Jang-woo

4 ms·

1.
by Jang-woo 6mo ago
Boundaries sounds complex, but it's just one of your answers. If AI is going to control your device, you define the minimum it needs to know. Example: Caution: heat risk
2.
Show HN: Answer 9 questions, get a Matter firmware with safe boundaries(anna.software)
3 points by Jang-woo 6mo ago|1 comments
3.
by Jang-woo 6mo ago
That's a great connection. You're right that current systems aren't close to that level of reasoning. What I'm wondering is whether we can approximate some of it structurally — by defining when execution is allowed or no
4.
by Jang-woo 6mo ago
One thing we realized while building this: Commands don’t give control. Defining conditions does. Curious how others think about this — especially when AI starts interacting with physical systems.
5.
Show HN: Stop programming devices. Just answer questions(github.com)
4 points by Jang-woo 6mo ago|1 comments
6.
by Jang-woo 6mo ago
I think you're reading it more strongly than I intended. The point about "ownership" in that document is more about where authority over execution sits, not about restricting what AI is allowed to reason about. So it's n
7.
by Jang-woo 6mo ago
That's really interesting — thanks for sharing the notes. The "rebel agent" framing feels very close to what I'm trying to get at, especially the idea that refusal can be part of correct behavior rather than failure. One
8.
by Jang-woo 6mo ago
That's a good point. I think you're right that at the model level, competition pushes toward "always say yes." What I'm wondering about is whether control needs to exist at a different layer — not in the model itsel
9.
by Jang-woo 6mo ago
I've been thinking about AI systems acting in the physical world. Most discussions about control focus on what the system should do, and how to make execution reliable. But it seems like a lot of real-world failures aren't about i
10.
Should AI have the right to say 'No' to its owner?(github.com)
5 points by Jang-woo 6mo ago|33 comments
11.
We may not have a model problem – we may have an execution boundary problem(github.com)
1 points by Jang-woo 8mo ago|0 comments
12.
by Jang-woo 8mo ago
It’s encouraging to see the conversation shifting from model performance toward execution responsibility and security structure.
13.
by Jang-woo 8mo ago
Really appreciated this perspective. A lot of conversations focus on what models can become, but far fewer ask where judgment actually lives when these systems begin to act. The uncertainty you describe isn’t just philosophical — it’s a des
14.
by Jang-woo 8mo ago
Really interesting read. This feels less like a security bug and more like a missing execution boundary.
15.
by Jang-woo 8mo ago
AI is not removing entry-level roles — it’s exposing where judgment boundaries actually exist.
16.
by Jang-woo 8mo ago
The real question may not be whether AI serves society or shareholders, but whether we are designing clear execution boundaries that make responsibility explicit regardless of who owns the system.