pradeep1177

6 ms·

1.
by pradeep1177 16d ago
Big labs don’t expose logits, obviously. So I was wondering, if you wanted to play with distillation at scale, what other signals could you get from these models? Ha fun building this experiment: give Claude Code/Codex an enforced “wor
2.
Show HN: Give your agent somewhere to think loud watch its decisions unfold live(github.com)
1 points by pradeep1177 16d ago|1 comments
3.
Show HN: Use Spot Instance Fearlessly(github.com)
3 points by pradeep1177 20d ago|0 comments
4.
by pradeep1177 1mo ago
There is a big difference IMO. transcript shows a summary. The workbook shows you generated text, verbatim.
5.
Show HN: Agents Workbook watch Claude Code, Codex write down their working notes(github.com)
5 points by pradeep1177 1mo ago|3 comments
6.
by pradeep1177 1mo ago
These logs containing opaque blobs could accidentally contain secrets, the researchers decoded many of reasoning blocks from public repositories and reported finding PII and credentials. I was experimenting a bit how I could block these usi
7.
Show HN: AI agents gives its worst answers when you're overwhelmed(github.com)
2 points by pradeep1177 1mo ago|0 comments
8.
Show HN: Gitleaks for Codex and Claude Code(github.com)
2 points by pradeep1177 2mo ago|0 comments
9.
by pradeep1177 2mo ago
Interesting take. I also looked into this issue from a different angle https://github.com/softcane/hamza
10.
Show HN: Maybe AI agents shouldn't decide what's sensitive(github.com)
2 points by pradeep1177 2mo ago|0 comments
11.
Show HN: Hamza Mask secrets and PII before Claude Code or Codex sends them(github.com)
2 points by pradeep1177 2mo ago|0 comments
12.
Show HN: Giving Claude Code and Codex its voice(github.com)
2 points by pradeep1177 2mo ago|0 comments
13.
by pradeep1177 2mo ago
And what are those necessary controls?
14.
by pradeep1177 2mo ago
These data retention contracts are black boxes; you never know how your IP was leaked, and it could end up in the model's training data. It's like trusting META with your privacy settings.
15.
Ask HN: How companies are protecting Claude Code from reading IP and PII data
2 points by pradeep1177 2mo ago|6 comments
16.
by pradeep1177 2mo ago
I opened this feature request https://github.com/anthropics/claude-code/issues/68284 and hacked around a solution that I thought wasn't so hard to bake into the Claude code. There are other similar open
17.
by pradeep1177 2mo ago
If it’s a preference question then certainly yes, but how to push back on x productivity boost argument when a company enforce this as policy? Yes a soft policy but it’s a smart move to encourage the rat race.
18.
Show HN: How you auto recover your Claude Code workflow when quota resumes(github.com)
4 points by pradeep1177 2mo ago|1 comments
19.
by pradeep1177 2mo ago
Author here, I would happy to answer your question's
20.
by pradeep1177 2mo ago
Then why to even keep codex open source?
21.
Show HN: Giving Claude Code and codex its voice using kokoro(github.com)
2 points by pradeep1177 2mo ago|1 comments
22.
by pradeep1177 2mo ago
Many times those transcript contains sensitive credentials, so it’s redacted?
23.
by pradeep1177 2mo ago
Im working on a project to prevent Claude code from reading any secrets key
24.
by pradeep1177 3mo ago
But the whole point of this is to prevent the distillation and identify the list of blocked providers. If a provider is capturing the proxy, they can identify and modify that as well, so it only looks legitimate to the model. What am I miss
25.
by pradeep1177 3mo ago
I used my proxy https://github.com/softcane/cc-blackbox setup to capture this. This is how it looks. # userEmail The user's email address is <my email>. # currentDate Today's date is 2026-06-30.
26.
by pradeep1177 3mo ago
I generally believe the proxy route is best to understand any harness. I been building some thing similar.
27.
by pradeep1177 3mo ago
So, how are you handling read/write caching? I mean, if I keep routing the next prompt based on the task weights? How about if I'm sending every 5th query to opus, which do expensive write cache?
28.
by pradeep1177 3mo ago
I've been pondering over the same question, but also on an enterprise scale. For example, how could I scrub the customer data, like name, domain, etc., before sending it to LLM if my Claude code is using some MCP to fetch and debug an
29.
by pradeep1177 3mo ago
I think the expectation that AI will be the silver bullet for the domain where human behaviour change is just futile.
30.
by pradeep1177 3mo ago
Me: Thanks! Claude, you did a great job. Claude Code: Anytime. I'm glad we nailed it together. Happy to help. bla bla bla.... Me: Thanks! Codex, you did a great job. Codex: Understood. It finally clicked to me when Peter Steinberger me

More ›