The use of advanced AI coding tools is causing the cost of finding software vulnerabilities to plummet.

Two years ago, the primary AI security concern for enterprises was Data Loss Prevention (DLP) for chatbots, focusing on what employees were inputting into tools like ChatGPT.

Sarah Guo's venture firm, Conviction, purchased H100 GPU nodes directly for its portfolio companies due to concerns about compute access for startups.

According to Sarah Guo, there has been increasing stress in the startup ecosystem over the past two quarters regarding access to compute supply.

The current AI market, with its rapid growth and leadership turnover, is more analogous to the volatile internet era of the late 1990s (e.g., AOL, Yahoo, Netscape) than the more stable SaaS era of the...

Some AI-native companies in Sarah Guo's portfolio with over $100 million in revenue operate with fewer than 50 engineers while employing close to 100 salespeople.

Managing human attention and ensuring code quality in an environment of abundant AI-generated code is a major unsolved problem, presenting a market opportunity for companies like Jira or new startups.

Sarah Guo states that a dominant thesis among investors is anxiety that AI-native application companies will struggle because most of the value will be captured by the foundational model layer.

Sarah Guo observes a theory that "vibe coding" and code generation will enable companies to create many of their niche software applications on-demand internally, rather than buying standardized produ...