Sabitlenmiş Tweet
CodecFlow
500 posts
CodecFlow
@codecopenflow
Simulate, train, deploy, and fund robots from a single stack.
Katılım Ocak 2025
11 Takip Edilen7.4K Takipçiler
SimArena V1 is a big one. Getting into it on the stream, see you there.
peaq@peaq
Robots don't have to learn on the job anymore @codecopenflow is launching SimArena V1, where robots train in digital twins of real spaces like malls, warehouses, and restaurants, before they're deployed No engineering team needed. Pick an environment, drop in a robot, train, all in browser The team's joining peaq’s CBO @MartinElKhouri to break it down — don’t miss the stream 🗓 May 21, 1pm UTC 🔗 x.com/i/broadcasts/1…
English
Robots don't have to learn on the job anymore
@codecopenflow is launching SimArena V1, where robots train in digital twins of real spaces like malls, warehouses, and restaurants, before they're deployed
No engineering team needed. Pick an environment, drop in a robot, train, all in browser
The team's joining peaq’s CBO @MartinElKhouri to break it down — don’t miss the stream
🗓 May 21, 1pm UTC
🔗 x.com/i/broadcasts/1…
English
@lukas_m_ziegler launching simarena soon - browser tab to a working robot policy in under 1 hour
robotics is open to anyone now
English
@altantutar #5 is exactly what SimArena (incoming launch) exists for
Codec is the platform around it: build, run, deploy and fund. Goal is lowering the barrier so anyone can get into robotics.
Vercel for Robotics.
English
If I were starting a robotics company in 2026, I wouldn't
build another humanoid.
Here are the 7 wedges I'd explore instead:
1/ Actuators: China runs ~90% of magnet manufacturing and has a monopoly. Every humanoid needs ~40 motors. This will be the hardest robotics bottleneck of the decade.
2/ Ocean robotics: Saronic just raised $1.75B at a $9.25B valuation. Defense is the main buyer, but there are a lot of commercial buyers that people ignore.
3/ Space robotics: Launch costs are going to fall exponentially in the next decade. Orbit is going to become a labor market, and robots will do the work, not humans.
4/ System integrators: There's a lot of robots being built, but there is no one to go deploy them.
5/ Specialized VLAs: Actor Labs fine-tuned π0.5 on a real excavator. You could see that apply to different domains.
6/ Industrial inspection: Boring at first site, but probably one of the most overlook category. Buyers are big spenders, like oil and gas companies.
7/ Sim-to-real for unmanned missions: Almost all sim environments are created for ground & air physics. What if you build an ocean simulator?
What would you add?
English
You can now train @LGE_Global robots to work anywhere
peaq is proudly contributing to @codecopenflow's Simarena, a virtual training environment for robots
From hotel managers to restaurant owners and beyond, anyone can train their robots to operate in line with the layout and rules of the location, including any edge cases — no technical knowledge needed
→ Import your floor plan as a digital twin
→ Teach the robot the do’s and don’ts
→ Build a reinforcement learning harness that sticks
All that and more — without hardware bottlenecks or an engineering team involved
Launching in production soon
CodecFlow@codecopenflow
SimArena integrated @LGE_Global's CLOi with @peaq. You drop a robot into a digital twin of any venue, set its zones, tasks and boundaries, and train it with RL until it follows them. With the addition of peaqOS, task orchestration, fleet coordination and pay-per-skill payments all happen inside the sim. This unlocks the full commerce loop in sim, before a robot even interacts with its first real guest. peaq showed CLOi completing its first onchain transaction, and SimArena is where that workflow gets built and tested first.
English
SimArena integrated @LGE_Global's CLOi with @peaq.
You drop a robot into a digital twin of any venue, set its zones, tasks and boundaries, and train it with RL until it follows them. With the addition of peaqOS, task orchestration, fleet coordination and pay-per-skill payments all happen inside the sim.
This unlocks the full commerce loop in sim, before a robot even interacts with its first real guest.
peaq showed CLOi completing its first onchain transaction, and SimArena is where that workflow gets built and tested first.
English
CodecFlow retweetledi
Physical API is the most underweighted part of this talk. Most of robotics today is handcrafted per-robot glue and it doesn't compose into a fleet.
That's exactly the layer we're building.
Jim Fan@DrJimFan
I promise this will be the best 20 min you spend today! Robotics: Endgame, the sequel to my last year's Sequoia AI Ascent talk, "Physical Turing Test". I laid out the roadmap for solving Physical AGI as a simple parallel to the LLM success story. Be a good scientist, copy homework ;) And stay till the end, more easter eggs and predictions for your polymarket! 00:30 DGX-1 origin story at OpenAI, I was there in 2016 signing with Jensen and Elon. Heading to the Computer History Museum! 01:42 The Great Parallel 03:31 Robotics, the Endgame 03:39 Why VLAs fall short 04:32 Video world models as the 2nd pretraining paradigm 06:09 World Action Models (WAM) 07:46 Strategies for robot data collection and the FSD equivalent to physical data flywheel for robot manipulation 11:06 EgoScale and the Dexterity Scaling Law we discovered recently 14:00 Physical RL: bridging the last mile 15:39 DreamDojo: an end-to-end neural physics engine for scaling RL in silico 17:00 Civilizational Technology Tree and my predictions for the near future. Spoiler: it's closer than you think. Thanks to my friends at Sequoia for inviting me back to AI Ascent this year! I had a blast! Last year's talk is attached in the thread if you missed it.
English
