frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Ask HN: Robotics engineers – how painful was setting up GPU sim infra?

5•nikhilol•6d ago
I'm exploring a service that auto-runs robot physics simulations on cloud GPUs when you push code, think Vercel CI/CD but for Isaac Sim, Gazebo, or MuJoCo. GPU spins up on push, runs headlessly, posts metrics to a hosted results dashboard, tears down. For those of you who've already solved the GPU infra problem: if you had to start over from scratch today with a new job, new team, new project, would you dread what lies ahead? Or was it a one-time annoyance you've forgotten about?

Genuinely trying to understand if this is painful enough to pay to avoid or just annoying for a few weeks and then fine.

(I have a landing page but no product yet, posting to validate before building!)

Comments

ghostnexus•5d ago
Former ML engineer here who ran IsaacGym and MuJoCo sims in the cloud for 2+ years. The pain is real and very specific:

1. Cold start latency killed iteration loops. Spinning up a GPU VM to test a 10-minute sim run took longer than the sim itself — you'd wait 3-5 min for the instance, run 8 min, tear down. That per-iteration overhead crushes exploration.

2. Idle billing. If you're grid-searching over reward functions, you want to fire 20 parallel runs, collect results, tune, repeat — but most providers bill per-hour so even a 12-minute run costs you a full hour.

3. Physics sim + CUDA dependencies. Custom CUDA kernels (warp sim, etc.) often need specific driver versions. Docker helps but image build/push overhead adds another 5-10 min to the loop.

The "CI for sims" framing (push code → run on GPU automatically) directly addresses #1 and #3. Worth building.

On the infrastructure layer: we built GhostNexus (https://ghostnexus.net) to address #1 and #2 — per-second billing, <30s cold starts on RTX 4090 hardware, Python SDK with 3 lines to submit a job. Might be worth using as the GPU backend if you don't want to manage the infra layer yourself. (Disclaimer: I'm the founder.)

nikhilol•5d ago
This is really interesting, thanks for this insight. Ghost nexus looks awesome, would love to chat some more!

Ask HN: What skills are future proof in an AI driven job market?

14•sunny678•2h ago•22 comments

Ask HN: How to solve the cold start problem for a two-sided marketplace?

136•alegd•18h ago•153 comments

Ask HN: How did you land your first projects as a solo engineer/consultant?

288•modelcroissant•1d ago•134 comments

Getting back into photography, ditching the phone camera in 2026?

4•eigenhombre•9h ago•5 comments

Ask HN: Is GitHub's #1 repo fake? build-your-own-x 491K stars

3•throwawaysoso•16h ago•1 comments

Ask HN: Building a solo business is impossible?

68•fnoef•4d ago•91 comments

Warning: If you search "Claude Desktop" on Google, the 2nd ad result is a scam

5•diogenescynic•18h ago•2 comments

Opus 4.7 vs. 4.6 after 3 days of real coding side by side from my actual session

8•agentseal•1d ago•6 comments

Recommended GPU Repairshop in Europe (Germany)

2•DogRunner•1d ago•0 comments

Ask HN: What makes a good Product Manager

8•chairhairair•1d ago•7 comments

Ask HN: May be a basic question, but how can I use AI well?

8•jdw64•1d ago•4 comments

Tell HN: 48 absurd web projects – one every month

78•absurdwebsite•4d ago•26 comments

Why don't we just ask AI to write assembler?

12•canterburry•3d ago•16 comments

Ask HN: Anyone know of that "levels of AI programming" blog post?

6•tuvix•2d ago•6 comments

Ask HN: How did you get your first users with zero audience?

16•arikusi•4d ago•18 comments

I built a real-time AR plane spotter, here's the math that makes it work

6•ananddhruv29•1d ago•4 comments

Ask HN: Getting depressed day by day, how to cope?

25•throwaw12•3d ago•25 comments

Ask HN: How do you maintain flow when vibe coding?

31•fny•4d ago•30 comments

Ask HN: Does magic link authentication use HTML canvassing?

5•trinsic2•2d ago•5 comments

Aliens.gov Resolves – To a WordPress "Site Not Found" Error

14•ascarola•4d ago•5 comments

Ask HN: ChatAi web-based session notation?

2•xtiansimon•2d ago•0 comments

Durable Object alarm loop: $34k in 8 days, zero users, no platform warning

31•thewillmoss•5d ago•3 comments

Ask HN: How are you using LLMs in production?

14•Anon84•4d ago•12 comments

Advice for tracking down a listening device?

11•comrade1234•4d ago•8 comments

Ask HN: Who is your favourite Entrepreneur/Visionary?

14•wasimsk•5d ago•34 comments

Do I Stop Learning Coding? DSA?

6•s_u_d_o•3d ago•15 comments

You've reached the end!