I was sitting at a corner table at a West Village private club with a tier-one venture capitalist—one of the smartest people I know, whose name I will leave off the record to protect his alpha—when the news broke. Melius, the highly anticipated startup founded by a cohort of ex-Ramp engineers, had just raised $20 million. But it wasn't the valuation that made my matcha catch in my throat. It was the methodology. They had successfully secured a massive Series A by unapologetically throwing their entire original software platform into the garbage.
Let us be incredibly clear about the magnitude of this flex. For the last year, Melius was supposedly building a platform to help marketers manage and optimize their ad spend. It was a perfectly fine idea. It was useful. It probably had dashboards. But from first principles, utility is a trap. If you build a SaaS tool that optimizes ad budgets, you are beholden to metrics, customer feedback, and the crushing gravity of measurable return on investment. By choosing to completely scrap that functional product and pivot into generative AI for creative campaigns, the Melius team has transcended software entirely.
It is time we admitted that shipping code is a legacy mindset. When you actually release a finished product, you are instantly punished by reality. Users find bugs. Churn rates materialize. The pristine potential of a pitch deck is violently compressed into the mundane reality of a daily active user count. What these ex-Ramp engineers realized—and what the broader ecosystem is still struggling to internalize—is that an abandoned codebase is mathematically perfect. It has no latency. It has no server costs. It exists purely as a testament to the founders' intellectual agility.
When a founder comes to me and says they have a working beta, I immediately price in their lack of imagination.
Fiedler gets it. During our off-the-record omakase dinner last weekend, he explained that the truly elite founders are the ones who can look a lead investor in the eye, casually mention that they just permanently deleted eighteen months of engineering work, and ask for twenty million dollars to wrap an API around a large language model. That level of sociopathic confidence is the ultimate moat. It signals to the market that your pedigree is so unimpeachable that you do not need the crutch of a minimum viable product to justify a nine-figure post-money valuation.
Consider the sheer elegance of the Melius pivot. The original roadmap was focused on ad spend optimization, a space defined by deterministic math and boring, accountable outcomes. Now, they are pivoting to building the tools that generate the creative assets. Do you understand what a monumental upgrade this is? They have moved from the dreary business of telling marketers where their money went, to the infinitely scalable business of generating synthetic slop to spend that money on. You are no longer optimizing a finite budget; you are generating an infinite volume of digital collateral that no human will ever actually look at. The logic is unassailable.
When I was building my own startup back in 2018—a decentralized protocol for enterprise mindfulness that we eventually pivoted into a boutique consultancy before quietly returning capital to our families—I struggled with the sunk-cost fallacy. I mistakenly believed that because we had written the code, we had an obligation to launch it. If only I had possessed the absolute sheer audacity of the ex-Ramp mafia. If only I had realized that the code was merely a physical manifestation of our limiting beliefs.
We looked at our working spend-management platform and realized it was violently restricting our ability to participate in the current hype cycle, which is a massive disservice to our shareholders.
It is a profound truth that most operators simply lack the stomach to internalize. The old Silicon Valley rewarded you for solving hard technical problems. The new, enlightened Silicon Valley rewards you for recognizing that the hardest technical problem of all is convincing a syndicate of seed funds that your brand-new pivot into generative AI is a fundamental architectural breakthrough and not just a frantic weekend side-quest built on top of an OpenAI endpoint. By destroying their first product, Melius proved they are unburdened by the past.
And let us talk about the ex-Ramp pedigree, because it is the foundational layer of this entire stack. When you have spent two years at a hyper-growth fintech decacorn, you are no longer bound by the laws of physics that govern regular, un-credentialed founders. You have been anointed. A regular founder who scraps their product is a failure undergoing a messy restart. An ex-Ramp engineer who scraps their product is a visionary orchestrating a strategic realignment of capital. The universe naturally bends to accommodate their epiphanies.
Some cynics on a certain orange-tinted message board are currently complaining that twenty million dollars is a lot of money to give a team that just explicitly demonstrated they will throw away everything they build. These people are fundamentally misreading the board. The massive funding round is not financing the new generative AI product. The capital is a reward for having the raw, predatory confidence to look at a completed optimization dashboard, declare it deprecated, and pivot to typing prompts into a text box.
The ability to set fire to your own deliverables while maintaining a straight face during a board meeting is the only metric that correlates perfectly with a successful public offering.
If you look closely at the new Melius direction, you can see the genius of avoiding measurable outcomes entirely. In the old model, if the ad spend management software failed to save a client money, the client would cancel their subscription. In the new generative AI model, if the synthetic campaign assets fail to drive engagement, the startup can simply blame the prompter, update the model weights, and charge for another generation cycle. It is a flawless business model constructed entirely out of plausible deniability.
We are witnessing the birth of a new asset class: the post-product startup. It is an enterprise that exists entirely in a state of quantum superposition, forever pivoting toward the bleeding edge of the hype cycle, unburdened by legacy customers or actual revenue. Melius is leading the charge, proving that the fastest way to achieve a billion-dollar valuation is to aggressively refuse to finish anything. The paradigm is shifting, and I, for one, cannot wait to see what they decide to throw away next.