ATechGuy

6 ms·

31.
by ATechGuy 8mo ago
I'm not saying sandboxes are not needed, I'm saying VMs/containers already provide the core tech and it's easy to DIY a sandbox. Would love to understand what value E2B offers over VMs?
32.
by ATechGuy 8mo ago
I will ask what I've asked before: how to know what resources to make available to agents and what policies to enforce? The agent behavior is not predefined; it may need access to a number of files & web domains. For example, you s
33.
by ATechGuy 8mo ago
In the last one year, we have seen several sandboxing wrappers around containers/VMs and they all target one use case AI agent code execution. Why? perhaps because devs are good at building (wrappers around VMs) and chase the AI hype.
34.
by ATechGuy 8mo ago
> allowNet: ["api.openai.com", "*.anthropic.com"], How to know what domains to allow? The agent behavior is not predefined.
35.
by ATechGuy 8mo ago
What about VMs? They offer strong isolation, as they don't share kernels, and have long been a foundational piece for multi-tenant computing. Then, why would we put an extra layer on top and rebrand it as an AI agent sandboxing solutio
36.
by ATechGuy 8mo ago
Congrats on launching, and great testimonials! What problem does it solve compared to bazillion code execution sandboxing agents (and containers/VMs)? Overall, a lot of people are building their own code execution sandboxing agents aro
37.
by ATechGuy 8mo ago
> They fail because the environment is flaky: missing deps, slow setup, weird state, unclear feedback loops. Why can't agents install missing deps based on the error message?
38.
by ATechGuy 9mo ago
> There's no "OAuth for syscalls" yet. This exists today in OSes in form of discretionary/mandatory permissions (e.g., SELinux, AppArmor, LandLocked).
39.
Ask HN: Have your views about AI / LLMs changed? What triggered it?
4 points by ATechGuy 9mo ago|1 comments
40.
by ATechGuy 9mo ago
What was the first concrete thing you needed that existing sandboxing tools (Docker/VMs/bwrap) just didn't provide?
41.
by ATechGuy 9mo ago
> As long as there is a full OS running, you are one libc function away from a sandbox escape. Does this mean, all software in the world is just one function away from escape?
42.
Ask HN: Why are so many rolling out their own AI/LLM agent sandboxing solution?
32 points by ATechGuy 9mo ago|18 comments
43.
by ATechGuy 9mo ago
That's a fair criticism. My thought process was that the original comment was based on their personal experiences and since ChatGPT is trained on a large dataset, it may offer a different perspective derived from experiences of a lot m
44.
by ATechGuy 9mo ago
When scaling out, edge latency will overshadow kernel boot-up times: speeding up boot-up from 1.5s to 150ms will not have any perceived impact on app performance when scaling on edge to meet the demand.
45.
by ATechGuy 9mo ago
Fast boot up means nothing if your agent/app is slow at runtime (due to virtualization tax or QEMU emulation). Fast boot up is a PR term, which can easily be optimized for compared to designed a better virtualization layer that perform
46.
by ATechGuy 9mo ago
@nl But this uses nested virtualization on GCP, which severely hurts performance. Faster boot up means nothing if everything else is slow.
47.
by ATechGuy 9mo ago
Genuine question: why not just use GCP/AWS VMs for agentic execution? What is missing?
48.
by ATechGuy 10mo ago
Scrapers use residential IP proxies, so blocking based on IP addresses is not a solution.
49.
by ATechGuy 10mo ago
It is really free? Genuinely asking.
50.
by ATechGuy 10mo ago
From ChatGPT: This approach can stop very basic scripts, but the claim that “99.9999% of scrapers can’t execute JS or handle cookies” isn’t accurate anymore. Modern scraping tools commonly use headless browsers (Playwright, Puppeteer, Selen
51.
by ATechGuy 10mo ago
Genuine question: if traffic to websites is down and likely to continue the trend, what would be the value? Do you plan to address agentic browsing in the near future?
52.
by ATechGuy 10mo ago
Why not use KVM?
53.
by ATechGuy 11mo ago
Having used this for our research, I highly recommend Rtrvr.
54.
by ATechGuy 11mo ago
Maybe we should write an agent that writes an agent that writes an agent...
55.
by ATechGuy 1y ago
Then why only apply to VMs, why not apps?
56.
by ATechGuy 1y ago
Side-channel attacks apply to multi-tenant cloud environments, not local.
57.
by ATechGuy 1y ago
Why not use VMs from AWS (EC2) or GCP? E2b is built on top of GCP anyway.
58.
by ATechGuy 1y ago
Startups would build on big tech, so are likely to add their margins. Have you looked into (bulk) discounts from GCP/AWS?
59.
by ATechGuy 1y ago
Thanks! While I agree with you on "saving seconds" and overall latency argument, according to my understanding, most agentic use cases are asynchronous and VM boot up time may just be a tiny fraction of overall task execution time
60.
by ATechGuy 1y ago
> It does take 1-5 seconds to boot the environment (firecracker vms). I'd say 1-5 secs is fast. Curious to know what use cases require faster boot up, and today suffer from this latency?

More ›