The concept of building holds mythic status in the world of new venture creation, and the mythos often obscures the difficulty. In construction, there are no such illusions. Construction is a $2 trillion U.S./$13 trillion global industry where automation has barely scratched the surface of possibility. Beneath that surface is dirt—unfathomable amounts. For new housing subdivisions, commercial real estate developments, data centers to power the AI boom, highway expansions, and countless other projects, American abundance begins with dirt.
Ask anyone who’s ever begun a term paper the night before it’s due: there are stark tradeoffs between research time and research quality. This is especially true of primary user research, one of the fundamental ways large enterprises can ensure their products resonate with customers. Historically, this has required a choice between user interviews, which are high-fidelity but slow, manual, and expensive, and surveys, which are scalable, cheaper, and faster, but often result in unrefined, low-signal data. Although there have been some strong companies on the survey side, e.g. Qualtrics and Medallia, this uneasy compromise has always persisted—until now.
We were thrilled to feature Charles Srisuwananukorn from Together AI at January’s Chat8VC. Charles is the Founding Vice President of Engineering at Together AI, where he leads the company’s work on AI infrastructure and clusters. Previously, he was Head of Applied Machine Learning at Snorkel AI and held engineering roles at Apple. He studied Computer Science at Stanford and has helped steer Together from an early contributor to open-source AI to a full-stack infra platform.
We were thrilled to feature Joe and Jonathan from Upwork at March's Chat8VC in San Francisco. We covered their journey from teams like Google Brain and Cruise, and their own startup, to leading AI efforts at Upwork—building Uma, a suite of specialized LLMs powering workflows for freelancers and clients across the platform.
Human-quality workflows need human-quality data, an axiom that has only grown truer in the AI-first enterprise. However, access to complete, high-signal data remains a limiting factor, given steep data provider fees, inflexible schemas, AI hallucinations, and scattered, inconsistent, and mutating sources. Customers don’t need to be data scientists to recognize shovel-ready datasets, but if they need to be data scientists to generate them reliably, data will always be rate-limiting.
It’s widely agreed that GenAI will transform software development, and GenAI dev tools have emerged as cornerstones of 8VC’s portfolio and broader AI productivity thesis. Up to now, however, hard data on the scale and specifics of this shift have been missing from the equation. In competitive industries, the speed and efficiency gains promised by GenAI coding tools could well mean the difference between market leadership and obsolescence. Companies can’t afford to select the wrong tools and end up on the wrong side of the AI adoption curve.