I read "Surrounded by Idiots" by Thomas Erikson about a year ago, and it's one of those books that rewires how you see every meeting, every team dynamic, every awkward Slack thread. The core idea is simple: people fall into four behavioral types (Red, Blue, Yellow, Green), and most of the friction in teams comes from these types misunderstanding each other, not from anyone actually being an idiot. What's been stuck in my head lately is how AI changes the equation. Not the personality types themselves (those are pretty hardwired), but how much each type's strengths and weaknesses get amplified when everyone on the team has access to tools that can execute, analyze, write, and build at a pace that was impossible two years ago. The four colors, quickly If you haven't read the book, here's the version that'll click in about 30 seconds. Red is the person on your team who hears the idea at 9 AM and has a working prototype by lunch. They don't wait for alignment meetings or strategy docs. They just go. Think of the coworker who responds to a 20-message thread with "I just built it, here's the link." That's Red. Red Blue is the one who reads the entire 80-page requirements doc before the kickoff call and shows up with three questions that nobody else thought to ask. They go deep. They find the flaw on page 47 that would've blown up in production. They're slow, but they're rarely wrong. Blue Yellow is the person who knows everyone on the floor by name, sells the vision before the deck is even finished, and somehow gets the skeptical VP to sign off on a project that has no business getting funded yet. They run on energy and relationships. Yellow Green is the calm one. The person who makes sure everyone's been heard in the meeting, who notices when someone's getting steamrolled, who quietly follows up on the thing that was about to fall through the cracks. They don't grab the spotlight, but teams fall apart without them. Green Most people are a blend (I'm probably some mix of Red and Blue, if I'm being honest), but one color usually dominates, especially under pressure. And pressure is where AI changes everything. How AI Amplifies Each Color Reds: the most dangerous and the most valuable AI is the best thing that's ever happened to Reds. The gap between "I have an idea" and "here's a working version" used to be days or weeks. Now, it's hours, sometimes minutes. A Red with Claude Code or Cursor can go from concept to deployed prototype in a single sitting. I've watched it happen on my own teams. Someone decides to just build the thing, prompts their way through it, and by the end of the day, there's a working demo that would've taken a three-person team two weeks a year ago. This is incredibly powerful and incredibly dangerous. Because Reds don't naturally stop to ask, "Should we build this?" They ask, "How fast can I build this?" AI removes the last bit of friction that used to slow them down, and that friction was sometimes the only thing preventing a bad idea from becoming a shipped bad idea. The Reds who thrive in this environment are the ones who learn to pair their execution speed with just enough validation to avoid shipping disasters. In practice, that usually means having a Blue nearby. Blues: the quality layer that AI can't replace Here's what's interesting about Blues in the AI era. A lot of their traditional work (research, analysis, data processing, building detailed models) is exactly the kind of thing AI is getting good at. You might think that puts Blues at risk. I think it does the opposite. When AI can generate a financial reconciliation or a compliance assessment in minutes, the bottleneck shifts from "who can produce this analysis?" to "who can tell me if this analysis is actually correct?" That's Blue territory. The ability to look at an AI-generated output and spot the subtle error, the wrong assumption, the data that doesn't quite add up, that's a skill that becomes more valuable as AI-generated output becomes more common. The Blues I'd be worried about are the ones who use AI to go even deeper into analysis rabbit holes without ever surfacing with a recommendation. AI gives Blues the ability to analyze forever, and some of them will. The Blues who thrive are the ones who position themselves as the validation layer, the person who writes the evals, reviews the AI's work, and gives the team confidence that the output is actually trustworthy. In a world where anyone can generate a deliverable in minutes, the person who can tell you whether that deliverable is right becomes the most important person on the team. Yellows: the least automatable skill set I keep coming back to this: AI can build, analyze, write, code, and design. It cannot read a room. It cannot build trust over a coffee that turns into a $2M project. It cannot sense that the CFO is nervous about something she hasn't said out loud yet and navigate the conversation accordingly. Yellows have the most automation-proof skill set of any color, and in a world where execution is cheap, the ability to sell, persuade, and build relationships becomes proportionally more valuable. The startup that can build the product in a weekend still needs someone who can get it in front of the right 50 people. The consulting team that can deliver an AI-powered assessment in 3 weeks instead of 12 still needs someone who can sell that engagement and make the client feel confident about a radically different delivery model. The risk for Yellows is different. It's not obsolescence, it's over-promising. When AI makes everything feel buildable, Yellows can sell visions that outrun their team's ability to deliver quality. "Yeah, we can absolutely do that" becomes a very easy thing to say when you've watched an AI prototype something in 20 minutes. The gap between demo and production-ready is still real, and Yellows need Blues and Greens to keep them honest about it. Greens: the highest risk, highest upside color This is the one I've thought about the most, and I want to be honest about it. If you self-identify as a Green, this section is for you. In a pre-AI world, Greens thrived because the pace of work was human-paced. There was time to build consensus, check in with everyone, and make