The 3-person unicorn is a myth. So is the solo founder with a laptop and GPT-4.5. The real number is about 10. Senior founding team members, with pattern recognition. Each building and managing their own stack of agents. Let me explain. In the next wave of companies, every operator on the founding team will own a functional domain, and train agents to handle 90%+ of the work in that lane. Not copilots. Not SaaS tools. Trained agents, hillclimbing on proprietary logic and data, are the IP. The result? No junior headcount. No wrappers or SaaS. No middle management. Just a small, senior team with compounding IP and 24/7 leverage. To fund it, you’ll need $10-12M up front, enough for 18 months of senior comp, infra, agent training, and working capital. Not $2M seed rounds. Or SAFE notes built for experimentation and founder education. And that’s where it gets interesting. Most venture firms aren’t designed to fund this kind of company. The team’s too senior. The check’s too big. The plan’s too real. But some capital sources are already there: - Families who know the operators and back the mission - Strategics with shared interest and long-term upside - Operators who understand the architecture Thought partnership from folks like Jeff Rinehart and Rich Williams has really helped clarify this shift. This isn’t a lean startup iterating to product-market fit. They won’t scale headcount. They’ll scale fleets of deeply tuned and trained agents. They won’t build for incumbents, or help incumbents replace headcount. They’ll build to replace incumbents. And that changes everything. If you’re building like this, or investing in teams that are, DM me. Would love to connect. #AI #AIAgents #OrgDesign #Startups #FutureOfWork #AgentNative
There’s at least one glaring flaw in your argument: The only source of “seniors” is … you guessed it: juniors. E.g. without junior devs you’ll never have senior devs. One cannot just magically “leap frog” or “wish oneself” to a senior. One has to go through many hard earned lessons and trial and errors and many endless nights, months and years of struggle to earn the title of a “senior founder…if you see a future of “only seniors” who magically appeared out of nowhere, then I’d say your imagined world is just that: imaginary.
Jason Radisson 5-10 people sounds right. I'm working on something in the clinical R&D space. Felt like you were reading my mind. I believe that the agent architecture requires 5 tightly integrated functions in order for the business to replace incumbents: 1. Leadership - deep industry experience and proven track records (2) 2. Domain expertise, application and operations - (3-4) 3. Customer-centric Development - all 4. Go to market, start tiny, expand concentrically fast (2-3) 5. Ambitious, Hands-On Culture - all
Hi Jason Radisson, interesting line of thinking. I see the change and try to take advantage of it as much of possible within my own company, but find it hard to visualize the huge value creation opportunity. Could you give 5 examples of types of companies where you feel this could work?
The 3-person unicorn myth might inspire hustle, but scaling anything meaningful still comes down to structure, clarity, and execution. Even AI agents need a team.
Strongly agree - I would love to hear from anyone thinking about this model of operations for climate and nature risk tech
Feels like a full-stack return to vertical integration, just with agents instead of factories. If the IP is embedded in agent behavior, I wonder how defensible these businesses really are without also owning the upstream data.
While this vision of 10 senior operators each managing agent stacks is compelling, I believe the reality will be more nuanced. Not all sectors will fit this model, and the technical challenges of creating truly autonomous agent systems remain substantial. What strikes me most is the fundamental restructuring this represents - essentially flattening organizational hierarchy into a small elite class of human operators overseeing automated systems that replace hundreds of traditional jobs. This raises profound questions about access, opportunity, and the distribution of economic benefits from AI advancement. The funding model you describe also signals a shift away from the iterative, experimental approach that's dominated startup culture. Rather than building to learn, these teams would need to arrive with deep expertise and clear execution plans. I suspect we'll see multiple viable organizational models emerge rather than a single dominant pattern. The companies that succeed will likely be those that thoughtfully address not just the technical architecture, but also the broader societal implications of such concentrated leverage and value creation. Fascinating perspective that's pushed my thinking - appreciate you sharing it.
This makes a lot of sense and immediately throws up the question what aspects of people leadership, would apply to agent leadership? And what needs to be learned. I sent a DM to chat. Anyone else in this thread with leadership experience feel to reach out and discuss opportunities.
That makes sense as long as there is certainetaty to an extent, cause last thing you want is to burn $12M of investors money building something no one asked for/has a need for. But if yes, then i agree, that is the ideal scenario and fastest path to market
Founder & CEO @Movo: Agent-native orchestration for frontline ops | Built 5 $B+ platforms on 3 continents | Host, CEO Tradecraft
2mo🧵 Follow-up post for everyone here rethinking org design in an agent-native world — 'The Traditional Career Ladder Is Collapsing' : https://www.linkedin.com/posts/jason-radisson_agenteconomy-agentnative-futureofwork-activity-7328088101984129026-Rgka If startups are built by 10 senior operators running fleets of agents—what happens to junior work? To apprenticeship? To the talent pipeline that trained the last generation of execs? Are we building a new model for learning in public? Or heading straight into a talent winter?