tifa2up
Re:Show HN: vLLM Studio – A macOS app for using vLLM models
on Saturday March 21, 2026 @04:07PM (#47470766)
Interesting project. Curious why Electrobun over Tauri here? Tauri has a much larger ecosystem and rust based for improved performance.
The Next Coding Interface Is a Canvas
Hackable Software
Re:Ask HN: What Are You Working On? (February 2026)
on Monday February 09, 2026 @07:11AM (#46944390)
https://agentset.ai/
Open-source RAG infrastructure.Every team I talk to has the same experience: RAG works in the demo, breaks in production.
We handle ingestion through retrieval with optimizations baked in. 97.9% on HotpotQA vs 88.8% for standard RAG. Model-agnostic, 22+ file types, built-in citations, MCP server. MIT licensed.
Do things like Oh My OpenCode work?
Show HN: EasyClaw – Seamless Installation for OpenClaw
Re:Ask HN: Could you share your personal website here?
on Thursday January 15, 2026 @04:45AM (#46630311)
Semantic Thesaurus
Vector DB Comparison
Always Do a Work Trial
Re:28M Hacker News comments as vector embedding search dataset
on Saturday November 29, 2025 @04:01AM (#46086119)
https://agentset.ai/leaderboard/embeddings good rundown of other open-source embedding models
Opus 4.5 is the best model for RAG
Gemini 3 vs. GPT 5.1 for RAG
Embedding Model Leaderboard
Re:Ask HN: What Are You Working On? (Nov 2025)
on Monday November 10, 2025 @08:32AM (#45875842)
Right now it's single shot, we're looking into building an "Agentic Retrieval" based on Claude ADK. tbd how it'll work
Re:Ask HN: What Are You Working On? (Nov 2025)
on Monday November 10, 2025 @02:09AM (#45873248)
I'm building https://github.com/agentset-ai/agentset, RAG as a service that works quite well out of the box.
We achieve this performance by baking in the best practices before any tweaking
Best Reranker for RAG: We tested the top models
Re:Developers are choosing older AI models, and the data explains why
on Wednesday November 05, 2025 @08:59AM (#45822880)
Think it varies by use case. It didn't do well with long context
Re:Developers are choosing older AI models, and the data explains why
on Wednesday November 05, 2025 @04:14AM (#45820980)
For large context (up to 100K tokens in some cases). We found that GPT-5:
a) has worse instruction following; doesn't follow the system prompt b) produces very long answers which resulted in a bad ux c) has 125K context window so extreme cases resulted in an error
Re:Developers are choosing older AI models, and the data explains why
on Wednesday November 05, 2025 @03:17AM (#45820651)
We tried GPT-5 for a RAG use case, and found that it performs worse than 4.1. We reverted and didn't look back.
Reranker Leaderboard
Re:Ask HN: How to boost Gemini transcription accuracy for company names?
on Sunday October 26, 2025 @01:41PM (#45713743)
Don't solve it on the STT level. Get the raw transcription from Gemini then pass the output to an LLM to fix company names and other modifications.
Happy to share more details if helpful.
Re:Hacker News Front Page: what 26 hours of traffic got us
on Thursday October 23, 2025 @01:42PM (#45684619)
Yes, we got 187 self-serve users (all on the free plan). And are in talks with an enterprise now.
Hacker News Front Page: what 26 hours of traffic got us
Re:Production RAG: what I learned from processing 5M+ documents
on Tuesday October 21, 2025 @03:53AM (#45653486)
You typically add a lot of metadata with each chunk text to be able to filter it, and do to include in the citations. Injecting metadata means that you see what metadata adds helpful context to the LLM, and when you pass the results to the LLM you pass them in a format like this:
Title: ... Author: ... Text: ...
for each chunk, instead of just passing the text
Re:Production RAG: what I learned from processing 5M+ documents
on Tuesday October 21, 2025 @03:51AM (#45653478)
Quite a decent hit. Local models don't perform very well in long contexts. We're planning to support a local-only offline set-up for people to host w/o additional dependencies
Re:Production RAG: what I learned from processing 5M+ documents
on Tuesday October 21, 2025 @03:49AM (#45653469)
OP. The way you improve it is move away from single shot semantic/keyword search and have an agentic system that can evaluate results and do follow-up queries.
Re:Production RAG: what I learned from processing 5M+ documents
on Tuesday October 21, 2025 @03:47AM (#45653457)
OP. We migrated to GPT-5 when it came out but found that it performs worse than 4.1 when you pass lots of context (up to 100K tokens in some cases). We found that it:
a) has worse instruction following; doesn't follow the system prompt b) produces very long answers which resulted in a bad ux c) has 125K context window so extreme cases resulted in an error
Again, these were only observed in RAG when you pass lots of chunks, GPT-5 is probably a better model for other taks.
Re:Production RAG: what I learned from processing 5M+ documents
on Monday October 20, 2025 @02:58PM (#45647756)
text similarity finds items that closely match. Reranking my select items that are less semantically "similar" but are more relevant to the query.