About / Categories / Guides / Subscribe / Twitter

Venture Capitalist at Theory Ventures

What Level of AI?

Which level do I want to use AI?

I find myself asking this question more & more frequently & I think the answer means at work I’ll be using many AIs - not just one or two.

AI Level Use Case Description
Chat-Based AI Find the best Italian restaurant in the North Beach neighborhood of San Francisco.
In-App AI Find a document or generate an overview paragraph within Notion.
Browser-Based AI Deep research queries, such as estimating the market size of data center construction.
Computer-Based AI Transcribe a video call and upload the notes to an investment memo.
Multiple AI Agents Newer coding agents (e.g., Codex & Jules) work in parallel on the same codebase.

Why are there so many levels? It depends on the context I want the AI to have.

An individual chat is a Google query. I have a very specific question I want to answer. No other details necessary.

On the other hand, the in-app AIs are context specific. For example, I find myself using Gemini to find questions within my Gmail - ferreting a particular clause from our employee handbook.

When working with deep research queries, I want a browser swarm to scour the internet, finding chestnuts within warrens on the web. “Run a market size calculation on data center construction in the US and estimate how much software spend will be applied there.” A deep research agent might process more than 150 websites.

Sometimes I want to operate across multiple applications in my computer, which is why I want the operating system AI. Take notes from this video call, summarize it, send it to the CRM, and then add my tasks to my task manager.

The last one, the multiple coding agents, they operate in parallel and can work together. So it’s like managing a team and I can accomplish much more than I could managing an individual coding agent.

Like working with a person, an AI succeeds with clear direction & the relevant related information to the task. At some point, one AI may be able to switch between these layers easily (potentially through MCP).

I may also need to decide how fancy an AI I would like to use : a faster, less expensive model like Flash or Sonnet or Mini or a premium model like Pro or Opus or 4o.

Over time, these choices will simplify but for now, I need to decide.