Interview
Matt Cutts on the US Digital Service and Working at Google for 17 Years
Y CombinatorMatt Cutts, Craig Cannon, Michael Wang, Spencer Clark, Stephan Sturges, John Doherty, Vanman0254, Ronak Shah, Adam Hoffman, Raphael Ferreira, Tim Woods, Snehan Kekre
- Google's operational culture is expected to have evolved from a startup model of informal project delegation to a structured entity where employees no longer work weekends.
- Advertising strategies are predicted to shift from cost-per-mille (CPM) to cost-per-click models driven by self-service prototypes and A/B testing, which significantly improved click-through rates.
- Legal precedents establish that search results are protected by the First Amendment, allowing manual ranking adjustments, though future challenges regarding "evil unicorns" and misinformation may lead to common carrier classification.
- Handling deep fakes is expected to become increasingly difficult due to the ease of synthesizing new content, necessitating a hybrid approach of algorithmic detection, manual review, and established policies rather than full automation.
- The US Digital Service (USDS) operates on a two-to-four-year tour of duty model, expecting members to form civic tech startups or join state digital services after government service.
- USDS functions as a rapid-response "swat team" for failing government systems, with a team composition of one-third engineers, one-third designers, and one-third product managers supported by talent and procurement experts.
- Hiring processes at USDS are predicted to remain continuous with streamlined applications taking approximately 90 seconds, aiming to place even a single technical professional in departments lacking IT talent to improve outcomes.
- Government contracting is expected to evolve from paper-based evaluations to code-based grading by engineers, with bug bounties and vulnerability disclosure policies becoming standard practices similar to the private sector.
- Technology adoption in the government is estimated to be improving at a rate of three years per year, reducing the lag from 23 years behind to 13 years behind modern standards, despite bureaucratic hurdles.
- USDS activities are anticipated to focus on data authenticity issues like fake comments on agency sites, while policy regarding deep fakes is managed by the Office of the Federal CIO.
- Communication regarding algorithm changes is predicted to emphasize "good design" and future trends to partners to avoid leaking spam tactics, while legislators are encouraged to focus on outcomes rather than specific mechanisms to create evergreen laws.
- High-performing tech professionals are increasingly seeking mission-driven roles, with a significant portion considering government service as a long-term career goal.
- Trust in government is predicted to be heavily dependent on user interactions, with USDS aiming to create value by addressing issues at the "seams between silos" rather than within isolated departments.
- Tech professionals are expected to contribute to policy by participating in local legislative meetings or running for office, with opportunities available in agencies like the FTC to serve as technology coordinators.
- The decision to build in-house versus outsourcing software will largely be determined by federal partners, with USDS acting as a discovery team to evaluate off-the-shelf SaaS solutions.
- Society is expected to become more mindful of digital consumption, potentially swinging the pendulum toward "digital wellbeing" and away from intrusive user experiences.
- User feedback drives feature development in government agencies, such as Veterans Affairs, leading to products that "just work" to engender goodwill without relying on optional "whimsy."
- Technologists in government are expected to support encryption security standards, participating in policy processes while abiding by final decisions even if they personally disagree with backdoors.