10 Comments
User's avatar
Deepak Jha from Quantum Mosaic's avatar

David — thanks! I want to add one layer you didn't cover that I think determines whether either path compounds or resets. A context layer captures WHAT people do. It doesn't capture WHY they decided to do it that way. When the CEO replaces half the exec team — as you rightly prescribe — the new leaders inherit processes without reasoning. They know what the company does. They don't know why, what was tried and abandoned, what assumptions must hold, or what would trigger a reversal.You also say: "Get really good at escalating contentious decisions to unblock progress." I've lived this. The problem is that in a 12-month death march with 50 hard calls per quarter, the reasoning behind Decision #7 is forgotten by the time Decision #22 contradicts it. Nobody notices because there's no record.

I've spent the last two years building the missing layer — what I call a Judgment Ledger. Version control for institutional reasoning: the decision, the criteria, the assumptions, the dissent, the reconsideration triggers. Not a knowledge base. Not a wiki. A system of record for the WHY behind consequential calls.Without it, both your paths create the same long-term risk: a brilliant 12-month refounding that produces zero retrievable reasoning for the team that inherits it 24 months later.Speed without reasoning capture is just faster amnesia. And both your paths run very, very fast.Thank you for writing this. It clarified my own thinking.

— Deepak Jha, Founder & CEO, Quantum Mosaic

Author, The Third Balance Sheet (https://quantum-mosaic.ai/book)

Vaibhav Maheshwari's avatar

I think a lot of the product discovery needs to be about content as a service, (CAAS). So when software is cheap and easily accessible, it is what content it is paired with that defines it's success, so when you are doing product discovery - you are doing pitch discovery for customers. The right pitch discovery will lead to right content which is where most of the PMs need to be focused on. Software still has a lot of potential to increase revenue for customers by embedding intelligence the right way, so please choose A and be the wartime CEO you were always meant to be.

David Mensah's avatar

You comment made me think about my companies Mission https://www.boxrewards.shop/mission

Is it relevant to what you are thinking Vaibhav

Box will allow this unique Product Knowledge content to be utilised for People and Agents through any tool they use to find exactly what they want.

gibson Jiang's avatar

SBC is too much in SaaS world

I agree

the rest is full of nonsense

Dylan's avatar

This is incredibly illuminating

Dhruv Jain's avatar

The "no middle path" framing cuts through a lot of founder noise. Most SaaS companies stay stuck between the two: not growing fast enough to justify VC multiples and not profitable enough to be defensible. Pick a lane early and stop confusing revenue traction with actual growth.

David Mensah's avatar

"too slow to deserve a premium growth multiple, too diluted to deserve a fortress multiple". Insightful writing....

Rameez MeeraSahib's avatar

Without reinvention SaaS companies would die eventually. You need to rewire the SaaS company to become the Agentic OS for whatever use cases they are focused on. You can read more about this approach here https://rmeerasahib.substack.com/p/personalized-software-is-coming-the?r=1n9qpn&utm_medium=ios

liujjw's avatar

Operating margin growth should be easier you literally have Anthropic OpenAI and the hyperscalers behind you subsidizing experiments in increased productivity

niyamic's avatar

For an AI-native product engineering studio, this is your "Atomic Unit." Instead of a bloated account team, you deploy a high-leverage pod that "refounds" a specific product line for the client - A small high-leverage, product-minded team.