intelligence
4
391
Capital is desperate for prophets.
My main point is that you have significant influence over the market. Posters run the world. The most important money managers religiously watch everything you put out. The financial media is now a collection of individual X accounts and you have earned your spot at the top of the hierarchy by creating exceptional content.
To think otherwise would be discounting your influence. Not here to gravedance or celebrate anyone's downfall, I'm just covering the market and trying to connect dots live.
2
671
Workshopping an idea that we should think about this like F1.
Conventional wisdom is that winning an F1 championship can be attributed to something like 50% engine, 25% chassis, and 25% driver.
The rough analogy here is model as the engine, harness as the chassis, and prompter as the driver.
A great model can produce valuable outputs independent of its harness but to perform at the frontier, it needs the post trained harness, and to a lesser extent, a truly capable prompter.
It doesn't make sense to think about model performance independent of the harness because they've become indistinguishable.
On harnesses, I vacillate between three beliefs:
- the less harness, the better. Models are the magic
- post training a model and harness is dramatically better and the model providers win
- harnesses have real independent value from the model
I have no idea which is right.
1
2
177
Nvidia seeding a new frontier lab as Jensen said he wished they had done with Anthropic but weren’t financially positioned to do so.
We are announcing a long-term strategic partnership with NVIDIA. NVIDIA is making a substantial investment in SSI that will let us 10x our compute in the next 12 months. We reached the point where our research is worth scaling and with this partnership we will be able to. We are honored by NVIDIA’s conviction.
6
1,128
People choose Palantir because they need a system that forces their people to stop doing stupid, expensive, custom stuff and the people revolt because they justify their existence through stupid, expensive, custom stuff
5
403