Our editorial approach to AI startup coverage has evolved since we started this section, and we think the specific changes are worth stating rather than leaving implicit.
We raised our bar for what counts as a verifiable traction signal
Early coverage sometimes accepted a startup's self-reported customer or usage numbers with less scrutiny than we're now comfortable with. We now weight independently verifiable signals more heavily than self-reported figures we can't confirm.
We stopped treating valuation as a primary story hook
We used to lead startup coverage with valuation figures fairly often. We've shifted toward leading with what the company actually does and what evidence supports its traction, treating valuation as one data point among several rather than the headline.
We got more disciplined about follow-up coverage
We didn't consistently revisit startups we'd covered early to report on how they actually performed later. We now try to do this more systematically, since early-stage coverage necessarily involves some uncertainty that follow-up reporting can resolve.
Why this evolution is worth sharing directly
Startup coverage is particularly susceptible to uncritical amplification given how much startups depend on press coverage for credibility, and we think being explicit about how we've tightened our own standards is more useful than simply claiming rigor.