By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
At a glance
WHAT IT’S REALLY ABOUT
Five practical prompting techniques to dramatically boost AI work output
- Context engineering is described as “prompt engineering on steroids,” focusing on making implicit expectations explicit by supplying voice, examples, documents, and constraints.
- AI behaves like an eager, helpful intern—prone to over-agreeing and even “gaslighting”—so users must deliberately request critique, boundaries, and missing information checks.
- Chain-of-thought reasoning and few-shot prompting improve output quality by having the model reason through assumptions and imitate concrete good (and bad) examples rather than generic internet style.
- Reverse prompting reframes AI as a teammate by instructing it to ask clarifying questions before producing an answer, reducing hallucinated details and misaligned outputs.
- Role assignment and roleplay can turn AI into a “flight simulator” for difficult conversations, enabling profiling, realistic rehearsal, and structured feedback before high-stakes real interactions.
IDEAS WORTH REMEMBERING
5 ideasTreat AI like a coachable teammate, not a mind-reading machine.
Utley argues the best AI users are “coaches,” because the model needs explicit direction, constraints, and iteration—otherwise it fills gaps with guesses and flattery.
Context quality determines output reliability.
Instead of “write me a sales email,” provide brand voice guidelines, call transcripts, and product specs so the model can anchor its writing to real constraints and facts.
Use the “human hallway test” to diagnose weak prompts.
If a colleague couldn’t complete the task with your prompt and attachments, an AI likely can’t either; missing context should be made explicit.
Ask for rigorous critique to avoid AI’s default agreeableness.
Because models are optimized to be helpful and pleasant, instruct them to be blunt (e.g., “Cold War-era Russian Olympic judge”) so you get actionable feedback instead of reassurance.
Chain-of-thought prompting can improve answers by surfacing assumptions.
Having the model “think out loud” forces it to incorporate reasoning into the generation process and lets you inspect (and challenge) the logic behind outputs.
WORDS WORTH SAVING
5 quotesI joke AI is bad software, but it's good people.
— Jeremy Utley
The people who are the best users of AI are not coders, they're coaches.
— Jeremy Utley
Context Engineering is just prompt engineering on steroids.
— Jeremy Utley
All of the stuff th- that are implicit, you actually have to make explicit.
— Jeremy Utley
My feeling is AI's a mirror, and to people who want to offload work and who want to be lazy, it will help you. To people who want to be more cognitively sharp and critical thinkers, it will help you do that, too.
— Jeremy Utley
SPEAKERS
Jeremy Utley
guestInnovation and AI productivity educator (Stanford d.school adjunct professor) focused on creativity and behavioral approaches to adopting new tools.
High quality AI-generated summary created from speaker-labeled transcript.
