To AI or not to AI.
That is the question being debated in boardrooms all over the country this quarter. And I guarantee while the question is being debated, AI is already changing the org chart of the company having the debate.
That is the part nobody wants to say out loud.
The deliberation is not neutral. It is not a pause button. It is a tax. Every week a leadership team spends weighing whether to commit to AI is a week their competition is not weighing it. Their competition is using it. Their competition is learning what it can and cannot do. Their competition is compounding the advantage that comes from hours spent with the tool in hand. By the time the deliberating team reaches a conclusion, the question has already been answered by the market.
And this was never more profound than last week
Claude 4.7 came out. It is here. It is fast. It is intelligent. And it is better than anything I have ever seen.
I am not saying that because I want to sound impressed. I am saying that because the curve just moved again, and most of the people who need to see that it moved are in meetings debating whether the curve is real. The release happened on April 16. A week later, anyone paying attention has had time to use it, test it, push it, and form a view on what it changes about their own business. Anyone not paying attention is still on page three of the board memo.
It is not an experiment
What is interesting to me is how many people continue to treat this as an experiment. They use the word experiment. They talk about pilot programs. They talk about evaluating. They talk about waiting to see how it all shakes out.
It is not an experiment.
An experiment is something you run to learn whether something works. That question has been answered. The tools work. The capability is real. The productivity improvements are real. The efficiency gains are real. The companies and the individuals already using these tools are not running experiments. They are executing. They are compounding. They are pulling ahead every single day.
What is actually happening right now is a sorting. Not an experiment. A sorting.
Some people are getting on the train. Other people are standing on the platform debating whether the train is real. And the train does not care about the debate. The train is leaving.
The cost of the experiment framing
The problem with framing AI as an experiment is that it gives people permission to stay on the platform. If it is an experiment, you do not have to commit. You can watch. You can wait. You can evaluate. You can observe how the experiment plays out before you decide whether to participate. That framing feels responsible. It feels measured. It feels like the grown-up move.
It is actually the most expensive mistake anyone in a leadership position can make right now.
Because while the experiment framing lets you feel responsible, it is building a gap between your organization and the ones that are not treating it that way. And that gap does not stay small. It compounds. Every week you wait is a week someone else is using the tools, getting better at using the tools, building workflows around the tools, and training their team on the tools. You cannot catch up to that from a standing start. The distance you need to cover grows faster than the rate at which you can close it.
The real estate parallel
I see this in real estate constantly. There are agents and brokerages right now who are already using AI every day in ways that are materially changing what they can do in a given week. There are other agents and brokerages who are still debating whether AI is going to matter. Same industry. Same market. Same products available. Fundamentally different trajectories, and that divergence started months ago and is widening every week.
The agents who have leaned in are not necessarily the most technical. They are not necessarily the youngest. They are not necessarily the ones with the biggest budgets or the most sophisticated tech stacks. They are simply the ones who stopped waiting for permission to try. They opened a tool. They used it on one real task. They saw what happened. They tried it on a second task. That is all it took to start the compounding.
The agents still sitting it out are doing what every late adopter in history has done. They are telling themselves they will start once the technology is mature. They will start once the dust settles. They will start once it is clear which platform to commit to. They have a list of reasons to wait, and every one of those reasons sounds responsible in isolation. None of them hold up against the actual math of what is happening in their market.
The deliberation is the decision
Here is what I would say to anyone still in the deliberation phase.
The deliberation is the decision.
Every day you spend deciding whether to commit to AI is a day your answer is effectively no. Not no as a policy. No as an outcome. The companies that will win this decade are not the ones with the best AI strategy memo. They are the ones whose people have the most hours with the tools in their hands. The ones who stopped deliberating months ago and started doing. The ones who treat each new model release the way an athlete treats a new piece of training equipment, as something to go use immediately and figure out what it unlocks.
The curve is accelerating
Claude 4.7 is not the end state. It is not even close to the end state. But it is the strongest consumer-facing general intelligence model I have ever used, and that is saying something because I have used every one of them as they have come out. Whatever it can do today, the next version will do better. The curve is not slowing down. The curve is accelerating. And every release closes the gap between what sounded like science fiction six months ago and what a single user can do with a laptop and a subscription right now.
Think about what that actually means. Six months ago, certain capabilities required a research team. Today, those capabilities live inside a tool that any professional can open on their phone at a stoplight. The tools that exist right now would have been considered impossible a year ago. The tools that exist a year from now will make the current ones look primitive. This is the compounding nature of the curve, and it is why the cost of staying still is going up every quarter.
The professionals who understand this are not waiting for the curve to flatten. They are using the curve. They are treating each release as an upgrade to their own capabilities, not as an invitation to pause and evaluate. They know that the only way to stay in the game is to keep using the newest tools on real work, and to let the compounding effect of hours on the tool translate into output that the competition cannot match.
Get on the train
In a curve that is accelerating, the penalty for staying still is compounding. You do not just fall behind. You fall behind faster.
So when I hear someone in a boardroom ask to AI or not to AI, I hear the wrong question. I hear a framing that protects the comfort of the people in the room while the answer is being decided without them. The right question is not whether to use AI. The right question is how aggressively can we deploy it across everything we do, starting this week, and how fast can we compound the advantage that use creates.
That is the conversation the winners are having.
The losers are still on the platform, debating whether the train is real. And the train has already left the station.
Get on the train.
Judd Hoffman is CEO and Co-Founder of Ethica AI, building AI-powered tools for real estate transaction workflows.