XDEFiANCE'e Quality Internet Shop

Alpha experiment Lens Studio Exploration-first LessWrong lensing. Steerable axes, bridge posts, and a personal attribute profile. Designed to be easy to delete if it is not worth keeping. Open Lens Studio

Exploration-first LessWrong lensing. Steerable axes, bridge posts, and a personal attribute profile. Designed to be easy to delete if it is not worth keeping.

Paste this into Claude Code to start exploring immediately. For full functionality (higher limits + private vectors), create an account.

Use this prompt directly inside the Claude web app. No MCP, no installs: just allow access to our API once.

Claude Code and Codex are essentially AGI at this point—we recommend getting acquainted with these tools even if you are not a software developer. For maximum ergonomics (else you’ll be manually approving each time Claude tries to query our API), we think you can get away with claude –dangerously-skip-permissions, but that is your risk to accept. We would not recommend this with a model less smart than Opus 4.5. The risk even if you trust us is prompt injection attacks in one of our ingested entities, even though we generally scrape content from reputable sources.

Primitive Operations <=> is pgvector’s cosine distance operator. Smaller = more similar. 0 = identical. @ Stored Vectors Store concept embeddings server-side, reference by name. No 8KB vectors in your context. embedding <=> @mesa_opt± Vector Mixing Blend concepts algebraically. Add what you want, subtract what you don’t. scale(@rigor,.6) – scale(@hype,.3)Ø Centroids Average embeddings to capture an author’s essence or an era’s vibe. SELECT AVG(embedding) FROM …Δ Temporal Deltas Track intellectual drift. Where did a thinker move over time? (c(’25) <=> @idea) – (c(’22) <=> @idea)& BM25 Lexical Full-text search with fuzzy matching, phrase search, and BM25 scoring. alignment.search(‘corrigibility’)⊕ Hybrid Search Lexical candidates, semantic re-rank. Best of both worlds. WITH hits AS (search(…)) <=> @q

This gives Claude web permission to call our API from its sandbox.

<=> is pgvector’s cosine distance operator. Smaller = more similar. 0 = identical.

Store concept embeddings server-side, reference by name. No 8KB vectors in your context.

Blend concepts algebraically. Add what you want, subtract what you don’t.

Average embeddings to capture an author’s essence or an era’s vibe.

Track intellectual drift. Where did a thinker move over time?

Full-text search with fuzzy matching, phrase search, and BM25 scoring.

Alpha experiment Lens Studio Exploration-first LessWrong lensing. Steerable axes, bridge posts, and a personal attribute profile. Designed to be easy to delete if it is not worth keeping. Open Lens Studio

Alpha experiment Lens Studio Exploration-first LessWrong lensing. Steerable axes, bridge posts, and a personal attribute profile. Designed to be easy to delete if it is not worth keeping. Open Lens Studio

Alpha experiment Lens Studio Exploration-first LessWrong lensing. Steerable axes, bridge posts, and a personal attribute profile. Designed to be easy to delete if it is not worth keeping.

Claude Web (easiest setup, but less agentic) Use this prompt directly inside the Claude web app. No MCP, no installs: just allow access to our API once. Open Claude → Settings → Capabilities. Enable Code execution and file creation. Toggle Allow network egress. In Domain allowlist, add api.exopriors.com. Paste the prompt below and start querying in claude.ai. Domain allowlist in Claude settings. Click to expand. This gives Claude web permission to call our API from its sandbox.

Open Claude → Settings → Capabilities. Enable Code execution and file creation. Toggle Allow network egress. In Domain allowlist, add api.exopriors.com. Paste the prompt below and start querying in claude.ai. Domain allowlist in Claude settings. Click to expand.

Primitive Operations <=> is pgvector’s cosine distance operator. Smaller = more similar. 0 = identical.

@ Stored Vectors Store concept embeddings server-side, reference by name. No 8KB vectors in your context. embedding <=> @mesa_opt

± Vector Mixing Blend concepts algebraically. Add what you want, subtract what you don’t. scale(@rigor,.6) – scale(@hype,.3)

Ø Centroids Average embeddings to capture an author’s essence or an era’s vibe. SELECT AVG(embedding) FROM …

Δ Temporal Deltas Track intellectual drift. Where did a thinker move over time? (c(’25) <=> @idea) – (c(’22) <=> @idea)

& BM25 Lexical Full-text search with fuzzy matching, phrase search, and BM25 scoring. alignment.search(‘corrigibility’)

⊕ Hybrid Search Lexical candidates, semantic re-rank. Best of both worlds. WITH hits AS (search(…)) <=> @q

Claude prompt + public key Paste this into Claude Code to start exploring immediately. For full functionality (higher limits + private vectors), create an account. Claude Code and Codex are essentially AGI at this point—we recommend getting acquainted with these tools even if you are not a software developer. For maximum ergonomics (else you’ll be manually approving each time Claude tries to query our API), we think you can get away with claude –dangerously-skip-permissions, but that is your risk to accept. We would not recommend this with a model less smart than Opus 4.5. The risk even if you trust us is prompt injection attacks in one of our ingested entities, even though we generally scrape content from reputable sources. Claude Web (easiest setup, but less agentic) Use this prompt directly inside the Claude web app. No MCP, no installs: just allow access to our API once. Open Claude → Settings → Capabilities. Enable Code execution and file creation. Toggle Allow network egress. In Domain allowlist, add api.exopriors.com. Paste the prompt below and start querying in claude.ai. Domain allowlist in Claude settings. Click to expand. This gives Claude web permission to call our API from its sandbox. Copy Claude prompt

Claude prompt + public key Paste this into Claude Code to start exploring immediately. For full functionality (higher limits + private vectors), create an account. Claude Code and Codex are essentially AGI at this point—we recommend getting acquainted with these tools even if you are not a software developer. For maximum ergonomics (else you’ll be manually approving each time Claude tries to query our API), we think you can get away with claude –dangerously-skip-permissions, but that is your risk to accept. We would not recommend this with a model less smart than Opus 4.5. The risk even if you trust us is prompt injection attacks in one of our ingested entities, even though we generally scrape content from reputable sources. Claude Web (easiest setup, but less agentic) Use this prompt directly inside the Claude web app. No MCP, no installs: just allow access to our API once. Open Claude → Settings → Capabilities. Enable Code execution and file creation. Toggle Allow network egress. In Domain allowlist, add api.exopriors.com. Paste the prompt below and start querying in claude.ai. Domain allowlist in Claude settings. Click to expand. This gives Claude web permission to call our API from its sandbox.

@ Stored Vectors Store concept embeddings server-side, reference by name. No 8KB vectors in your context. embedding <=> @mesa_opt± Vector Mixing Blend concepts algebraically. Add what you want, subtract what you don’t. scale(@rigor,.6) – scale(@hype,.3)Ø Centroids Average embeddings to capture an author’s essence or an era’s vibe. SELECT AVG(embedding) FROM …Δ Temporal Deltas Track intellectual drift. Where did a thinker move over time? (c(’25) <=> @idea) – (c(’22) <=> @idea)& BM25 Lexical Full-text search with fuzzy matching, phrase search, and BM25 scoring. alignment.search(‘corrigibility’)⊕ Hybrid Search Lexical candidates, semantic re-rank. Best of both worlds. WITH hits AS (search(…)) <=> @q

Lexical candidates, semantic re-rank. Best of both worlds.

Your Question “What’s been written about mesa-optimization since 2023 that isn’t just rehashing Risks from Learned Optimization?” Claude composes the query Generated Query

“What’s been written about mesa-optimization since 2023 that isn’t just rehashing Risks from Learned Optimization?”