First-hand: integrating embeddings myself, they cluster images by overall vibe
ciguleva · x · 2026-10-07
The author admits she's no embeddings expert—she'd worked on a team that built a comprehensive embedding system, but this is her first time integrating embeddings herself.
What fascinated her, even within the first few iterations, was how the model grouped images by their overall vibe, beyond just the specific objects in them. Midjourney founder David Holz replied asking whether she found it noticeably different from the embedding she used before; no detailed comparison followed in the thread.
More from Models
- 字节 Seed 论文实锤：KV 压缩相位盲区致 DeepSeek 长文能力周期性波动 — teortaxesTex · 2026-10-08
- Gary Marcus Slams OpenAI's Vague Math Proof Report: Zero Details, Won't Pass Peer Review — GaryMarcus · 2026-10-08
- Astra Designs Well On Its Own: Clean UI Without Bounding Box Hand-Holding — Pichonn · 2026-10-08
- Perplexity releases pplx-embed-v2-late: OCR-free late-interaction embeddings topping retrieval benchmarks — perplexity_ai · 2026-10-08
- Anonymous stealth LLM 'Space Bunny Alpha' tops OpenRouter with 22% usage share — maferase · 2026-10-08
- Check Point Breaks Decision Model Jev for About 50 Cents per Attack — evilsocket · 2026-10-08