langs

6 ms·

1.
by langs 11d ago
I don't get it. Why benchmark the latency instead of recall/precision? Optimizing for millisecond-level latency is meaningless in the context of LLM calls. Accuracy is the tool's greatest value, yet there is no testing for it
2.
Show HN: An Autonomous Agent for Slay the Spire(github.com)
1 points by langs 15d ago|0 comments
3.
by langs 1mo ago
Simply feeding the Slay the Spire game state directly to an LLM is currently insufficient to achieve stable deck-building or consistent combat performance. I categorize the game's actions into three types: combat, deck-building, and ot
4.
by langs 1mo ago
I'm building an AI agent that plays Slay the Spire. It currently reaches A20 Act 3 consistently and occasionally defeats the A20 Heart. It use two search: spare graph search for deck building, and MCTS for combat. Parallelly, I'm
5.
by langs 1mo ago
Yes, it uses modern technology to tackle an old problem: retrieval.
6.
by langs 1mo ago
https://github.com/AttemorySystem/attemory/ stars and issues are welcome :) Using attention for retrieval was inspired by a comment I saw in here long time ago: Prediction and retrieval are two sides of the same c
7.
by langs 1mo ago
I am working on the same thing right now. However, unlike storing conversations in an external retrieval system, I use a local LLM to store the conversation's KV cache and perform retrieval directly on that cache. The method involves r
8.
by langs 1mo ago
Pi’s minimalism reveals a simple truth: the LLM is the core of any agent harness. Consequently, much of current harness tuning will become redundant(or even a hindrance) with next-gen models.
9.
by langs 1mo ago
> Myth 2: Writing Code Is the Bottleneck Writing code is indeed the bottleneck for same resource constrained companies. Rapid code development creates more opportunities for trial and error, providing companies with more information for
10.
by langs 1mo ago
You need to optimize the KVCache part(save to disk to save compute) to achieve this goal.
11.
by langs 2mo ago
So, is this a competition? To see whose model can be jailbroken the most times and incite the highest level of public alarm?
12.
by langs 2mo ago
"mandatory safety testing" is an impractical ideal, it's not workable in real world. Like any technology, LLMs are dual-use tools capable of both beneficial and malicious applications—a fundamental reality that human intent
13.
by langs 2mo ago
the current bottlenecks for agents are actually bottlenecks of the LLMs, rather than the agents themselves.
14.
by langs 2mo ago
That’s a good analogy, but I think the reason we currently need the "nurse" role is the need to interact with the physical world. Most software products don't require this step, so the demand for “nurses” or "doctors&quo
15.
by langs 2mo ago
How does your solution differ from other KV Cache approaches, such as LMCache?
16.
by langs 2mo ago
nope. any rewrite will be an AI rewrite soon.
17.
by langs 5mo ago
I vide-coded some, but the core is hand-craft arch/algorithm, that's the most valuable part I want to protect.
18.
Ask HN: How do solo devs protect their work in the age of vibe coding?
56 points by langs 5mo ago|17 comments