Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jflynt76
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
1.
▲
AI Agent Safety Is Still Thinking Like Permissions
(jeffreyflynt02.medium.com)
5 points
by
jflynt76
1mo ago
|
0 comments
2.
▲
by
jflynt76
3mo ago
So what's the move?
3.
▲
A benchmark revealing an average memory-retrieval accuracy of 9%
(zendoric.com)
5 points
by
jflynt76
3mo ago
|
0 comments
4.
▲
A Deterministic Replacement for LLM-as-Judge in Stateful Agent Evaluation
(arxiv.org)
4 points
by
jflynt76
3mo ago
|
0 comments
5.
▲
Multi-Agent Simulation Framework for Verifiable Synthetic Corporate Corpora
(arxiv.org)
4 points
by
jflynt76
3mo ago
|
0 comments
6.
▲
What happens when OKF runs inside an AI tool
(tenureai.dev)
5 points
by
jflynt76
3mo ago
|
0 comments
7.
▲
by
jflynt76
3mo ago
For what it's worth, I didn't describe it as anything; just posted the link. It's a paper with open code, no product behind it.
8.
▲
Why most AI evals would miss the Linear sales email failure
(tenureai.dev)
6 points
by
jflynt76
3mo ago
|
2 comments
9.
▲
Two AI judges scored our agent's answer 0.85, but it never opened the file
(tenureai.dev)
6 points
by
jflynt76
3mo ago
|
0 comments
10.
▲
AI Governance Cannot Be a Tool Call
(tenureai.dev)
5 points
by
jflynt76
3mo ago
|
0 comments
11.
▲
AI Memory Is Still Thinking Like Search
(medium.com)
5 points
by
jflynt76
3mo ago
|
0 comments
12.
▲
Show HN: Synthetic corporate dataset generator for AI agent evaluation
(github.com)
3 points
by
jflynt76
3mo ago
|
0 comments
13.
▲
Why MCP is the wrong abstraction for memory
(tenureai.dev)
4 points
by
jflynt76
3mo ago
|
0 comments
14.
▲
Show HN: Tenure – Traceable AI memory with configurable memory modes
(tenureai.dev)
3 points
by
jflynt76
3mo ago
|
0 comments
15.
▲
You Don't Need a GitHub Copilot Subscription to Use VS Code AI Features
(medium.com)
5 points
by
jflynt76
3mo ago
|
1 comments
16.
▲
AI Memory Proves Inefficient: Tenure Project Detects 95% Error Rate
(zamin.uz)
5 points
by
jflynt76
3mo ago
|
0 comments
17.
▲
by
jflynt76
3mo ago
I think it's because too many people have released tools that's clearly not ready for production because they don't know what to actually check. So it's now just easier to pattern match away any good tools that might sur