Why Sequoia's AI Automation Strategy Misses the Mark for Most Companies

I think Sequoia got it wrong where to start automating with AI

Julien Bek from Sequoia argues you should start with "autopilot" AI tasks since they offer the fastest feedback loops. Sounds logical on paper, but reality is messier.

Even his "autopilot" examples still need 5-10% human oversight today. Meanwhile, "copilot" tasks might need 50-80% supervision. The lines aren't as clean as VCs like to draw them.

Here's what I've seen work better: Start where you're already outsourcing work. Your customer service team using ChatGPT to draft responses? That's your sweet spot - not because it's "autopilot," but because you already have the workflows and quality checks in place.

Smart mental models from VCs are great starting points, but don't let them box you in. The messy middle is often where the magic happens.

Where have you found the biggest gap between AI theory and practice in your company?