How does the Pangram tool estimate if a text was generated by a machine?
Replied byChris Best
Co-founder & CEO at Substack
Niche: Technology
Revenue: Not Publicly Disclosed /month
Location: San Francisco, California, United States
Started: 2017
Pangram works by identifying specific statistical distributions and word choices that large language models naturally favor. While it can be bypassed by advanced prompt engineering, it provides a transparent signal that helps writers and readers discuss authorship.
0
From the Full Interview
This answer is part of a full interview with Chris Best, Co-founder & CEO at Substack.
Share this Answer
Found this insight valuable? Share it with your network to help others learn from Chris Best's experience.
Cite This Answer
Use this answer in your research, article, or academic work
Related Answers
How does Brex use the open-source Crab Trap tool to manage AI agent security?
By Garry Tan
Technology
Not Publicly Disclosed/mo
How does Substack's new Pangram integration help users navigate AI content?
By Chris Best
Technology
Not Publicly Disclosed /mo
How does Baumol's cost disease apply to writing when machine text becomes free?
By Chris Best
Technology
Not Publicly Disclosed /mo
How does generative artificial intelligence impact the volume of corporate noise and institutional credibility?
By Jim VandeHei
Technology
Approx. USD 8 Million/mo
How does Axios cultivate retention and engagement among its newsletter readership?
By Jim VandeHei
Technology
Approx. USD 8 Million/mo
What role does local journalism play in your broader long-term corporate vision for Axios Local?
By Jim VandeHei
Technology
Approx. USD 8 Million/mo
How does Microsoft balance its partnership with OpenAI against building internal AI models and capabilities?
By Satya Nadella
Technology
Approx. USD 20 Billion/mo