In AI, every approach has performed better when more scale is applied, especially as foundational problems are addressed over time.

When scaling innovations, it's important to balance between cooking up ideas and scaling them, as moving too fast can lead to failure.

Once you've found product market fit, scale aggressively to capture market opportunities before competitors do, as advantages always accrue to the number one player in a competitive market, allowing for faster product development and increased revenue.

The main goal of doing things that don't scale is to maximize learning. By engaging directly with customers, you can ensure that what you're building provides real value before scaling up.

The main goal of doing things that don't scale is to maximize learning and delight early customers, even if it requires manual and one-off efforts.

YC encourages doing things that don't scale, like Airbnb did by taking high-quality photos of homes to increase bookings. Don't wait for processes to be perfect before scaling; just get started in an unscalable way to unlock potential.

When deciding to scale rapidly, it's important to identify the right moments to accelerate growth, whether due to competition, market opportunities, or achieving critical mass.

To maximize improvement in AI, we need to scale existing techniques to their limits while also investing in future innovations that could lead to significant advancements.

You should not scale a company past 20 people before achieving product-market fit; once you find it, scale aggressively.