Founding Product · Liminal · Temasek
Blink AI /
Project Genesis
01 — Problem
Building on Web3 was too slow.
Developers wanting to prototype decentralised applications had to wire up smart contracts, configure wallets, and navigate chain-specific infrastructure before writing a single line of product logic. Days of setup before any validation.
The hypothesis: if we could let developers describe what they want in natural language and generate a working dApp prototype automatically, we could collapse the feedback loop from days to minutes.
02 — Methodology
Human-centred design, from zero.
Before writing a line of code, I ran a full discovery process grounded in HCD — treating developer pain as the product constraint, not the blockchain stack.
01
Discover
400+ developer interviews. Personas, JTBD, pain point synthesis.
›02
Define
Hypothesis framing. Prioritised pain → opportunity space.
›03
Prototype
Rapid iteration. Behavioural analytics. Prompt-to-dApp MVP.
›04
Test
Live demos at EthDenver. 50% qualified conversion.
›05
Scale
GTM foundation. Waitlist. Ecosystem partnerships. Seed raise.
HCD research synthesis — context setting, divergent ideation, affinity mapping, prioritisation
03 — What we learned
The signal from 400+ conversations.
Pain #1
Setup friction was the biggest barrier to prototyping — not the actual dApp logic. Developers wanted to skip to the interesting part.
Pain #2
Most Web3 tooling was built for production, not experimentation. There was no "sketch mode" for on-chain product ideas.
Insight
The highest-value workflow wasn't code generation — it was idea-to-testable-prototype in a single session, sharable with non-technical stakeholders.
Pivot
Early feedback shifted our framing from "developer tool" to "product prototyping platform" — broadening the addressable persona set meaningfully.
04 — Outcome
Concept to seed funded.
I left the team after securing the seed to pursue the next venture — taking the same methodology into Laguna.