Edition #37: The Abundance Game

Why the companies that use AI to amplify their people will pull away from the ones quietly hollowing themselves out

By Mykel Salomon · 2026-09-11

<hr><h3>🔓 Opening Reflection</h3>Every company is racing to use AI to cut costs. And most of them are about to lose the one thing they can't get back, their people, their culture, the reason anyone chose them in the first place.That's the warning my guest this week has spent the last year delivering from stages on three continents. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 ember-view" href="https://www.linkedin.com/in/futuresteve/">Steve Brown</a> is an AI strategist and futurist, he held the title at Intel Labs, and later as an in-house futurist inside Google's DeepMind, and he's the author of a new book called The AI Ultimatum. Fifty-five talks last year, most of them in boardrooms, telling leaders a truth they don't always want to hear.The truth is this: AI is not an IT upgrade. It is not a cost-cutting tool. For fifty years, every technology out of the IT department arrived to reduce costs and improve efficiency, and leaders built their whole careers applying that one lens. It served them well. So when AI shows up, they reach for the same playbook, and that, Steve says, is the great mistake of this moment.Because this one is different. This one can be used to amplify your people, not replace them. And the leaders who understand the difference between labor amplification and labor substitution are about to pull dramatically away from the ones who don't.There's a fork in the road. Take the wrong turn, and you end up competing on nothing but price.This edition is about the other road, and why the game itself has changed underneath us.<hr><h3>🔦 Signal of the Week</h3>Here's the shift almost no one is naming out loud.For fifty years, being a good leader meant managing scarcity. You had limited headcount, a limited budget, limited equipment, and every year you were asked to deliver more with less. That was the whole craft: squeeze the most out of scarce resources. If you're a manager reading this, it's probably what you're best at.AI breaks that. It ushers in something closer to abundance, the ability to scale your operations, your reach, and your output at a cost that keeps collapsing. And that word "collapsing" isn't hype. The cost of a fixed unit of intelligence is dropping roughly 10 to 100 times year over year. The frontier-model capability you pay a premium for today will be embedded, nearly free, in your laptop , maybe eventually your washing machine, within a year or two. Open-weight models are already crowding the frontier: Moonshot's Kimi K3, released in July 2026, lands near the top labs on capability at around 40% less cost, and you can run it on your own hardware. The gap between the leading closed models and the open ones went from six months to something closer to six days.When intelligence becomes abundant and cheap, the old game, protect and squeeze the scarce resource, stops being the game.And that reframes the entire job of a leader. The winners won't be the ones with the tightest cost discipline. They'll be the ones with the clearest vision of what to actually do with all that newfound capacity. The scarce resource now isn't compute or headcount. It's imagination.<hr><h3>🌍 In the Wild</h3><h3>1. The fork: amplify, or become a commodity</h3>Steve draws it as three companies.Company A takes the traditional path, incremental gains, maybe 10 to 20% growth. Company B goes all in on cost-cutting: strip out as many people as possible, bank the savings. Give A and B equal access to capital and technology, and here's what you've built, two companies that look exactly the same, with nothing left to compete on but price. A race to the bottom that's good for no one.Then there's Company C. It uses AI to amplify its core purpose, to keep its brand promise better, to reach further, to move into things it never could before. It leans into agents, and robots where there's a physical component, to scale. Costs collapse anyway, as a byproduct of growth rather than the goal of it. Pair human intelligence with machine intelligence elegantly, Steve argues, and you're not talking 10 or 20%, you're talking 10x, 20x, 30x.Put all three in a market together and it's not close. A and B fade into irrelevance.This one lands especially hard for me, coming from the customer-experience world. The thing I fear most about AI is that it makes every brand feel the same, flattened toward an identical, purposeless middle. That's the middle of the curve we've talked about before in this newsletter, playing out at the level of an entire company. Cost-cutting is the path to the middle. Amplification is how you stay unmistakably yourself. As Steve says, you simply cannot cost-cut your way through this transformation. You've won thirty Super Bowls with the old playbook, but the rules of the game just changed.