What if the biggest mistake leaders can make with AI is spending too much time worrying about the future and too little time building it?
That was one of the strongest themes in my recent conversation with Reid Hoffman, co-founder of LinkedIn, co-founder of Inflection AI and Manas AI, Microsoft board member, pioneering technology investor, author, podcaster, and one of the most influential voices in entrepreneurship and artificial intelligence.
Hoffman has never been shy about big ideas. He has helped shape the internet economy, invested in some of the world’s most important technology companies and now spends much of his time thinking about what AI means for humanity, business, health, work and society.
What struck me most in our conversation was his framing of AI optimism. This is not blind enthusiasm. It is not Silicon Valley cheerleading. It is a practical, strategic argument that leaders need to spend more energy asking what could go right.
As Hoffman put it, “You only get to the good futures by steering towards them.” That sentence captures the leadership challenge of AI beautifully. Avoiding risk is important, yet avoidance alone does not create progress. Companies, governments and individuals need a destination worth moving toward.
Why Strategic Optimism Matters
AI conversations often begin with fear. Will AI take jobs? Will it make mistakes? Will criminals misuse it? Will companies lose control of their data? These are valid concerns, and any responsible leader needs to take them seriously.
Yet Hoffman’s point is that fear cannot be the whole strategy. If a company only focuses on avoiding bad outcomes, it can easily end up avoiding the future itself. That is a dangerous position when AI is becoming a general-purpose capability across almost every area of business.
The more useful question is, what would we do if we were deliberately steering toward the best possible outcomes?
For Hoffman, that includes AI medical assistants on every phone, legal assistants that help people understand their rights, educational assistants that help people reskill and government interfaces that make public services easier to access. These are great examples of how AI could expand access to expertise that has historically been expensive, scarce, or difficult to navigate.
This is where I believe many leaders need to shift their mindset. AI strategy should not begin with a tool selection exercise. It should begin with ambition. What would better customer service look like? What would better decision-making look like? What would better access to expertise look like? What would it mean for every employee to have a capable assistant that helps them think, create, analyze, and act?
That is where the real opportunity lies.
The End Of Pilot Purgatory
One of the most practical parts of our conversation was Hoffman’s view on why so many companies struggle to move beyond pilots.
Many organizations are still treating AI like a conventional technology rollout. They assign a small team, run a limited experiment, prepare a report and then wonder why momentum stalls.
Hoffman argues that this model often fails because AI changes how people work. It is hard to understand that shift from a narrow pilot in one corner of the business. He gave a very simple recommendation: start with meetings.
As he put it, “You should go all in on meetings right away.” The idea is straightforward. Companies should use AI to record, transcribe, summarize, identify follow-ups, assign action items, highlight missing questions and share relevant information with people who were not in the room.
That might sound mundane, but it is exactly the sort of everyday workflow where AI adoption becomes real. Meetings are where decisions are discussed, knowledge is shared, action is agreed and accountability often becomes messy. Improving that information flow can teach an organization a huge amount about where AI can create value next.
I see this all the time. The companies making progress with AI are the ones building learning loops. They encourage people to experiment, share what worked, discuss what failed and keep adjusting.
Hoffman suggested adding a simple question to weekly team meetings: What did you try with AI, and what was the result? That simple habit can drive more real AI adoption than another top-down initiative.
Everyone Will Manage Agents
The biggest shift in work may be that AI turns more people into managers of digital capability.
Hoffman said something that should make every executive pause: “I don’t think we’ll have individual human contributors anymore.” His point was not that people disappear from the workplace. His point was that almost everyone will be “managing and leveraging a set of agents.”
That has huge implications for skills, job design, education, leadership and organizational structure.
If AI agents can draft, research, analyze, code, summarize, test and generate ideas, the human role shifts toward orchestration. People will need to decide what to delegate, how to combine agents, how to check outputs, how to frame good questions and how to turn machine-generated work into meaningful outcomes.
Hoffman called this out directly: “Agent management will now be a basic skill that everyone needs.”
