A summary and commentary. Source: the Russian translation by Sergey Pimenov, pimenov.ai/articles/smert-kompanii-organizatsionnaya-singulyarnost (in Russian). Original by Peter H. Diamandis: x.com/peterdiamandis/status/2092667837529694400. The idea builds on a masterclass by Salim Ismail.
The core claim
The author argues that companies with traditional structures (layered hierarchies, annual plans, departments that cannot reorganize quickly) will eventually become non-viable. The reason: AI-native teams, where people work alongside AI agents, can ship results many times faster. The author calls this transition organizational singularity.
The 70% test
The starting point is a question Salim Ismail asks every CEO he works with: does your company have a high-margin unit that two people with AI agents could rebuild in 60–90 days? According to the author, about 70% of leaders say yes. The figure comes from his conversations with CEOs rather than an independent study, so treat it as an estimate.
What the author proposes
- An intelligence stack. An AI-native company is built in layers: sensing data, interpreting it, making decisions, orchestrating execution, and learning from outcomes. A governance layer sits on top and sets the boundaries.
- The periphery strategy. Rebuilding a company from the inside is hard, so the author recommends starting with peripheral processes where a hypothesis can be tested quickly without risking the core business.
- Productivity gains. The author cites pilot programs with ten companies and estimates roughly 100x productivity with about a fifth of today’s headcount. These are his pilot-based calculations, not a verified result.
- The edge strategy in practice. A new AI-native operation is built alongside the old one, processes are moved over one at a time, and the old part is retired once the new one outperforms it.
What survives and what dies
The author expects the company’s purpose, its legal shell, its own data and learning loops, and human judgment to survive. Static org charts, five-year plans and a large middle-management layer that exists to coordinate will not. He also gives a 60–80% cut in middle management as his own estimate.
What it means for different roles
- CEO: start asking which units could be rebuilt around agents, instead of postponing it to next quarter.
- Founder: see dominant high-margin incumbents with slow structures as a target for a small team.
- Investor: favor companies that own their data and have fast learning loops; be cautious about companies with static org charts and five-year plans.
- Employee: routine office work will be automated first. Skills that are harder to replicate, such as judgment, handling uncertainty and experience, will hold their value better.
My take
I agree with the core idea: the speed at which a company can reorganize is becoming a competitive advantage, and it can be tested in practice without waiting for a large reorganization. I would still treat the numbers with caution. Much of the article is projection and generalization from a masterclass rather than research data. The most useful takeaway is the question “what would we rebuild in 90 days with a two-person team and agents?”, answered for your own business.
See the links above for the full text and all sections.