021 AI for Operators

Bryce Murray, Director of Technology at Permanent Equity, 9 links

Author

Tom Guthrie
October 29, 2025

Hi there, 

Welcome back to AI for Operators. Here’s what we’ve got for you this week:

  • The Operator: Bryce Murray from Permanent Equity (listen here)

  • The Essay: The AI Backlash is Just Beginning

  • The Links: 9 links including AI data center backlash and job cuts, a 20-agent playbook, and an AI Minister with 83 children…?

The Operator

This week’s episode is with Bryce Murray, Director of Technology at Permanent Equity

Bryce Murray, Director of Technology at Permanent Equity, shares how he applies AI pragmatism from defense research to portfolio use cases and daily life.

  • His AI roots in building programs to detect explosives at scale.

  • The pros and cons of an academic mindset: convert narrow academic theory into usable products, accepting messy data, dealing with legacy systems, and working around real-world constraints.

  • Basics of “Prompt Theory” to achieve better results: set a goal, invoke emotion, weigh benefits and risks, enforce facts, then brainstorm in separate passes.

  • Playing on AI’s “emotions” - a key to more accurate results?

  • How to use evals effectively: the shift to favoring UX-style metrics (resolution, satisfaction, engagement) over abstract accuracy when use cases go beyond coding and customer service.

  • The key to adoption (especially in tech-skeptical companies): start with safety and clarity, then demystify models, set permissions, and normalize usage at the edges before chasing “low-hanging” ROI projects.

  • How process documentation becomes leverage: well-defined workflows become prompts; documenting them often doubles output even before automation.

  • Bryce’s personal systems: long-form meal-planning prompt (with shopping list and built-in randomization!), persistent work thread that auto-generates weekly stakeholder updates.

The Essay: The AI Backlash is Just Beginning

With AI data centers facing local opposition and headlines about current and future job loss mounting by the day (yes, the headlines that Amazon is replacing 600,000 jobs with AI are misleading, but since when are we as a society good at news literacy?) , it’s safe to say that we’re only in the first inning of the anti-AI backlash.

What does that mean for operators?

Well, for some operators, it might mean that they’re going to lose their jobs. Amazon also announced a corporate headcount reduction of 14,000 people this year, with more on their way next year. Whether companies are seeing actual efficiency gains or not (some are, some aren’t), executives are always looking for ways to boost profits, and this is a believable excuse that signals to the market that the company is AI-savvy.

If you’re among the laid off, it’s going to be a tough slog. Adding tens of thousands of newly-fired folks to an already-tough job market will make competition that much more difficult, and the introduction of AI to hiring processes has only made things less personal. Still, your best bet is to hone your skills, show that you know how to take advantage of AI tools to multiply your impact, and build relationships broadly and deeply so that you have a shot at jobs that aren’t publicly advertised.

If you still have a job, you should grab the AI bull by the horns and make yourself indispensable to your organization. Be the company expert on AI. Identify improvements - both cost reductions and ways to increase revenue - and build real plans around them. Make sure that you’re putting yourself in a position to drive the most meaningful metrics and priorities for your business.

It’s also worth thinking about how you’re going to navigate messaging to your team and your customers if (when) anti-AI pressure increases. Are you going to lean into the “innovative, AI-first” messaging and risk bigger backlash? Position yourself as human-first and try to take care of everyone? Chart a middle course?

In this case, I think transparency is - more or less - the best policy. If your company’s leadership is all in on AI, say so, but don’t leave people behind. Evangelize around your AI organizational transformation plan and give people ample opportunities to use the tools and show how valuable they are. Make clear that expectations are increasing and competition is all around you.

You will have some people that opt out - that’s fine. The people that you need to stay are the ones who are curious, proactive, and dedicated. I assume that you’re one of those people. Create the conditions such that you can clearly identify which of your teammates are, too.

AI Fellowship Starts TOMORROW

  • Deliverable: your company’s AI Transformation Plan

  • 4 classes with an 8 hour total commitment

  • Next cohort starts tomorrow, October 30

The Links

Practical

Perspectives

News

  • Albania’s AI Minister is pregnant…with 83 ‘children’: apparently the experiment with AI governance is going so well that they decided to roll it out across all of their departments.

  • AI sets ups Kodak moment for global consultants - the argument makes logical sense, but contrast that with the fact that companies like Accenture are seeing record revenue from ‘AI Consulting’ work and the picture becomes fuzzier.

  • OpenAI launched ChatGPT Atlas, its entrant in the AI browser ‘wars’. Use cases are limited and reception has been mixed. OpenAI also launched “Company Knowledge”, bringing your org’s information into a single place so you can query across Slack, Sharepoint, Google Drive, Github, and more.

  • Netflix goes “all in” on AI: they’re pitching it as a way to make ‘great creatives more efficient’, but they’re already creating AI-generated scenes and visual effects jobs (among others) are likely to be negatively impacted.

  • AI data centers under fire: Locals around the globe are protesting the rapid building of new data centers in their areas, citing water usage and broader environmental impacts, along with purported rises in electricity prices. Globally positive, locally negative is a hard political needle to thread, especially when the AI buildout hinges on data centers that have to be located next to someone’s neighborhood.

Thanks for reading,

Tom Guthrie

If you liked this edition of AI for Operators, share it with another operator you know!