19 min read
Elon Musk cut about half of Twitter's 7,500 employees during his first week as owner. He started before ChatGPT launched publicly.
That timing matters. The first Twitter cut was not an AI workforce transformation. It was concentrated authority, severe financial pressure, and an owner willing to absorb risks that most hired executives are not rewarded to take. Musk's ownership group had paid roughly $44 billion for the company, and he said Twitter was losing more than $4 million a day. Waiting one year at that rate meant watching about $1.46 billion disappear.
He owned the upside and the loss. Most executives do not.
The Incentive System Prefers a Staircase
Start with 8,000 employees and reduce the workforce by 15% each year. After five years, about 3,550 remain. After ten years, the number is about 1,575, close to the workforce Musk said Twitter had reached by April 2023.
A conventional company can approach the same operating-model change. It may simply take a decade, even with today's AI, because 10–15% is an excellent corporate result.
Consider a business unit with $500 million in controllable annual cost. A 10% improvement creates $50 million in recurring benefit. A 15% improvement creates $75 million. That can fund a maximum bonus, support a favorable board discussion, and leave another credible improvement for next year.
The leader who attempts $200 million at once receives capped upside and uncapped career risk. If the restructuring fails, the outage, lost customers, unvested equity, and next job carry the executive's name. If it succeeds, the savings become next year's baseline. The executive is not necessarily trying to win one race by five laps. They are trying to score well across the season without making another championship the new minimum.
AI Can Be an Accelerant or an Alibi
AI can make a smaller organization more capable. It can also give companies modern language for correcting overhiring, duplicated work, weak margins, and management layers that accumulated before the current tools existed.
Amazon's reductions and management-ratio target illustrate that distinction. Box offers the counterargument that lower-cost work can create more demand, more products, and new roles instead of turning every productivity gain into a layoff. Both outcomes are possible. The phrase AI-driven does not prove which one occurred.
None of this makes Twitter a clean operating playbook. People lost jobs and security with almost no preparation. X absorbed outages, advertiser flight, regulatory scrutiny, and reputational damage. A company continuing to run does not prove that it is healthy, safe, or creating durable value.
The Charge Belongs to the Rest of Us
You cannot care more than the leadership team does. Do good work, take the bonus available to you, learn how value moves through the business, document the economic outcomes you create, and build skills that remain useful outside the current org chart. If your company chooses the ten-year path, you should not finish the decade with the same skills.
The first time you become the CxO, your outlook will change. Yesterday, 500 people were colleagues and names you knew. Today, they are also a line in the operating plan you signed. The outage becomes your board memo. That perspective does not excuse bad leadership. It explains why you should learn the incentive system before you inherit it.
And if you are already a board member or executive with a different view, I would love to learn where this theory is wrong. What incentive am I missing? What have you seen make it rational to deliver a decade of responsible change now?
Companion
Most readers also read: If you cannot afford the tokens, can you afford to build it?
