Interview: Brett Bliss Dustin — 2026-05-01
Key Themes
Validation-first discipline vs. builder instinct. Brett’s central message was that Dustin and Bliss are at risk of the classic founder trap: designing a solution in search of a problem. He distinguished between the Andy Radcliffe “build-first” school and the customer-discovery-first school, arguing that the latter is far more appropriate for a company without a pre-existing technological breakthrough. His framing of “mini learning loops” rather than parallel tracks was a useful conceptual reframe — build only to answer a specific question, then stop.
The digital twin hypothesis needs interrogation, not execution. The team has made an unvalidated leap: that the oil & gas financialization model (hedging, insurance, benchmarks) maps to semiconductors, and that all of it rests on a digital twin data asset. Brett flagged this clearly — the question isn’t “is a perfect digital twin valuable?” but “why don’t these capabilities exist today, and what makes semis different from oil & gas?” That gap analysis must come from primary research, not secondary or AI-assisted analysis.
Earned secrets as the north star for interviews. Brett pushed toward surfacing insider, tribal knowledge that only practitioners hold — not general industry overviews. The PE investor anecdote (no visibility into supplier alternatives, no market-making mechanism, all phone calls and relationships) is an example of an earned secret surfacing organically. More of these are needed before any building.
Persona mapping is the immediate next step. Brett recommended mapping 15 specific role titles across the semiconductor value chain who would benefit from digital twin capabilities, then running one deep conversation per persona type. The goal is to find whose job or life gets concretely better — and to remain open to the possibility that the highest-value wedge is only 5% of the envisioned product.
Notable Quotes
- “Building is contagious. Building is easy. But you will rabbit hole so hard on building.”
- “You interviewed yourself and you said, my insight is this should exist and this is a big problem.”
- “Find the pain, find the problem, build conviction in that, solve it.”
- “Ask what am I trying to learn before building anything.”
Surprises
Brett’s framing of the Andy Radcliffe model as specifically for companies at the frontier of a technology inflection — not a general-purpose startup methodology — was a sharper critique than expected. The team’s self-awareness that they’ve been “designing a solution in search of a problem” was notably candid and suggests genuine intellectual honesty about where they are. The PE investor signal (no market visibility, phone-call-dependent capacity allocation) emerged organically mid-conversation and may be the strongest unvalidated signal to date.
Open Questions
- Why haven’t risk transfer mechanisms been built for semiconductors already? Who has tried?
- What specifically makes the semiconductor supply chain structurally different from oil & gas in ways that break the analogy?
- Who actually owns risk management today at a fab, an OEM, a distributor — and how do they do it?
- Is the digital twin the product, or is it the enabling infrastructure for a more specific wedge (compliance, benchmarking, insurance)?
- What is the MBA 2 team’s DoD-facing approach and what can be learned from their path?