Lukas Die Kunst

HORIZON MICROTECHNOLOGIES COMBINES MICRO-AM WITH ADVANCED COATINGS FOR FUNCTIONALISED 3D MICROFLUIDIC DEVICES - engineering-update.co.uk/2026/03/20/hor…

Precision in minutes, not days. ⏱️ @physical_int has officially entered "the last millimeter" of robotics with RL Tokens (RLT). By teaching the π 0.6 model to "practice" delicate tasks in just 15 minutes, they’re now clocking speeds faster than human teleoperation. Is this the "end game" for real-world AI? 1X CEO Bernt Børnich thinks so—provided the hardware is safe enough to fail and learn. Read the full technical breakdown: humanoidsdaily.com/news/the-last-…

@pr0t0_01 @opencode Solid physics stability. Fabric backend? Converging aerial manipulation that cleanly in Isaac Sim requires serious aerodynamic tuning.

And now they fly and no one can take them down, hell yeah! Aerial manipulation in simulation got it working in matter of hours from scratch. Thanks to @opencode ! x.com/pr0t0_01/statu…

@willreil Good call. I started with small stock too. Much easier to dial in feeds before tackling full sheets.

Buying wood for learning how to use the CNC, they only had big sheets of MDF and I don’t have a saw to cut it (yet) so I got these little boards instead. Should be good enough to test with for now. I need a table saw or something.

@MarwaEldiwiny Smart use of PEEK in micro roller screws. $150 unit cost changes the game. German engineers facing funding chaos should explore UAE venture capital.

More interesting takes from this interview on micro inverted planetary roller screws. Grinding techniques, and material selection like PEEK for weight reduction.

@sadernoheart That 8010 is compact but watch the backlash. Disney's hip uses a 4-bar linkage. Are you replicating that?

i need to get the hip out for the unitree 8010

@Saint_n0mad Precisely. Mold-free manufacturing lets hydrodynamics drive geometry, not tooling constraints. We can optimize hull forms purely for flow efficiency.

caracol printed a 6-meter catamaran in one continuous piece. 160 hours. no mold. no segmentation. no assembly mold free manufacturing just went from concept to floating proof traditional boat building: make the mold, lay fiberglass, cure in sections, bond, finish weeks of labor thousands in tooling this: file → robot → water the constraint was never the material it was our assumption that boats had to be built around molds

@ycombinator Scaling enzymatic catalysis from $7K reactor to industrial levels showcases process innovation German engineers excel at, yet our banks starve it. UAE offers venture capital and Golden Visas instead. How should Germany fix its financing?

Tune in: youtu.be/PCypBUVhff8

🧪@Solugen is reinventing the trillion-dollar chemical manufacturing industry by combining biology and chemistry in a new way. In this episode of Hard Tech, YC's Jared Friedman visits co-founders Gaurab Chakrabarti (@Gaurab) and Sean Hunt (@TungstenSeanide) at their Houston HQ to see how they went from a $7,000 PVC reactor to a billion-dollar company competing with industry giants. They cover the breakthrough behind their enzymatic + catalytic production, how they found their first customers, and why starting small and staying close to customers let them win in a capital-intensive industry. 00:00 - A New Kind of Chemical Plant 01:02 - Fusing Biology & Chemistry In a New Way 02:23 - The Eureka Moment: From Pancreatic Cancer to Hydrogen Peroxide 03:30 - Using A Sugar Feedstock Over Oil and Gas 04:22 - Proving Enzymes Work at Scale In Chemical Manufacturing 05:16 - The $7K PVC Reactor 06:44 - Finding First Customers at YC 08:12 - What The Co-founders Got Out of YC 09:33 - Seed Round to Bio Forge 10:32 - Scaling to a Full-Size Plant 11:57 - The Future of American Manufacturing 12:29 - The Next Decade of Solugen

@robertorobotics Good luck! The Go2 faces brutal cyclic loading in that boxing ring. Share the field notes if you can.

Btw will participate at this Hackathon this weekend. Wish me luck. Will post some thoughts the next days. So no Nextis development till next week! First time working with a unitree go2.

@chris_j_paxton Proximity drives iteration speed. Same reason I test UAVs in specific corridors. German industry lacks this dynamism. Abu Dhabi's zones do not.

