shad42 on Hacker News

1

How we know if our agent is right (opens in new tab)

(mendral.com)

2shad421h ago0

2

How we know if our agent is right (opens in new tab)

(mendral.com)

5shad426d ago0

3

The agent harness belongs outside the sandbox (opens in new tab)

(mendral.com)

180shad428d ago121

4

We decreased our LLM costs with Opus (opens in new tab)

(mendral.com)

106shad4212d ago31

5

Multi-player agents don't fit in the sandbox (opens in new tab)

(mendral.com)

2shad4215d ago0

6

We built our AI agent, for analyzing CI logs (opens in new tab)

(mendral.com)

1shad4225d ago0

7

Same LLM, different agent: a CI debugger built on Claude (opens in new tab)

(mendral.com)

2shad4226d ago0

8

Agent Harness: Inside vs. Outside the Sandbox (opens in new tab)

(mendral.com)

3shad4229d ago0

9

Same LLM but different output: we built a CI specialist (opens in new tab)

(mendral.com)

1shad421mo ago0

10

We upgraded our agent to Opus and our costs went down (opens in new tab)

(mendral.com)

2shad421mo ago0

11

Same LLM, Different Agent: What Changes When You Specialize for CI (opens in new tab)

(mendral.com)

3shad421mo ago0

12

We decreased our LLM costs by switching to Opus (opens in new tab)

(mendral.com)

2shad421mo ago0

13

What CI looks like at a 100-person team (PostHog) (opens in new tab)

(mendral.com)

56shad421mo ago30

14

We upgraded to a frontier model and our costs went down (opens in new tab)

(mendral.com)

1shad422mo ago0

shad42

Recent submissions

How we know if our agent is right (opens in new tab)

How we know if our agent is right (opens in new tab)

The agent harness belongs outside the sandbox (opens in new tab)

We decreased our LLM costs with Opus (opens in new tab)

Multi-player agents don't fit in the sandbox (opens in new tab)

We built our AI agent, for analyzing CI logs (opens in new tab)

Same LLM, different agent: a CI debugger built on Claude (opens in new tab)

Agent Harness: Inside vs. Outside the Sandbox (opens in new tab)

Same LLM but different output: we built a CI specialist (opens in new tab)

We upgraded our agent to Opus and our costs went down (opens in new tab)

Same LLM, Different Agent: What Changes When You Specialize for CI (opens in new tab)

We decreased our LLM costs by switching to Opus (opens in new tab)

What CI looks like at a 100-person team (PostHog) (opens in new tab)

We upgraded to a frontier model and our costs went down (opens in new tab)

Recent submissions

How we know if our agent is right (opens in new tab)

How we know if our agent is right (opens in new tab)

The agent harness belongs outside the sandbox (opens in new tab)

We decreased our LLM costs with Opus (opens in new tab)

Multi-player agents don't fit in the sandbox (opens in new tab)

We built our AI agent, for analyzing CI logs (opens in new tab)

Same LLM, different agent: a CI debugger built on Claude (opens in new tab)

Agent Harness: Inside vs. Outside the Sandbox (opens in new tab)

Same LLM but different output: we built a CI specialist (opens in new tab)

We upgraded our agent to Opus and our costs went down (opens in new tab)

Same LLM, Different Agent: What Changes When You Specialize for CI (opens in new tab)

We decreased our LLM costs by switching to Opus (opens in new tab)

What CI looks like at a 100-person team (PostHog) (opens in new tab)

We upgraded to a frontier model and our costs went down (opens in new tab)