3 ms·
Ask HN: What's stopping you from running autonomous agents today?
I’ve been experimenting with autonomous agents and tried setting one up on a basic VPS yesterday. To my surprise, the entire stack was running in under 15 minutes.
It can already:
– browse
– execute tasks
– run tools
– automate workflows
– interact with APIs
What’s interesting is that this now works without GPUs or expensive infra.
However, the setup process is still rough:
– security
– API key management
– monitoring
– uptime
– scaling
– model switching
– cost tracking
I’m considering building a managed platform where:
• You bring your own VPS
• You own your data
• Setup + maintenance are automated
• Agents can run continuously
Before committing, I wanted to ask:
Would you actually use something like this?
If yes:
– What use cases?
– Biggest friction today?
– What would make this 10x more useful?
– Would you trust a third-party layer for this?
Trying to validate if this is a real need or just early excitement.
- tim-tday 8mo agohttps://www.mitre.org/sites/default/files/2026-02/PR-26-00176-1-MITRE-ATLAS-OpenClaw-Investigation.pdf https://www.mitre.org/sites/default/files/2026-02/PR-26-0017...
- zweeki 8mo agoThanks for this article however I am working with zeroclaw (openclaw alternative which is lightweight rust based cli tool). Here is the link to GitHub https://github.com/zeroclaw-labs/zeroclaw https://github.com/zeroclaw-labs/zeroclaw