Investing Secrets from Silicon Valley’s Best VCs

With AI & I on hiatus this month, we’re re-upping some of our best interviews. Today we're featuring a digest from three episodes in which Dan Shipper interviewed three of Silicon Valley's most influential venture capitalists—Sarah Tavel, Mike Maples, and Nabeel Hyatt. Their opinions vary, but they share a focus on founders who can see and capitalize on the paradigm shift that AI has brought about. Whether you're building, investing in, or simply curious about AI's trajectory, their insights prove essential.—Kate Lee

Was this newsletter forwarded to you? Sign up to get it in your inbox.

Sarah Tavel, Mike Maples, and Nabeel Hyatt have invested millions into the next generation of AI startups. Their bet isn’t on raw technology, though; it’s on products that feel magical, new business models, and solutions to wants people don’t even realize they have.

Over the first half of 2025, Dan Shipper interviewed the three on our podcast AI & I. Tavel is a venture partner at Benchmark; Maples is an iconic Silicon Valley investor who wrote early checks to Twitter, Twitch, and Lyft, and now invests through Floodgate, the fund he cofounded; and Hyatt is general partner of Spark Capital, who made early investments in Discord, video editor Descript, and, more recently, AI note-taking app Granola.

Read on for how Tavel, Maples, and Hyatt think about investing in AI. We cover:

1. Their big ideas for the next chapter of AI
2. The qualities of a remarkable founder
3. The kind of startups they’re drawn to

You can also check out the episodes in full here:

Tavel, Maples, and Hyatt on where we go next

As tech matures, the products get friendlier

Tavel believes the average person’s experience with AI could be far better. She gives a simple example: needing to manually add custom instructions in ChatGPT just to get it to ask clarifying questions instead of jumping straight to an answer. “It shouldn’t be this hard,” she says.

As the technology matures, Tavel expects richer product experiences to emerge. She points to Google as an early example: One of the internet’s first consumer breakthroughs, it was built by a deeply technical team. Over time, the next wave of products—Facebook, Instagram, Snap, Pinterest (where Tavel was an early employee)—came from teams that were less technical, but deeply product-savvy.

She sees the same shift happening in AI. As the technical layer solidifies, the real differentiation will come from founders who can make intuitive, delightful user experiences. OpenAI’s recent move to launch GPT-5 inside ChatGPT with an “auto-switcher” is a step in that direction: It handles the complexity of model selection behind the scenes, sparing users from needing to remember which confusingly named model does what best.