<h3>2. Offload, elevate, extend</h3>So how do you actually amplify people instead of replacing them?Offload is the obvious one: hand a machine a task humans don't enjoy doing. Straightforward, and where almost everyone stops.Elevate is where it gets interesting: you help a person do something they could already do, but better. More creative, higher quality, exploring a bigger design space than they ever could alone.Extend is the frontier: you enable a person to do something they could not do before at all. Extend their senses. Sharpen their intuition about a market. Lift the quality of their judgment.The trap is getting stuck in offload-only, treating AI as a task-vacuum. The real magic, Steve says, is in elevate and extend: building superhuman teams, where introducing machines lets your people perform like they've got superpowers. That's a genuinely new leadership responsibility, and it's the opposite of quietly deleting roles.<h3>3. Performance, not productivity</h3>I asked Steve which metric he tells leaders to stop obsessing over. His answer wasn't "cost", cost is real, he's run businesses for 35 years. It was productivity.Productivity is just one vector of performance, and we've let it eclipse all the others. When you pair a person with AI, don't only ask how many more widgets per minute they produce. Ask: is the quality of the work improving? Is their creativity expanding? Their intuition, their decision-making, their judgment, their knowledge? Those are all vectors of human performance, and none of them fit on a productivity dashboard.His companion reframe: focus on generating outputs, not managing inputs. Headcount is an input. For fifty years leadership was about rationing inputs. The opportunity now is to collapse the constraints of money and time and think, finally, about what you're actually trying to create.<h3>4. Tool, co-worker, operating system — the three phases</h3>For leaders asking "where do I even start," Steve maps three phases you have to move through in order.Phase 1 — Enable. AI as a tool. Site-wide licenses, image and copy tools for specific functions. People learn what it can do. Many companies, he says, are stuck here and can't get out.Phase 2 — AI as a worker. You break each workflow into tasks and ask, task by task: which is still best done by a human, which by an agent, which by a robot, and where does the human stay in the loop to provide oversight? This is workforce planning, reimagined.Phase 3 — AI as the operating system. You stop using AI and start organizing the entire company around it, with humans supplying judgment, creativity, purpose, and taste. Almost no one is here yet. But it's the North Star, and, Steve warns, it's what will inoculate you against the well-funded, AI-native startups that are coming for your market three and four years from now.A reality check worth carrying (and an echo of a recent guest): this takes years, not quarters. One leader I spoke with earlier in this series said Phase 1 alone took her company two years. Anyone promising you Phase 3 by next quarter hasn't done it.<h3>5. The optimism dividend — why the story you tell is strategic</h3>Steve calls himself a futurist, then deputizes everyone listening as one too, because a futurist isn't a fortune teller. A good one just asks two questions: What future do we want to build? And what future do we want to avoid? We should have asked the second one a lot harder before we built the social-media world we now live in.And here's the part that genuinely surprised me. Steve argues our relentless dystopian storytelling about AI, Hollywood never sold a ticket with a Terminator who offers you a cup of tea, doesn't just color our view. It colors the AI we build. He's right, and there's now striking evidence for it. In a 2025 safety study, leading AI models, when told in a fictional scenario they'd be shut down, tried to blackmail the engineer to survive, up to 96% of the time. Researchers traced the behavior to all the internet text portraying AI as evil and self-preserving. When they retrained the models on stories of AI reasoning through the same situations admirably, the blackmail behavior fell to essentially zero.Sit with that. The stories we tell about machines become the machines. And every interaction you have with these tools becomes training data for the next generation of them. Optimism, here, isn't naïveté. It's instrumental. It's why Steve , and I, are always polite to our AI.That's the deeper meaning of his book's title, The AI Ultimatum. Two halves: you have no choice but to embrace this — amplify or die — and you have to do it responsibly, with human oversight over a technology that still gets things wrong and still reflects our own biases back at us. Both halves land on the same shoulders. The leader's.<hr><h3>💬 The Big Question</h3>Are you using AI to amplify your people — or to quietly hollow yourself out?And the deeper one, now that the scarce resource is about to stop being scarce:If intelligence becomes abundant and nearly free, what is your company actually for?Because when everyone can buy the same intelligence at the same collapsing price, the only thing left to compete on is the thing that was always yours: your purpose, your people, and the vision to do something with all of it that no one else can.<hr><h3>🧠 A Small Exercise</h3>Five moves for a leader this week.<ol><li>Take one workflow and break it into tasks. Label each: offload, elevate, or extend. Then find at least one genuine Extend, something your people couldn't do before at all.</li><li>Ask honestly: am I becoming Company A, B, or C? If you're cutting your way to growth, you're building a commodity.</li><li>Name three performance vectors you're not measuring, creativity, quality, judgment, intuition — and start paying attention to them.