Harvey, an early OpenAI investment, is now creating proprietary models on open-source foundations. This signals a major trend where vertical AI companies are protecting their valuable training data and avoiding dependency on platforms that could become competitors.
AI companies subsidize compute costs to attract users, much like Uber and Lyft subsidized rides. This creates impressive top-line growth but hides the true cost-to-serve and user churn. The industry faces a reckoning when capital dries up and profitability becomes paramount.
Instead of asking an AI for a list of potential podcast guests (a goal), build a persistent agent that researches, identifies, and shares new prospects every 48 hours (a system). This shifts from manual, periodic tasks to automated, continuous value creation.
Frontier models from giants like OpenAI force enterprises to share sensitive data, creating platform risk. The future of corporate AI lies in private, fine-tuned, open-source models that keep a company's "intelligence" in-house, preventing it from training potential competitors.
During a seed fundraise, a student founder can request to sell a small portion of their common stock (e.g., $100k) to pay down student debt. This is pitched to investors not as cashing out, but as removing a major personal distraction, thereby increasing focus on the company.
A small startup (like Willow) gains credibility by framing itself as an alternative to an established player (Whisper). Conversely, the established player legitimizes the challenger by engaging. This asymmetric dynamic is a core rule of competitive positioning.
The ability for an undergraduate to create a working autonomous vehicle with off-the-shelf parts suggests the barrier to entry is collapsing. This could lead to a highly fragmented market with many niche players, rather than a few winners like Waymo or Tesla.
Willow's founder anticipates basic dictation will become a free feature on major platforms. Instead of being disrupted, he is accelerating that trend by offering a superior free product to build a user base he can upsell on more advanced, defensible features.
AI is enabling academic triage where students automate work for courses they see as obligations, allowing them to engage deeply with material that interests them. This behavior mirrors how professionals use automation to focus on high-value work.