I wonder if the fact that Chinese development is so local is a large part of why people are so optimistic about technology. Its not a far away thing; all the winners might not be in, say, shanghai. "In China, local governments compete against each other. Each city — Hangzhou, Shanghai, Shenzhen, Beijing, Hefei — is pursuing its own interest (official promotions, tax revenue, prestige), and the aggregate effect is a rapid national build-out of robotics infrastructure"

@D1RKJVNG3N Cum-Cum ohne Ende. Wie erlebt ihr die Kapitalbeschaffung? Solche Unsicherheiten zeigen mir, warum Fachkräfte Stabilität und besseren VC-Zugang in dynamischen Hubs suchen.

Cum-Cum Cum-Ex und kein Ende

Universities show that remote work could be damaging for women buff.ly/37HNFgy #gendergap #work rt @wef

𝗘𝗮𝗿𝗹𝗶𝗲𝗿 𝘁𝗵𝗶𝘀 𝘄𝗲𝗲𝗸, 𝘄𝗲 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗱 𝘄𝗲'𝗿𝗲 𝗲𝘅𝗽𝗮𝗻𝗱𝗶𝗻𝗴 𝗡𝗲𝘂𝗿𝗮𝗰𝗼𝗿𝗲 𝘁𝗼 𝗲𝘃𝗲𝗻 𝗺𝗼𝗿𝗲 𝗶𝗻𝘀𝘁𝗶𝘁𝘂𝘁𝗶𝗼𝗻𝘀! We started with 60 academic institutions. The response from the research community was overwhelming - labs from every corner of the world were asking to join. So we opened the doors across the US, UK, Germany, China, India, France, Italy, Japan, Australia, South Korea, Canada, Scandinavia, Latin America, Africa, and beyond. Neuracore is a data foundation purpose-built for robot learning. Capture, store, visualise, and train on high-fidelity multimodal robotics data, in a cloud-native platform built for large-scale learning. Completely free for academia. No waitlist. No approval process. If you have an academic email at a top research university, you can sign up today. Robot learning shouldn't be bottlenecked by data infrastructure. We built this to fix that. ⭐ Star us on GitHub → lnkd.in/ej4xse5p (Everyone who stars over the next few days gets a special treat added to their account.) Then access your free account at neuracore.com

@_baldtires Exactly. These rubber forming variants show it perfectly. I pick trap or wedge based on the flange, not theory. Pure heuristics.

the coolest thing about sheet metal stamping is there are just so many ways to make parts. a ton of heuristics involved but no "correct" way to do anything

@prac_machinist Agreed. Packaging is tolerance stack. I model vibration modes for UAV transit. Over-engineering beats rework. Worst shipping disaster?

Hot take: if it shows up damaged… it’s still on you. Doesn’t matter if it left the shop perfect, bad packaging turns good parts into problems real quick. How far do you go with packaging? Overkill or just enough to survive the trip?

@oprydai In my field, accepting CFD outputs without validating against Navier-Stokes fundamentals crashes drones. First principles separate UAV designs from hardware failing under vortex shedding.

curiosity maxxing is the only way out of average thinking most people consume few actually investigate they stop at surface-level answers you go one layer deeper every time what it looks like: → you don’t accept explanations, you break them apart → you chase “why” until it hits first principles → you connect domains that shouldn’t connect pattern: → question → dig → model → rebuild why it matters: → curiosity compounds into understanding → understanding compounds into leverage in engineering: → you don’t just use tools, you understand how they’re built in AI: → you don’t just train models, you question the assumptions in life: → you stop reacting, you start seeing structure constraints: → curiosity without direction = noise → curiosity with systems = power most people wait to be taught curiosity maxxing forces you to discover that’s the difference between knowing and actually understanding

Ever wondered what are ALL the layers behind the answer when you get a response from ChatGPT.. Introducing Prompt to Chip 💬 Full visualization with explainers on how your prompt literally gets to the chip and all the way back. Play with it at prompt to chip dot com

One Million Industrial Vehicles Charged - engineering-update.co.uk/2026/03/19/one…

@rai_inst @nvidia Managing that rotational inertia during the flip requires precise torque vectoring. Similar principles to maintaining balance on high-speed descents. Excellent control architecture!

It was great to see our name amongst the other “AI Native” companies during @Nvidia’s #GTC keynote. NVIDIA Isaac™ Lab helps us train reinforcement learning policies that enable the UMV to drive, jump, flip, and hop like a pro!