Ask HN: Why aren't local LLMs used as widely as we expected?

On paper, local LLMs seem like a perfect fit for privacy‑sensitive work: no data leaves the machine, no margin cost, and can access local data. Think law firms, financial agents, or companies where IT bans browser extensions and disallows cloud AI tools on work machines. Given that, I’d expect local models to be everywhere by now—yet they still feel niche.

I’m trying to understand what’s in the way. My hypotheses (and I’d love corrections):

1) People optimize for output quality over privacy. 2) Hardware is far behind. 3) The tool people truly want (e.g., “a trustworthy, local‑only browser extension”) have yet to emerge. 4) No one has informed your lawyer about this—for now. 5) Or: adoption is already happening, just not visible.

It’s possible many teams are quietly using Ollama in daily work, and we just don’t hear about it.