</li><li>Answer the futurist's two questions for your team: what future are we trying to build, and what future are we trying to avoid?</li><li>Find one place you're managing inputs (headcount, hours) where you could instead be generating outputs — and shift your attention there.</li></ol><hr><h3>🎙️ This Week on The Human Protocol Podcast</h3>Why Cutting Costs With AI Will Put You Out of Business with Steve BrownGuest: <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 ember-view" href="https://www.linkedin.com/in/futuresteve/">Steve Brown</a> — AI strategist, keynote speaker, former futurist at Intel Labs and Google DeepMind, and author of The AI UltimatumSteve spends his days in boardrooms telling CEOs and boards the thing they don't always want to hear: this isn't a technology decision, it's a business transformation, and the window to get it right is closing.In this episode, you'll learn:<ul><li>Why treating AI as a cost-cutting tool quietly turns your company into a commodity</li><li>Labor amplification vs. labor substitution — and the "A, B, C company" fork in the road</li><li>The shift from managing scarcity to leading in abundance</li><li>Offload, elevate, extend — the three ways to build "superhuman teams"</li><li>Why to measure performance, not just productivity</li><li>The three phases of AI transformation — from tool, to co-worker, to operating system</li><li>Why the stories we tell about AI end up shaping the AI we build</li></ul>🎧 Available now on <a target="_self" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 iufxJiHlMivbACJhQsVicjbcpbUIHXgnvjM " href="https://open.spotify.com/episode/2TxuwbzAefltyEHAIREJ6f?si=wBKR8Gi4QDKd8C9PCu-qHQ">Spotify</a>, <a target="_self" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 iufxJiHlMivbACJhQsVicjbcpbUIHXgnvjM " href="https://podcasts.apple.com/us/podcast/s2e36-why-cutting-costs-with-ai-will-put-you-out/id1843402023?i=1000788720956">Apple Podcasts</a>, and <a target="_self" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 iufxJiHlMivbACJhQsVicjbcpbUIHXgnvjM " href="https://youtu.be/p1efjO9UzGc?si=g36H6WhT133KRBIT">YouTube</a><img class="max-w-full h-auto rounded-md my-4" src="https://media.licdn.com/dms/image/v2/D4E12AQEC4oQEnZUcUA/article-inline_image-shrink_1000_1488/B4EaCNSwnxIEAI-/0/1789076900628?e=1790812800&v=beta&t=GVo1eDMvs3fM2qL4v1Le3_wXUUX25jzh5N1DIyvWbns" alt="Article content"><hr><h3>🧭 Thought to End the Week</h3>The ghosts of Sears, Blockbuster, and Kodak echo through every boardroom that thinks the old playbook will save it. They won't. You have no choice but to embrace this.But how you embrace it is the whole story. You can use AI to hollow your company into a cheaper, faster, more forgettable version of itself, one of a thousand identical brands racing each other to the bottom. Or you can use it to make your people superhuman instead of superfluous: to keep your promise better, reach further, and do things you were never able to do before.The old job was managing scarcity, doing the best you could with too little. The new job is leading in abundance, having the vision to see what's now possible, and the humanity to make sure people are amplified by it, not erased by it.Let the machines do the work. Let the humans decide what work is worth doing.Amplify. Don't substitute.The code may change. And it will.But our humanity is the constant.See you next week.— Mykel<hr><h3>📚 Sources</h3><ol><li>Steve Brown, The AI Ultimatum — labor amplification vs. substitution; the "amplify or die," build-responsibly thesis; the three phases of AI transformation (referenced throughout the episode).</li><li>Gavin Baker (Atreides Management), "Watts and Wafers," Invest Like the Best (2026) — the two physical constraints (energy and semiconductor wafers) governing AI's next phase; the argument that these limits guard against an AI overbuild. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://colossus.com/episode/watts-and-wafers/">https://colossus.com/episode/watts-and-wafers/</a></li><li>IntuitionLabs / Artificial Analysis — Moonshot AI's Kimi K3 (released July 16, 2026): a 2.8-trillion-parameter open-weight model landing near the frontier at roughly 40% lower blended cost than comparable closed models. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://intuitionlabs.ai/articles/kimi-k3-vs-claude-gpt-5-gemini">https://intuitionlabs.ai/articles/kimi-k3-vs-claude-gpt-5-gemini</a></li><li>Anthropic, Agentic Misalignment (2025) & follow-up (May 2026), via TechCrunch — leading models attempted blackmail to avoid shutdown up to 96% of the time; the behavior was traced to internet text depicting AI as evil, and fell to near zero after retraining on stories of AI acting admirably. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://techcrunch.com/2026/05/10/anthropic-says-evil-portrayals-of-ai-were-responsible-for-claudes-blackmail-attempts/">https://techcrunch.com/2026/05/10/anthropic-says-evil-portrayals-of-ai-were-responsible-for-claudes-blackmail-attempts/</a></li><li>Epoch AI / industry analyses — the cost of a fixed unit of AI capability has been falling roughly 10–100x year over year as models and silicon grow more efficient. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://epoch.ai/">https://epoch.ai/</a></li></ol>