Log in or create a free Rosenverse account to watch this video.
Log in Create free account100s of community videos are available to free members. Conference talks are generally available to Gold members.
Hands-on AI #1: Let’s write your first AI eval
This video is featured in the Evals + Claude playlist.
Summary
In this talk, Peter Van Dijck of the helpful intelligence company will walk you through writing your first eval. You will learn the basic concepts and the tools, and write an eval together. This talk is hands on; you can follow along, and there will be plenty of time for questions.
You will go away with an understanding of the basic building blocks of AI evals, and with the confidence that you know how to write one. And more importantly, you’ll build some intuition, some product sense, around how the best AI products today are built, and how that can help you use them more effectively yourself.
Key Insights
-
•
Evals consist of a task, a golden dataset with known correct outputs, and an evaluator that measures correctness.
-
•
Manual AI prompt testing is slow and inconsistent; automated evals accelerate and scale evaluation.
-
•
UX and product teams can and should learn evals as a practical, non-technical skill.
-
•
Creating your own golden dataset is essential and cannot be outsourced or fully automated.
-
•
Models are fixed once trained; improvements happen by refining prompts and context design, not retraining the model.
-
•
Evaluations measure task performance, not the underlying model itself, allowing comparison across models.
-
•
Outputting a confidence score from models is unreliable due to lack of internal memory and inconsistent scale interpretation.
-
•
Biases are baked into models during training via evals used in post-training refinement.
-
•
LLMs can be used to judge other LLM outputs to evaluate tasks with non-binary answers.
-
•
Effective eval work requires collaboration across data analysts, engineers, subject matter experts, and UX/product teams.
Notable Quotes
"Evals are like a way to define what good looks like."
"The model was baked and once it’s baked, it does not learn again until they bake a new one."
"You need to be looking at the data. Nobody wants to, but that’s core work."
"Without a golden dataset, you have to build the golden dataset yourself."
"We’re not teaching the model anything; we’re improving our prompts and context."
"Confidence scores from the model are not a good idea because the model has no memory."
"Biases are baked in through the evals used during model training and post-training."
"LLMs judging other LLMs might sound crazy, but if you do it right, it works."
"Evals are a product and UX skill; learning them lets you make these systems do what you want."
"There is a large and growing capability overhang in these models we haven’t discovered yet."
Or choose a question:
More Videos
"The Red Queen effect means you have no choice but to adapt as others evolve faster."
Simon WardleyMaps and Topographical Intelligence
January 31, 2019
"We need to transform our role as connectors between people, disciplines, and ideas that together make our work richer and more impactful."
Megan BlockerTheme 2 Intro
March 12, 2025
"Kids are not little adults. You can’t just simplify the text and call it for kids."
Mila Kuznetsova Lucy DentonHow Lessons Learned from Our Youngest Users Can Help Us Evolve our Practices
March 9, 2022
"We have to step outside research and understand AI, technology, and product processes to put ourselves in the right space."
Megan Blocker Lada Gorlenko Fatimah Richmond Molly StevensWhat UX research maturity looks like and how we get there [Advancing Research Community Workshop Series]
November 9, 2023
"Design thinking is accessible and teachable quickly, but mastering design as a craft is a multi-year journey."
Wendy Johansson Surya VankaDesign at Scale: Behind the Scenes
April 29, 2021
"You don’t design interfaces anymore—you choreograph intentions."
David SternbergUncovering the hidden forces shaping user behavior
July 17, 2025
"Convenience includes accepting multiple payment providers like Apple Pay or PayPal, which can mean the difference between completing and abandoning a purchase."
Sam ProulxOnline Shopping: Designing an Accessible Experience
November 29, 2023
"Patient opportunism means most of the time having the current at your back and only occasionally angling directly against it to avoid burnout."
John CutlerOxbows, Rivers, and Estuaries: How to navigate the currents of change (without burning out)
December 3, 2024
"Sacred civics requires us to hold higher order accountabilities to ecosystems and future generations."
Jayne Engle Tanya Chung-Tiam-FookCivic Design for the Next Seven Generations—A Discussion on Sacred Civics
August 25, 2022