Investing in Pangram to Stop AI Slop on the Internet

The internet is currently facing a massive influx of automated content. Estimates show that nearly 35 percent of new websites consist of machine-generated text, with significant portions of academic papers and civil complaints now authored by AI. This trend creates a collapse in the cost of producing language, which in turn causes the cost of trusting that same language to skyrocket toward infinity. When users can no longer distinguish between human thought and automated output, the foundation of civic and commercial culture begins to erode.

To address this, Menlo Ventures has invested in Pangram, a firm dedicated to identifying automated content with high precision. While many detection tools fail to maintain accuracy against newer models, Pangram has demonstrated a false-positive rate of one in 10,000 in independent academic audits. Their approach relies on mechanistic interpretability, allowing the system to distinguish between text generated by different AI models without requiring constant retraining. Major platforms including the Internet Archive and Quora already rely on this technology to maintain content integrity.

Founders Max Spero and Bradley Emi developed this system out of a clear necessity to protect the digital ecosystem. Their refusal to build tools that help bypass detection underlines their commitment to restoring the trust contract of the internet. As AI usage expands into image and video generation, the mission to separate genuine human input from artificial synthesis becomes even more urgent.

This is not a matter of rejecting technological progress. It is about transparency. Readers deserve to know the origin of the information they consume. By providing the tools to verify authorship, Pangram is building a necessary layer of infrastructure for a functional information landscape. The company continues to scale its operations to meet the demand for verification across text and eventually visual media.