Pangram: we are "dangerously close" to the dead internet theory becoming reality

The available content corresponds to the show notes for an episode of TechCrunch's Equity podcast, not to a full article or a complete transcript. What follows is therefore a brief summary, strictly limited to the data appearing in the downloaded text; the specific nuances and arguments the guest develops in the audio are not available in writing and cannot be reproduced here without inventing them.

The episode centers on Pangram, a startup dedicated to detecting AI-generated content. According to the text, Pangram has recently raised $9 million for its AI detection system and has struck a partnership with Substack, a platform that already uses this technology to show readers which authors turn to AI to write their newsletters. The company has also launched a new tool for detecting AI-generated images.

The guest is Max Spero, co-founder and chief executive of Pangram, interviewed by journalist Rebecca Bellan. The text previews several topics addressed in the conversation, though without developing them: why Spero believes the internet could be "dangerously close" to the so-called dead internet theory becoming reality in the coming years; why measuring how much AI there is in a text may be harder, but also more useful, than simply labeling it "AI" or "human"; the risks of AI detection failing, particularly when false positives affect sensitive images; and the idea that the "bottom rung" of writing jobs may have disappeared permanently, while genuinely good human writing could gain value.

Taken together, the piece points to a broader trend that several startups are trying to capitalize on: becoming a kind of "trust layer" for an internet where it is increasingly difficult to distinguish human content from machine-generated content, whether in texts, job applications, product reviews or even insurance claims. Pangram presents itself as one of those players, with recent funding and concrete commercial partnerships (Substack) backing its pitch.

Since the source is only the episode's promotional show notes and not its transcript, it is not possible to offer here the figures, examples or detailed reasoning that Spero shares on the dead internet theory, the technical workings of graduated AI detection, specific cases of false positives or the future of the writing job market. Anyone wanting to explore those points should turn to the original Equity podcast audio.