A construction site. A mine pit. A solar farm. @DroneDeploy captures it all from the sky.
DroneDeploy turns drone imagery into maps and models teams can trust. Tracking progress, spotting safety risks, and verifying work across millions of acres.
Underneath every flight: centimeter-accurate positioning from @GEODNET RTK. The map is only as good as the data holding it together.
Regular GPS is fine when “close enough” works.
But for drones, surveying, mapping, agriculture, robotics, and autonomous systems, close enough is not enough.
@GEODNET delivers centimeter-level positioning through the world’s largest RTK network, giving teams the accuracy they need to work faster, smarter, and with more confidence.
GPS shows the area.
GEODNET helps pinpoint the exact location.
Aerodrome is changing how it pays for liquidity.
Emissions are paid in AERO while fees come in as dollars, so when AERO rises, the protocol gets less profitable if fees don't outpace the price. Every week it mints new tokens that are paid to liquidity providers, which is a real cost to the business.
Aerodrome aims to fix this with the AER Engine, which pegs emissions to revenue instead of the token price and caps inflation within a set band. Spend follows what the protocol earns rather than what the token does, so a rising AERO price no longer mechanically pushes earnings down.
It also changes the shape of the cost line. A fixed weekly mint held spending flat regardless of what the protocol brought in. Linking emissions to fees replaces that with a cost that scales with the business.
If your kid has that entrepreneurial spark, we're hosting Founders Weekends nationally through the rest of the summer.
It's a free weekend where they'll learn:
- Idea validation
- Vibe coding
- Sales
- Pitching
With likeminded students + expert mentorship.
Limited spots!
Just closed $11M from @firstround Capital... @proceptionAI! The same day Tesla dropped its trade secret lawsuit against them.
I caught up with CEO @JayLiStanford in Palo Alto earlier this year:
Backstory: Jay led hand engineering on Tesla's Optimus.
He left, started Proception, and got sued by his old employer six days later.
Almost a year of litigation. He called it a "resilience test."
Now it's over, the round is closed, and Proception just shipped its first batch of hands.
What are they building?
ProHand, a tendon-driven hand, 20 DOF plus a 2-DOF wrist, covered in full tactile skin. It reacts to contact like a real hand, because that's what trains the model underneath it.
The real unlock is the data:
Most teleoperation needs a robot in the loop and zero tactile feedback. Proception's ProGlove uses the same sensor skin as the hand: humans wear it, do tasks with their own hands, and it captures real contact data with no robot involved.
Hiring is one of the biggest bottlenecks, Jay says, not supply chain, not the tech. Standard founder problem, just at robot-hand difficulty.
Dexterous hands are the unsolved piece of humanoid robotics. Musk has called it one of the hardest engineering problems out there. Jay and his team think they'll get there faster than people expect...
(Longer episode out soon)
More about the hand here: https://t.co/NYuLFaeioZ
Congrats to @JayLiStanford, @_jaku_xu and the entire team for this milestone!!!!
——-
Weekly robotics and AI insights.
Subscribe free: https://t.co/9Nm01QUcw3