Signal Decisions Log
Chronological record of every significant decision made in Signal conversations between the co-founders (Jan 14 – Apr 11, 2026). Answers the question: “Why did we decide X?“
2026-01-15 — Listening mode at networking events
Context: Shield Capital happy hour; didn’t want to reveal direction prematurely. Decision: Stay in listening mode; only dive in if someone raises supply chain/critical manufacturing organically. Quote: > “stay in listening mode and if someone happens to talk about supply chain/critical manufacturing to dive in”
2026-01-16 — Bias for action on Malchow meeting
Context: Uncertainty about whether semiconductors vs. broader supply chain was the right focus. Decision: Schedule the meeting anyway — lean into bias for action. Semiconductors might be a red herring, but gathering data is worth it. Quote: > “I wasn’t sure whether to schedule that mtg rn — semis could be a red herring vs our broader supply chain direction — but decided to lean into the bias for action”
2026-01-20 — Founders finish their own conversation threads
Context: Dustin proactively pushed an email thread Bliss had started. Bliss asked him to let her finish her own strokes. Decision: Each founder owns the threads they initiate; ask before jumping in. This is about growth, not territory. Quote: > “appreciate you being proactive in pushing the email thread with Milani forward but in the future could you first let me try to finish the stroke on the threads I start”
2026-01-21 — AI as intentional strategic lens
Context: Both founders are contrarian about AI hype, but recognized the moment. Decision: Stop being passive about AI and lean intentionally into it — both for the business model and for the mission framing. Quote: > “it does feel like we’re living through a once-in-a-generation type of technological moment, and we should be thinking about how to lean in”
2026-01-22 — 3-business-day follow-up cadence
Context: Debate about whether 2 or 4 days was the right follow-up window. Decision: 3 business days as the standard; schedule-send follow-ups in advance.
2026-01-29 — Protect first half of February for research sprint
Context: Dustin realized they needed deep internal work before external meetings would be valuable. Decision: Rearrange entire February calendar; push all external meetings to late Feb/early March to protect the Blank research sprint. Quote: > “I’ve rearranged my calendar for the month of Feb to push all my mtgs to later in the month or early March and protect first half of Feb for this Blank research exercise”
2026-01-30 — Schedule Malchow for post-sprint (Feb 24)
Context: Malchow is “a pretty substance-driven guy” — need more depth before the conversation. Decision: Schedule for Feb 24 only, after research sprint produces concrete knowledge.
2026-02-02 — Thesis must survive policy cycles
Context: WSJ article on repatriation agenda raised the question of over-indexing on political tailwinds. Decision: Avoid building a thesis that requires supply chain repatriation as a policy outcome. Ensure the business model holds across administrations. Quote: > “do we want to (explicitly or implicitly) bet our careers on the repatriation of supply chains? Probably not.”
2026-02-05 — Divergent AI research queries
Context: Risk of getting confirmatory analysis from identical prompts. Decision: Founders use different GPT instruction sets to produce divergent research outputs; share but don’t merge. Quote: > “I think it’s beneficial for us to have different instructions (so we’re not getting the same outputs)“
2026-02-07 — Log Eoin O’Doherty insight as gold-standard customer signal
Context: Organic conversation with robotics startup ops person who complained about needing supply chain intelligence + OSINT overlay. Decision: Record in Notion as first direct customer signal — “absolute gold.” Quote: > “This kind of first hand insight from prospective customers about hair on fire problems is absolute gold.”
2026-02-12 — Programs serve the business, not vice versa
Context: Evaluating H4D, Botha Chan, LLP, RDI simultaneously. Decision: Shoehorn programs into their objectives, not the reverse. First principles: build something, spend summer doing it, protect time during year. Quote: > “we shoehorn these programs into our objective, not vice versa (we use these programs to serve our business goal, we do not let these programs dictate our business goal)“
2026-02-12 — Apply for Botha Chan ($1,500/week)
Context: Summer funding opportunity. Decision: Definitely apply. Small but meaningful funding; confirmed March 20. Quote: > “Botha Chan is $1500 per week. Not a huge number obv, but beats a sharp stick in the eye! We should for sure do that!“
2026-02-14 — RDI as primary program vehicle
Context: Multiple program options in consideration. Decision: RDI (Research Driven Inspiration) selected as the structured summer program for deep research leading to a company.
