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

    How can you tell whether an AI-assisted analysis supports its conclusion?

    A Vanishing Gradients panelist says the discussion covered programmatic checks, agents critiquing other agents and human judgment.

    GV
    HB
    2 Sources, ,

    TLDR

    A Vanishing Gradients panelist says Gaël Varoquaux has seen experienced data scientists get more done with agents, while people without that expertise think they’re getting more done. The panel discussed how to check agent-assisted analyses, including programmatic checks, agents critiquing other agents and human judgment. It also explored the possibility of business users questioning models and running simulations themselves.

    Combined views

    569

    2 Sources, first seen 3h ago

    Combined views

    569

    2 Sources, first seen 3h ago

    8 likes
    3h ago
    first seen 3h ago
    8 likes
    1 comments
    4 saves
    3 reposts
    1 comments
    4 saves
    3 reposts
    Featured Source

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    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

    2 Sources

    @hugobowne“I like to think that agents are like coffee. They can get you to do stupid things faster with more energy.” @GaelVaroquaux on the latest Vanishing Gradients 😂 I joined Gaël (@probabl_ai, @scikit_learn), Shipra Arora from Bain & Company, and @lfiaschi86 from @pymc_labs to talk about the state of agentic data science: what we can delegate, how we verify the results, and what we could do beyond speeding up our existing work. Gaël has seen experienced data scientists getting more done with agents, and people without that expertise thinking they’re getting more done. When your code runs and the accuracy looks beautiful, how do you check whether the analysis actually supports your conclusion? We get into programmatic checks, agents critiquing other agents, and where human judgment still matters. We also explore what becomes possible when business users can question models and run simulations themselves, and data teams can finally try the ideas sitting in their backlog. Here’s a taste of the conversation, including my attempt to get this very polite panel to argue with each other. Full episode in the replies 👇3h
    @GaelVaroquauxRT @hugobowne: “I like to think that agents are like coffee. They can get you to do stupid things faster with more energy.” @GaelVaroquaux…3h

    2 Sources

    @hugobowne“I like to think that agents are like coffee. They can get you to do stupid things faster with more energy.” @GaelVaroquaux on the latest Vanishing Gradients 😂 I joined Gaël (@probabl_ai, @scikit_learn), Shipra Arora from Bain & Company, and @lfiaschi86 from @pymc_labs to talk about the state of agentic data science: what we can delegate, how we verify the results, and what we could do beyond speeding up our existing work. Gaël has seen experienced data scientists getting more done with agents, and people without that expertise thinking they’re getting more done. When your code runs and the accuracy looks beautiful, how do you check whether the analysis actually supports your conclusion? We get into programmatic checks, agents critiquing other agents, and where human judgment still matters. We also explore what becomes possible when business users can question models and run simulations themselves, and data teams can finally try the ideas sitting in their backlog. Here’s a taste of the conversation, including my attempt to get this very polite panel to argue with each other. Full episode in the replies 👇3h
    @GaelVaroquauxRT @hugobowne: “I like to think that agents are like coffee. They can get you to do stupid things faster with more energy.” @GaelVaroquaux…3h