Simple products like DocuSign become massively complex at scale due to requirements for local data centers, country-specific standards (e.g., Japanese stamps), on-premise appliances for security, and compliance needs like FedRAMP. This complexity justifies a large engineering team.
For companies at the trillion-token scale, cost predictability is more important than the lowest per-token price. Superhuman favors providers offering fixed-capacity pricing, giving them better control over their cost structure, which is crucial for pre-IPO financial planning.
Superhuman's CTO credits a non-tech role managing submarine maintenance with teaching him to lead without technical legitimacy. By being forced to put his ego aside and drive change by asking fundamental questions, he learned to influence people far smarter in their domain.
AI coding assistants won't make fundamental skills obsolete. Instead, they act as a force multiplier that separates engineers. Great engineers use AI to become exceptional by augmenting their deep understanding, while mediocre engineers who rely on it blindly will fall further behind.
Observing that younger generations prefer consuming information via video (TikTok) and communicating via voice, Superhuman's CTO predicts a fundamental shift in user experience. Future interfaces, including email, will likely become more conversational and audio-based rather than relying on typing and reading.
Instead of generic benchmarks, Superhuman tests its AI models against specific problem "dimensions" like deep search and date comprehension. It uses "canonical queries," including extreme edge cases from its CEO, to ensure high quality on tasks that matter most to demanding users.
Superhuman adopted AI coding tools using a three-quarter plan: 1) Unrestricted experimentation with centralized budget approval. 2) Analysis and measurement using self-reported PR labels. 3) Observing a sustained increase in engineering throughput from 4 to 6 PRs per engineer per week.
Superhuman designs its AI to avoid "agent laziness," where the AI asks the user for clarification on simple tasks (e.g., "Which time slot do you prefer?"). A truly helpful agent should operate like a human executive assistant, making reasonable decisions autonomously to save the user time.