2026-02-15 — Three-step progression is the clearest vision
Context: Debate between pitching compliance only vs. full vision. Decision: compliance → intelligence → finance is the most compelling narrative arc. Quote: > “the most compelling and exciting vision is compliance to intelligence to finance”
2026-02-21 — Narrow pitch for Botha Chan: compliance wedge only
Context: Claude’s first draft thrashed between compliance and full 3-step vision. Decision: Pitch only the narrow compliance wedge for the application; keep broader vision in appendix. Quote: > “I bias towards nailing the smaller idea rather than pitching the 3-step plan at half depth”
2026-02-22 — “TierZero” as submission name (not “Project TBD”)
Context: Botha Chan application required a company name. Decision: Use “TierZero” for the submission only — judges might not respond well to “TBD.” Not a permanent name change. Quote: > “Kind of like what GPT spit out… Not for the overall project, but just for this submission.”
2026-02-22 — Reject Claude’s fabricated interview data
Context: Claude generated fake customer interview results in the Botha Chan deck. Decision: Hard reject — reprompt to use only desktop research and real interview evidence. Never submit fabricated data. Quote: > “It basically made up the fact that we had conducted user interviews and cited bogus results from those convos.”
2026-03-01 — Altana/Interos are vulnerable
Context: Competitive research surfacing that existing compliance software has broken UX. Decision: Don’t be intimidated by incumbents’ websites; exploit the gap when actual users complain about broken software. Quote: > “Easy to get scared by Altana and Interos’ sleek websites… but only when you talk to actual users you figure out the software is broken and we have some seams to exploit”
2026-03-09 — Compliance wedge as strategic underpricing
Context: Eyck meeting + synthesis of prior advisor feedback. Decision: Use compliance product to undercut market → gain customer access → discover real problems and build the digital twin on top. Quote: > “We’re gonna build this digital twin by undercutting the market on compliance and then monetizing otherwise”
2026-03-09 — Relentless push for “who would buy this?”
Context: New discipline adopted after Eyck meeting surfaced unexpected customer segment (defense primes, transshipment). Decision: In every future advisor/expert conversation, relentlessly push toward “who might be interested in this problem.”
2026-03-17 — Dramatic press release tone approved
Context: Dustin asked Bliss to make the press release more narrative/dramatic. Decision: Embrace the drama — “it’s actually quite illuminating from the standpoint of why this product matters.”
2026-03-18 — Null hypothesis stays on semiconductors
Context: Post-McMaster meeting with adjacent ideas surfacing. Decision: Stick with semiconductors as the focus unless evidence specifically points to a pivot. Don’t let interesting adjacencies dilute focus. Quote: > “our null hypothesis/framing right now is semis so let’s stick with that unless we are shown evidence otherwise to pivot”
2026-03-18 — Don’t reactively update positioning after single conversation
Context: Temptation to rewrite OnePager after Joe/Ali insights. Decision: Gather more before codifying. Chew on ideas first. Quote: > “we should put a little more thought into that, chew on some of these ideas more before codifying them”
2026-04-05 — Keep Dasher and Blank meetings; use them for pressure-testing
Context: Founders back from break, no tactical plan ready. Decision: Don’t reschedule — use meetings to run through tactical plan and ask for advice in real time. Quote: > “Should we reschedule our meetings with Blank and Dasher? I think not — let’s use this time to run thru our tactical plan in a very tactical manner”
2026-04-07 — Blank canceled → tactical plan becomes re-engagement artifact
Context: Blank canceled Apr 7 meeting. Decision: Use the polished tactical plan document as the prompt to get back on his calendar. Quote: > “Good thing we know how to play this game 😈 We can use this doc as a prompt to get back on the cal”
2026-04-07 — Product roadmap at product level, not engineering level
Context: Dustin’s first draft dove into technical implementation details. Decision: Four-phase roadmap (Data Integration → Ontology → Agentic Query → MVP UI) stays at product level; no code libraries or packages.
2026-04-10 — Ontology building is the most important immediate task
Context: Start of Q2 execution mode. Decision: Build ontology on publicly available data now — it frames the edges of the problem, reveals data gaps, and lends credibility to all external conversations. Quote: > “the most important thing for us to do right now, and quickly, is build this ontology using currently available data”