The Algorithm: How to Run at Radical Speed

The 5-step operational framework for controlled chaos.

I spend a lot of time iterating work processes, constantly looking for the fastest path to the best result. Recently, I revisited a framework that puts this exact pursuit into perspective. It is based on an operational playbook utilised by former Tesla President Jon McNeill and codified by Elon Musk, known simply as ‘The Algorithm’.

It is a framework designed for radical simplification and acceleration. To apply it, you must reframe your daily work and processes as a manufacturing line: Raw Inputs > Processing > Finished Throughput.

The Algorithm is the blueprint for achieving high volume without sacrificing standards. Here are the five steps, in order. The order matters.

1. Question every requirement

Never take a requirement for granted. Every step in your work process was put there by a person, at a specific point in time, for a reason that may no longer exist. Interrogate each one: Who decided this? What evidence do we have that it matters? How do we know? And so what?

Requirements are sticky. They survive long after the reason for them has died because nobody wants to be the one who removed a step and broke something.

2. Delete every step you can

Once you have questioned the requirements, delete ruthlessly. Deletion isn’t a sign of failure; it is the key that unlocks speed.

Take a client deliverable. Most decks go through a final polish pass before anyone sees them: aligning boxes, restyling charts, hunting for the perfect icon, rewriting headers for the third time. It can take hours. But who decided every deck needs it? There is no law. Do clients accept a recommendation because the boxes align? No. They buy the thinking, not the polish.

The rule of thumb: if you are not forced to add at least 10% of your steps back later, you didn’t delete enough. Everyone fears deletion, but very few process decisions are irreversible doors. You can almost always add a step back. What you can’t get back is the time.

3. Simplify and optimise what’s left

Only now, after deleting, do you optimise. Most professionals get this backwards and spend their energy perfecting steps that should never have existed in the first place.

For each surviving step, ask: where do we get the most return for the least effort? If slide polish takes four of your ten hours, that is 40% of your time. Does it deliver anything close to 40% of the final value? Unless the polish is the product, a final-round pitch deck, say, the answer is usually no, and it is a candidate for simplification, or a return trip to step two. Weigh the opportunity cost. Those four hours could go toward work that actually moves the needle.

4. Accelerate cycle time

Speed comes from shortening feedback loops, not working harder. If you want a team to move faster, increase the frequency of communication. Replace “update me at the end of the week” with “update me at the end of the day” and watch what happens. Work expands to fill whatever loop you give it.

Three questions drive this acceleration: Who is going to do what, by when? And when they give you a deadline, ask why it can’t be sooner. That is not micromanagement; it is surfacing blockers early while they are still cheap to fix. A daily update reveals on Tuesday the problem you would otherwise discover on Friday.

5. Automate last

This is the golden rule, and the one AI has tempted everyone to break. Automation is only as good as the process underneath it. Automate a good process, and you get fast, brilliant results. Automate a flawed one, and you produce failure at scale, catastrophically fast. Humans who understand what great looks like must perfect the process first. Only then do the robots come in.

Because AI is now cheap and accessible, people jump straight to step five, automating tasks that should have been deleted at step two. Remember: the only thing more efficient than automation is elimination. Choosing not to do a worthless thing beats doing it instantly.

Here is a simple litmus test. AI platforms bill you every month whether your output improves or not. Ask yourself: am I getting better at my job? Are we generating more revenue? If your subscription costs keep rising and the answers don’t change, you haven’t automated strategy. You have automated waste.

The Verdict

The Algorithm isn’t really about Tesla, and it isn’t about slide polish. It is a way to put your own work on trial. Question the requirements. Delete the stupid ones. Simplify what survives. Speed up the loops. And only then, automate what is left.

Most of us are running processes we inherited and never interrogated. Ten hours of actual work, three weeks of delivery time.

That gap is exactly where your growth is hiding.