What practice best gives AI models the right context?

Most answers emphasize the importance of providing richer, more relevant context—whether through prompt engineering, data enrichment, or personal information—while differing on the specific methods used.

8 answers

Answers

  • Pamela Fox

    Principal Cloud Advocate (Python) · Microsoft

    Prompt Engineering

    Pamela Fox says prompt engineering is the practice that helps — showing that a little prompt engineering, like breaking a sentence into words and having the model show its reasoning, got a better, correct response.

    AI Engineer · Watch on YouTube

  • ali dasdan

    Visiting Research Scientist · New York University

    Maintaining Up-to-Date Work Context

    Ali Dasdan says his team tries to keep work context collected, up to date, and permissions honored, so the information they provide to AI models is relevant.

    MIT Sloan Management Review · Watch on YouTube

  • Kieran Klaassen

    GM of Cora | EIR · Every Inc.

    Teaching AI About You

    Kieran Klaassen says teaching AI more about you is what gives you an advantage, not teaching it more about writing, because models already know writing but not your beliefs, experiences or tastes.

    Kieran’s Substack - The AI Marketing Generalist · Kieran’s Substack - The AI Marketing Generalist · Read

  • Thariq Shihipar

    Member of Technical Staff · Anthropic

    Prompting

    Thariq Shihipar says prompting is valuable for giving models context, but says it is more than a prompt—it includes skills, data, and prior harness and verification work that build up into context.

    The Developing Dev · The Developing Dev · Read

  • Rob Key

    CEO · Converseon

    Advanced Enrichment

    Rob Key says advanced enrichment—structuring data for specific brand and industry definitions through NLP—reached 92% accuracy against the human gold standard, versus about 65% for native enrichment.

    CEO Series · CEO Series · Listen at 15:57

  • Dan Bjornn, PhD

    Senior Data Scientist · Lease End

    Providing Better Context

    Dan Bjornn, PhD, says providing better context was the practice: in coding tasks, he found they never needed to change the model, just the skill, resources, and context, because better context yields better results.

    AI Engineer · Watch on YouTube

  • Chris Piech

    Associate Professor, Computer Science (Teaching) · Stanford University

    Interview Prep

    Chris Piech answers that his practice is 'interview prep': they run a saved recipe on a pre-interview call transcript, which surfaces the story, threads, and questions so they sit down already knowing where it should go.

    EO · EO · Listen at 9:09

  • Akshay Nathan

    Engineering · Openai

    Feeding It More Context

    Akshay Nathan says the more information you give the model about whatever domain you're in, the more valuable it becomes, adding that it needs to have access to that context first.

    Latent Space · Watch on YouTube

More Data Science & Machine Learning questions

View all →