I think this is one of the most important workforce messages for leaders today. AI literacy is not a niche technical skill. It is becoming a core business skill. In the same way that digital literacy became essential across functions, agent literacy will become part of how people get work done.
That does not mean everyone needs to become a programmer. It means people need to understand how to work with AI systems effectively, how to challenge them, how to improve them and how to use them responsibly.
What Young People Should Learn Now
One of the most urgent questions around AI is what happens to entry-level work. Many junior roles have historically involved research, administration, document review, basic analysis and other tasks that are increasingly automatable.
Hoffman’s advice to young people was refreshingly direct. Right now, being AI native can be an advantage. A young person entering the workforce can say, in effect, I understand these tools, I can help your organization adapt and I can bring new ways of working into the firm.
That is important. The threat to junior jobs is real, but so is the opportunity for young people who learn faster than the systems around them.
The apprenticeship model will change. A junior lawyer may spend less time digging through old documents and more time using AI to prepare summaries, stress-test arguments and red-team a partner’s thinking. A junior analyst may spend less time formatting spreadsheets and more time shaping the questions that matter. A junior marketer may spend less time producing first drafts and more time orchestrating campaigns, testing ideas and interpreting what the market is saying.
This is where education and training need to evolve. We should be teaching people how to ask better questions, evaluate AI-generated outputs, understand context, communicate clearly and exercise judgment. The rote tasks may shrink. The need for judgment will grow.
Competitive Advantage In The Age Of AI
A common question I hear from executives is this: if everyone has access to powerful AI models, where does competitive advantage come from?
Hoffman’s answer combined classic business thinking with AI-native thinking. Some old principles still matter. Network effects, customer relationships, deep integration into enterprise workflows, trusted brands, proprietary expertise and strong distribution can still create moats.
At the same time, AI opens new possibilities. Hoffman used Manas AI, the drug discovery company he co-founded, as an example. His point was that applying AI to biopharmaceuticals requires high-quality AI and high-quality biology. Software alone is not enough. Domain expertise matters enormously.
That is a crucial lesson for business leaders. The winners will not be the companies that simply add a chatbot to existing workflows. It will be the companies that combine AI with unique data, deep expertise, strong customer understanding and redesigned processes.
AI is powerful, but the advantage comes from how it is applied.
Speed, Safety And The Need To Learn
Hoffman is clear that AI comes with risks. He worries about criminals, cybercrime, rogue states, bioterrorism and the wider social disruption that can come from fast technological change.
Yet he also argues that zero risk is impossible. If companies demand perfect safety before experimentation, they will never start. The better approach is to distinguish between acceptable learning mistakes and serious harms.
A flawed internal summary is one kind of error. A cybersecurity vulnerability, dangerous medical misinformation or misuse by malicious actors sits in a different category. Leaders need governance that is strong enough to manage serious risks without freezing the organization.
In practical terms, that means clear policies, secure environments, human oversight, auditability, training and a culture where people learn from mistakes rather than hide them.
Hoffman quoted Ethan Mollick’s memorable line: “The worst AI you’re ever gonna use is the AI you’re using today.” That is a useful reminder. AI systems will continue to improve, and organizations need to build a habit of continuous learning.
A use case that fails today may work six months from now. A workflow that feels clumsy today may become transformative as models improve and employees gain confidence. The companies that keep learning will pull away from those that wait for certainty.
The Leadership Challenge Ahead
The most important takeaway from my conversation with Reid Hoffman is that AI leadership is about movement.
Leaders cannot outsource this to a small innovation team. They need to use AI personally. They need to encourage their people to experiment. They need to integrate AI into the everyday flow of work. They need to look for areas where their organization has unique capabilities that AI can amplify.
Hoffman’s advice was clear: get people engaged, integrate AI into meetings and information flows and think hard about where your organization’s knowledge and capabilities could create future industries.
That is exactly the right place to start.
The future of AI will be shaped by the leaders who ask better questions, create stronger learning cultures and steer toward useful, human-centered outcomes.
AI will change work. It will change organizations. It will change competition. The real question is whether leaders will treat it as another tool to bolt onto existing processes or as an opportunity to rethink what their organizations can become.
