mkagenius

5 ms·

61.
by mkagenius 3mo ago
Made the exact same thing when apple containers were released - https://github.com/instavm/coderunner#run-claude-code-inside...
62.
Self-host sandbox cloud and hypervisor for AI agents(github.com)
1 points by mkagenius 3mo ago|0 comments
63.
by mkagenius 4mo ago
> Free idea: I would like to be able to "jail" an agent inside a VM and send instructions to the harness from outside the VM to agent(s) installed inside. Ideally there is no Codex/Claude/etc. installed on the host. Y
64.
Show HN: Tarit – Self-host sandbox cloud and hypervisor for AI agents(github.com)
6 points by mkagenius 4mo ago|0 comments
65.
Burned out creating public software(mkagenius.substack.com)
4 points by mkagenius 4mo ago|1 comments
66.
Hey GLM 5.2, build me a hypervisor(technotes.substack.com)
6 points by mkagenius 4mo ago|0 comments
67.
How Claude Code and Codex approach sandboxing(instavm.io)
2 points by mkagenius 4mo ago|1 comments
68.
by mkagenius 4mo ago
One would. But then the results are even weirder as opus 4.6 scored more than opus 4.8 by a huge margin
69.
by mkagenius 4mo ago
It looks like the author is specifically avoiding model's name, because results are really weird. Opus 4.8/4.7 scored 28% Opus 4.6 score 37% So the author thought as let's not get into that just write Claude.
70.
by mkagenius 4mo ago
Token smuggler sounds like a profession coming soon. For distillation and stuff.
71.
by mkagenius 4mo ago
Have you tried https://instavm.io in US
72.
by mkagenius 4mo ago
Useful in splitting a big task - some parts are easy so give it to say Gemini. Some are harder so give it to gpt 5.5 and so on. Also the throughput kind of increases since providers are different.
73.
by mkagenius 4mo ago
We have created Murmur[1] which kind of works with your existing subscription (having API key is not mandatory). You can just tag @copilot @codex from claude code to delegate work to them. (it can also do it on its own too btw) 1. https:&#
74.
by mkagenius 4mo ago
Not so subtle plug for another sandbox provider, https://instavm.io : Apart from the above features. 1. We support more than 32GB disk (as a shareable device, ideal for agentic memory) 2. We provide egress control 3. W
75.
Taste and judgement are lies we tell ourselves(technotes.substack.com)
2 points by mkagenius 4mo ago|0 comments
76.
Taste and judgement are lies we tell ourselves(technotes.substack.com)
5 points by mkagenius 4mo ago|0 comments
77.
by mkagenius 4mo ago
AWS for AI agents - https://instavm.io Providing sandboxes through a CLI. Guardrails such as egress control and secret injection and audit trails built in. We can also be used as 3rd party sandboxes in Anthropic managed agent an
78.
Show HN: Manim-Studio – Prompt to 3blue1brown video(github.com)
3 points by mkagenius 4mo ago|0 comments
79.
by mkagenius 4mo ago
6 months a bummer, but we got it for apple sandbox - coderunner ( https://github.com/instavm/coderunner ) We got it yesterday, maybe they just started rolling it out and hence op posted this.
80.
by mkagenius 4mo ago
Yes it is funnily true but it was for fake news generation and not it's cyber capabilities. Another fun little gem of information, government has something called Mayhem > the autonomous Robo-Hacker AI called Mayhem that’s now in ch
81.
by mkagenius 4mo ago
Apple containers are great for providing a sandbox to your AI coding agents I have made it a MCP so that it's easily discoverable by all the coding agents https://github.com/instavm/coderunner
82.
by mkagenius 6mo ago
The article misses the point that you can have a third option - harness inside and outside both. The use cases will evolve so much more in coming weeks and months that the boundaries will blur. You can still contain the agent outside in a s
83.
by mkagenius 6mo ago
With instavm ( https://instavm.io ), you can provide an OCI image built from a dockerfile.
84.
by mkagenius 6mo ago
I also felt the need for a cloud based cron like automations, so decided to build it myself https://cronbox.sh with: 1. an ephemeral linux sandbox for each task 2. capability to fetch any url 3. can use tools like ffmpe
85.
by mkagenius 6mo ago
Not just sides, the vents are much sharper. Few years back, I tried to look on reddit for complaints regarding this - barely anything.
86.
by mkagenius 7mo ago
> The biggest problem is the coding agents don't "Fail fast and loud". They fail deceivingly. GPT 2 and 3 used to fail fast (and loud coz we could easily see it lying)
87.
Show HN: See How You Talk to Claude Code(prophetic-tacky-bison.instavm.site)
3 points by mkagenius 7mo ago|0 comments
88.
by mkagenius 7mo ago
Has to be taken in light of median income too right?
89.
by mkagenius 7mo ago
I had the exact same idea but for AI agent harnesses. I even created an app to explain it - https://news.ycombinator.com/item?id=47381803 (deleted the app as got no traction whatsoever) Idea was that, the ai models like op
90.
by mkagenius 7mo ago
I think you are absolutely right. (had to)

More ›