Devicetocloud

Engineer | Full stack - device to cloud
Specialize in AI + Robotics + Devices
Building in Public

[Devicetocloud @devicetocloud]

Building infrastructure for Physical AI means I spend my days talking to robotics CTOs about what’s breaking.

Everyone is obsessed with the AI models, but that's rarely what actually kills a robotics company. The real bottlenecks look like this: 🧵(1/8)

(@NVIDIARobotics has been super helpful with their showcase where I actually found people to talk too)

[Devicetocloud @devicetocloud]

@IliaLarchenko @ieee_ras_icra Looking forward to your video! How much data/episodes you had to collect? Did you have to breakdown your policy for sequence of tasks or was it a single shot?

[Devicetocloud @devicetocloud]

@NVIDIAAI Woah, this is amazing! Wish it was running on dgx spart cluster

[Devicetocloud @devicetocloud]

@pbshgthm @PopVaxIndia @sohamsankaran @LeRobotHF Those are great policies to start with. Also great arm to work with!

[Devicetocloud @devicetocloud]

@Ferbin08 Sup!?

[Devicetocloud @devicetocloud]

@Sanjil Multi-modal dimensionality and a level of depth that you don't get elsewhere!
That's what got me to start my own company and go all in

[Devicetocloud @devicetocloud]

@epichrisis @NVIDIARobotics @nvidiaomniverse @NVIDIAAI Interesting and timely! Will take a look at it and share how it goes

[Devicetocloud @devicetocloud]

Trying to get my sim setup working similar to my physical setup using april tags using NVIDIA Isaac sim

Reason for doing this is to have a proper sim-eval. You just hear "use sim for eval or RL", but didnt find a good way to do scene matching which is important for VLA

[Devicetocloud @devicetocloud]

@nerdytanay Originally from Pune! Would love to collaborate with you.

[Devicetocloud @devicetocloud]

@nerdytanay smolVLA, ACT till now. Going to try molmoact2 and groot as well

[Devicetocloud @devicetocloud]

Some progress on my continous improvement loop for improving and managing robots!

Was able to connect my SO-101 and my mujoco simulated SO-101 to my console and start generating the data.

Data visualization, train & deploy coming up next!

[Devicetocloud @devicetocloud]

Anyone with an alternative for @loom? Since Altassian acquisition, this service has become terrible. It cant even handle 2 min long videos

[Devicetocloud @devicetocloud]

@LeRobotHF @rabault_nicolas This looks like a great tool to help folks get past the initial hiccups!

Now if you want to graduate from this with a different arm or power use-cases, you can look at what I am building
https://t.co/t7EKCkPwDx

[Devicetocloud @devicetocloud]

Some progress on my continous improvement loop for improving and managing robots!

Was able to connect my SO-101 and my mujoco simulated SO-101 to my console and start generating the data.

Data visualization, train & deploy coming up next!

[Devicetocloud @devicetocloud]

@knowShubhangi Super interesting, registered!

[Devicetocloud @devicetocloud]

@decapostos @NVIDIAAI @ASUSUSA I bought one from asus (definitely not cheap) https://t.co/rDtWKF88aV

What you will need to also do is, make sure the firmware on the spark is uptodate else it might not give you the IP or show slower speed

[Devicetocloud @devicetocloud]

Just bumped by local AI setup!

Welcoming Asus GX10, should go nicely with my DGX Spark! Now waiting for the QSFP cable to show up 😀

[Devicetocloud @devicetocloud]

Interesting choice of Android. Especially with Google moving towards a more closed android I wonder if we are going to see forks?

[Tom Warren @tomwarren]

will be interesting to see where Microsoft goes with Project Solara. It has tried and failed to build platforms for many devices in the past - Band, Cortana devices, Windows Phone, the list goes on. This time it's using a version of Android, anticipating future AI agent hardware

[Devicetocloud @devicetocloud]

This is what makes me more bullish on software engineering. Even after Opus 4.8, Mythos and what not, we are still in an era where debugging in production is required.
If a frontier lab has to undergo this, imagine what a tier 2/3 company has to face as challenges. And I am saying this with absolute respect for the scaling that Anthropic team has to do.

[ClaudeDevs @ClaudeDevs]

We've reset 5-hour and weekly rate limits for all users on Pro and Max plans.

We fixed an issue that caused some Claude Code sessions to spawn excessive parallel subagents, burning through usage faster than expected.

[Devicetocloud @devicetocloud]

@NileshArnaiya Tell me about your interests and what are you looking for!

[Devicetocloud @devicetocloud]

Started off via reddit to get folks educated on robotics https://t.co/TP2onGhRqW , this is where my platform also comes up with plans to make it freely available for students (see my last few tweets to see a demo of it). Essentially owning your data flywheel for robotics whether sim or physical

[Devicetocloud @devicetocloud]

So they took the one thing DGX Spark was bad at which is poor bandwidth and made it into a laptop! Man @NVIDIARTXSpark is going to bankrupt me :)

[Tom Warren @tomwarren]

Nvidia just officially announced its Windows on ARM chips. RTX Spark is “the most efficient PC chip ever built,” apparently. Nvidia has worked with Microsoft on RTX Spark laptops, with up to 128GB of unified memory for running large AI workloads locally https://t.co/JjDykUZwyg