sure the team was aligned before moving forward. Greens were the glue that held teams together, and their patience and diplomacy were genuinely essential. In an AI-accelerated world, the pace changes fundamentally. Reds are shipping faster. Blues are analyzing deeper. Yellows are selling harder. And if you're a Green whose natural instinct is to slow down, seek harmony, and avoid confrontation, you can find yourself watching the rest of the team pull away from you. The Greens who thrive will be the ones who actively step into an orchestration role. Not the passive "let's make sure everyone's comfortable" version of Green, but the active "I'm going to make sure this Red doesn't ship something broken, this Blue doesn't disappear into a rabbit hole, and this Yellow doesn't over-promise to the client" version. That's a genuinely hard and genuinely valuable role, and it's the role that AI is worst at filling because it requires reading people, not data. But if you're a Green who doesn't make that shift, if you stay in the passive, non-confrontational, "I'll wait for the group to decide" mode, the group will decide without you. Not out of malice, but because AI-accelerated teams move too fast to wait for consensus on everything. Building Teams in the AI Era This is where it gets practical. Whether you're staffing a consulting engagement, building a startup, or picking a co-founder, the color mix matters more now than it did before, because AI amplifies the gaps between unbalanced teams. The mix isn't static. It shifts with the stage of the work: Build & Validate (0 to 1): Red and Yellow co-dominant. In modern product work, building and validating aren't sequential; they're concurrent. You're shipping something small while simultaneously getting it in front of real people to see if anyone cares. Reds build fast with AI tools (genuinely 10x here), while Yellows are out talking to users, reading rooms, and figuring out whether what's being built actually solves a problem anyone has. Blue is lighter here because you're optimizing for speed of learning, not depth of analysis. Build & Validate (0 to 1): Scale phase (1 to N): Blue and Green move to the front. The prototype works; now you need people who can make it reliable, documented, and sustainable. Blues own the quality control, the evals, and the edge case analysis. Greens own the process, the team health, and the stakeholder alignment. Reds get restless here (this is the "boring" part), and if you don't have Blues and Greens, the thing that Red built fast will break fast. Scale phase (1 to N): Growth phase (selling and expanding): Yellow-heavy. The product works, and it's stable. Now, you need people who can get it in front of clients, build partnerships, and expand the footprint. Yellows do their best work here, but they need Blues to keep the promises grounded and Greens to make sure the team doesn't burn out scaling too fast. Growth phase (selling and expanding): The key insight is that no color wins alone, and AI makes that truer, not less true. A team of all Reds with AI tools will ship a lot of broken things very quickly. A team of all Blues will produce the most thoroughly analyzed product that nobody ever launches. A team of all Yellows will sell a vision with nothing behind it. And a team of all Greens will have great team morale and no output. The Real Skill: Knowing Your Color The whole point of Erikson's book is self-awareness. Not "be more Red" or "stop being so Green." Just know what you are, know what you're not, and build accordingly. AI doesn't change that. It just raises the stakes. Your strengths get amplified, but so do your blind spots. A Red who was already impatient becomes reckless at AI speed. A Blue who already over-analyzed now has infinite analysis to hide behind. A Yellow who already over-promised can now over-promise with a working demo. A Green who already avoided conflict now has even less time before the team moves on without them. The people I've seen navigate this best, whether in consulting, startups, or internal teams, are the ones who know their color, actively compensate for their weaknesses (usually by finding complementary people, not by trying to change who they are), and use AI to amplify the thing they're genuinely good at rather than trying to be all four colors at once. If you haven't read the book, read it. It's a quick one. And then look at your team through this lens and ask yourself: are we balanced, or is AI about to make our imbalance a lot more visible? I'm a Senior Manager at EY, getting back into writing after a long break. If any of this resonated or you want to riff on how AI is changing the way we build teams, feel free to reach out on LinkedIn — always happy to chat. I'm a Senior Manager at EY, getting back into writing after a long break. If any of this resonated or you want to riff on how AI is changing the way we build teams, feel free to reach out on LinkedIn — always happy to chat. reach out on LinkedIn Disclaimer: The views in this post are entirely my own and do not represent the opinions of my employer. Disclaimer: The views in this post are entirely my own and do not represent the opinions of my employer.

This story contains new, firsthand information uncovered by the writer.

Slack

You're Building Your AI Team Wrong: It's Not a Skills Problem

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

UI/UX Fails and Wins in the Cryptocurrency Space

The Noonification: How Often Do NFTs Pass The Howey Test? (1/13/2023)

Darwin's Hybrid Intelligence to Align AI & Human Goals for Startups & VCs

The Noonification: White Man (11/26/2022)

The Noonification: The Metaverse is a Sh*tshow (11/2/2022)

100 Days of AI Day 1: From Newsletter to Podcast, Leveraging AI for Audio Transformation

UI/UX Fails and Wins in the Cryptocurrency Space

The Noonification: How Often Do NFTs Pass The Howey Test? (1/13/2023)

Darwin's Hybrid Intelligence to Align AI & Human Goals for Startups & VCs

The Noonification: White Man (11/26/2022)

The Noonification: The Metaverse is a Sh*tshow (11/2/2022)

100 Days of AI Day 1: From Newsletter to Podcast, Leveraging AI for Audio Transformation

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps