| Few VCs are experts in machine learning or building deep learning frameworks, but most of them are pretty good with unit economics. Which is why they’re laser-focused on generative AI’s tech stack. Whether it’s infrastructure, middleware, applications or something else, investors are looking for founders who can dig defensible moats and dominate. According to Leonard Wossnig, CTO of biopharmaceutical startup LabGenius, “the true value proposition of AI companies now lies not just within the models, but also predominantly in the underpinning datasets.” Due to “a noticeable lack of substantial differentiation,” he says these models “are rapidly becoming commodities.” In this TC+ column, he presents questions that will help nontechnical VCs gauge a company’s “data quality . . . and what could go wrong if the data's not up to scratch,” along with frameworks that show how each layer in the stack creates value. Don’t let the headline fool you: If you work inside an early-stage AI startup, you need to know which angles of attack investors are likely to take when probing your pitch for flaws. Image Credits: E. Slomonson The Photo Group (opens in a new window) / Flickr (opens in a new window) under a CC BY 2.0 (opens in a new window) license. TechCrunch Disrupt 2023 ended yesterday, and out of all the events I’ve attended since working here, this one was my favorite. I moderated three panel discussions with investors, hosted a Q&A with TC+ columnist Sophie Alcorn, and I had the great pleasure of meeting and talking to scores of early-stage founders in the halls at Moscone Center. We’re all still catching our collective breath as my co-workers fly home to places like Pittsburgh, Paris and Providence, but keep an eye out next week for our recaps from Disrupt. We uncovered a ton of actionable business intelligence and had some fun along the way. Cheers, Walter Thompson
Editorial Manager, TechCrunch+
@yourprotagonist Read More | 
