Open-interpreter: A natural language interface for computers

I finally got around to trying this out right now. Here&#x27;s how to run it using uvx (so you don&#x27;t need to install anything first):<p><pre><code> uvx --from open-interpreter interpreter </code></pre> I took the simplest route and pasted in an OpenAI API key, then I typed:<p><pre><code> find largest files on my desktop </code></pre> It generated a couple of chunks of Python, asked my permission to run them, ran them and gave me a good answer.<p>Here&#x27;s the transcript: <a href="https:&#x2F;&#x2F;gist.github.com&#x2F;simonw&#x2F;f78a2ebd2e06b821192ec919639959e6" rel="nofollow">https:&#x2F;&#x2F;gist.github.com&#x2F;simonw&#x2F;f78a2ebd2e06b821192ec91963995...</a>

I find the &quot;Can you ...&quot; phrasing used in this demo&#x2F;project fascinating. I would have expected the LLM to basically say &quot;Yes I can, would you like me to do it?&quot; to most of these questions, rather than directly and immediately executing the action.

It&#x27;s funny that we&#x27;re getting so much attention funneled towards the thought-to-machine I&#x2F;O problem now that LLMs are on the scene.<p>If the improvements are beneficial now, then surely they were beneficial before.<p>Prior to LLMs, though, we could have been making judicious use of simple algorithmic approaches to process natural language constructs as command language. We didn&#x27;t see a lot of interest in it.