Making the entire world’s data AI Native with @wondersearchai
Singapore
Joined November 2025
- Tweets422
- Following37
- Followers368
- Likes376
evokoa (YCF26) retweeted
80k is insane
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
evokoa (YCF26) retweeted
Search has been guessing since 1940.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today we're launching @wondersearchai: search everything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
No more guessing. Just search.
evokoa (YCF26) retweeted
Thank you for your support! This hit 62k 🫡
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
The Y Combinator W27 deadline is on Nov 2
If you're applying, we would love to help, comment what you're working on below. We'll share our yc application, help you review yours, and we'll personally refer the top ones.
Please comment here and dale will reach out!
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
evokoa (YCF26) retweeted
The Y Combinator W27 deadline is on Nov 2
If you're applying, I would love to help, comment what you're working on below. I'll share my yc application, help you review yours, and i'll personally refer the top ones.
Please comment here rather than dm me. I'll reach out to you instead.
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
evokoa (YCF26) retweeted
insane, this blew up like crazy
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
evokoa (YCF26) retweeted
RIP vector databases
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
evokoa (YCF26) retweeted
Today we’re launching WonderSearch (YC F26)
Search millions of unstructured documents without pre-embedding the corpus. With comparable/better retrieval recall than embedding-based search.
WonderSearch searches at query time:
> No corpus-wide embedding step.
> No vector index to maintain.
> More compute only when needed.
> Pay per search intead of embeddings you may never use.
Go from raw data to useful answers in seconds.
Link + FAQ below (+ free launch credits!) ↓
evokoa (YCF26) retweeted
We got zero-embedding semantic search before GTA 6
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
In two years nobody will admit they ever built a RAG pipeline
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
evokoa (YCF26) retweeted
Embeddings were a mistake.
We matched OpenAI’s best embedding model on FiQA without using a single embedding. No vector database. No chunking. No re-indexing.
Today, we're launching @wondersearchai searcheverything your company ever wrote and find the exact line.
@ycombinator @fdotinc @daltonmeon
Lfggg
You don’t need a vector database.
Introducing WonderSearch (YC F26): Semantic search with no embeddings. Instantly search through all of your data.
Plug it into your app or your AI agent through our API or MCP.
+ Secret welcome email surprise @wondersearchai
evokoa (YCF26) retweeted
You don’t need a vector database.
Introducing WonderSearch (YC F26): Semantic search with no embeddings. Instantly search through all of your data.
Plug it into your app or your AI agent through our API or MCP.
+ Secret welcome email surprise @wondersearchai
evokoa (YCF26) retweeted
Everyone said I was full of shit back then. This is only 6 months ago.
evokoa (YCF26) retweeted
You come from a long line of searchers.
We built the search. You bring the wonder.
1,073,741,824 ancestors. Every one of them wondered something.
Now you can just search it: wondersearch.com