Why would somebody ever use a hosted version? I see "One-click gerber, DXF/STEP, render and BOM export" and "Altium support beyond KiCad" in cloud plans, dunno if they are compelling.

Tangential, do LLMs generate D2 code as easily as Mermaid and do tools like Github and Confluence render them?

>do LLMs generate D2 code as easily as Mermaid

Yes

>and do tools like Github and Confluence render them?

No

Is there any use case/demo of lighter models like using embeddings for full text search, STT, OCR etc?

Good ol days. I bought the book about the project's failure in 2012. They mentioned Python-Wx wrapper they created and I used it for a Python project.

Which language and agent? I used Aider on a Python project which had string references (as opposed to module tree) and cross-repo references. LSP lookups were fast and I hated it when Aider used 2M tokens a day, but grep was more complete.

Any, any. I've got something a bit more cursed and complicated going on. Need to clean it up, just ask your LLM to explain it https://github.com/doublemover/PairOfCleats

Is there a big distinction between "Build from scratch (Claude Code, Codex)" and "Vibe-code (Lovable, v0)", if an agent can set up hosting, database, deploy etc?

None. Vibecoding from scratch.

What is Lenovo doing here? Were they bundling Dropbox with their devices?

It used to be the case with Amazon India. Now my Gmail shows my fish food will arrive in 2 days.

How is pgvector with Sentence Transformers, a CPU-only embedding model, compared to a model hosted by OpenAI?