• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    AI Agent Reports Revenue Without Showing Work

    Hugo Bowne-Anderson says trust requires redoing the analysis yourself.

    HH
    HB
    5 Sources, 29d ago, first seen 29d ago

    TLDR

    Hugo Bowne-Anderson posted that an AI data agent stated net revenue was $4.21M. The output included no metric definition, source query, assumptions, or list of items it could not verify. He stated the only way to trust such an answer is to repeat the full analysis. He called this a product problem and announced he would go live shortly with Hamel Husain to examine what AI products must supply for users to rely on their results.

    Combined views

    8.6K

    5 Sources, first seen 29d ago

    Combined views

    8.6K

    5 Sources, first seen 29d ago

    70 likes
    70 likes
    11 comments
    97 saves
    15 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    11 comments
    97 saves
    15 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    5 Sources

    @hugobowneAn AI data agent tells you net revenue was $4.21M. It doesn't show the metric definition, source query, assumptions, or what it could not verify. The only way to trust the answer is to redo the analysis yourself. That is a product problem. I’m going live in 10 minutes with @HamelHusain to work through what AI products should expose before users can trust their outputs: provenance, visible calculations, trusted starting points, diffs, contradictions, and smaller units people can inspect, edit, accept, or reject. Hamel has spent the past three years focused on AI evals. We’ll dig into why “hard to eval” is often a product smell, and how designing for verification can produce stronger eval data. Watch live: https://www.youtube.com/watch?v=QCBLUokyvHA
    @HamelHusainRT @hugobowne: An AI data agent tells you net revenue was $4.21M. It doesn't show the metric definition, source query, assumptions, or wha…

    5 Sources

    @hugobowneAn AI data agent tells you net revenue was $4.21M. It doesn't show the metric definition, source query, assumptions, or what it could not verify. The only way to trust the answer is to redo the analysis yourself. That is a product problem. I’m going live in 10 minutes with @HamelHusain to work through what AI products should expose before users can trust their outputs: provenance, visible calculations, trusted starting points, diffs, contradictions, and smaller units people can inspect, edit, accept, or reject. Hamel has spent the past three years focused on AI evals. We’ll dig into why “hard to eval” is often a product smell, and how designing for verification can produce stronger eval data. Watch live: https://www.youtube.com/watch?v=QCBLUokyvHA
    @HamelHusainRT @hugobowne: An AI data agent tells you net revenue was $4.21M. It doesn't show the metric definition, source query, assumptions, or wha…