What if most of your roadmap is work that should not exist at all?
Chris Franks, Anthony Franco, and Stephanie Hays take apart Elon Musk's five-step algorithm — make the requirements less dumb, delete, simplify, accelerate, automate — and stress test it against the reality of running an early-stage company. They dig into the Tesla fire-pad story that nobody could trace to an owner, why most founders run the sequence in reverse, and what changes when you treat cycle time as a learning instrument instead of a productivity metric.
The story that anchors the algorithm is small and absurd. Early in the Model S run, Tesla engineers had bolted a thick fire-suppression pad between the battery pack and the passenger compartment. It was expensive. It slowed assembly. The safety team said the noise team required it. The noise team said the safety team required it. Tesla removed the pad and neither metric moved. That is the entire first step of the algorithm in one anecdote. Make the requirements less dumb does not mean strip features out of a product. It means refuse to carry any requirement that cannot be traced to a named human being.
The second step is where most founders skip past the most important move. Deleting and simplifying are not the same operation. Deleting removes the thing entirely. Simplifying refines what remains. The single most common engineering mistake — and it shows up just as often in sales pipelines, onboarding flows, and internal reports — is optimizing a process that should not exist at all. A useful rule for spotting the difference: if you are not occasionally adding something back in, you are not deleting aggressively enough. Founders almost never do this because nothing visibly bad happens when you keep a step you do not need. The cost of over-process accumulates silently. The cost of under-process announces itself immediately. That asymmetry makes the wrong choice feel safe.
The algorithm assumes a system already exists to delete from. A founder walking into a freshly incorporated company has nothing to strip away. So the first job is honest: build something messy, get it in front of paying users, and only then apply the algorithm to what survives. This is where the planning fallacy does most of its damage. Founders convince themselves they have eight productive hours in a day, then plan four meetings, six deep-work blocks, and a strategy session into those hours. By Tuesday the plan is wreckage. The algorithm becomes a checklist against the wreckage — which requirement has no named owner, which step can be deleted entirely, which task is being optimized when it should have been killed, which cycle is running too slowly, and which is being automated before it was understood.
Step four — accelerate cycle time — gets misread as a productivity hack. It is not. It is a learning hack. Every iteration is a chance to be wrong faster and correct sooner. Five hundred AI-generated logos will not produce a great logo. Five thoughtful logos done by a designer might. But five hundred attempts give you five hundred chances to eliminate what is not working, which is a different and equally valuable thing. The dangerous version of this step is producing output for the feeling of output. Speed without a hypothesis is just heat.
Automation comes last for a reason. Musk has been public about the fact that he tried to automate the Tesla factory first and nearly broke the company doing it. Automating a process that has not been deleted, simplified, and timed produces a faster version of the wrong work. Founders attempting this in 2026 with AI tools face a sharper version of the same trap — the tooling is cheap enough to automate anything, including the steps that should not have existed in the first place.
The clearest warning sign that the algorithm is being run backwards is a team that keeps building tools to manage complexity instead of removing it. The five steps are not a process diagram. They are an order of operations. Question the requirement, delete the step, simplify what remains, accelerate the loop, and only then automate. Founders who run the sequence in reverse build elegant systems for work that should never have existed.
Watch the Full Episode on the Musk Algorithm below:
Follow us to watch live on YouTube and LinkedIn or listen to episodes on Apple Podcasts and Spotify.