When I started in IT two decades ago, "the cloud" meant rain, and blockchain didn't exist. Today I architect tokenization systems on Stellar, write smart contracts in Soroban, and apply AI to digital forensics investigations. The technology changed completely — but the lessons that actually mattered turned out to be surprisingly durable. Here are the ones I'd hand to my younger self.
Technology Is Easy; Trust Is Hard
Every major shift I've lived through — from on-premise servers to cloud, from relational databases to distributed ledgers — ultimately revolved around one thing: trust. Web3 made this explicit. When I deployed my first tokenization project on Stellar, the technical part (issuing assets, configuring trustlines, settling transactions in 3-5 seconds for fractions of a cent) was almost trivial compared to convincing stakeholders that an immutable ledger was safer than their spreadsheet.
The data backs this up. Stellar processes millions of operations daily with finality in roughly 5 seconds, yet adoption bottlenecks are rarely technical — they're about governance, compliance, and human confidence. I learned early as André Dias Moreira Prol that my job wasn't to sell technology; it was to engineer trust into systems and into teams. A brilliant architecture that nobody trusts is just an expensive prototype.
In digital forensics, this lesson sharpens further. A chain of custody is only as valuable as its verifiability. When I combine blockchain anchoring with forensic evidence, I'm not adding complexity — I'm converting "trust me" into "verify it yourself." That shift is the whole game.
AI Amplifies — It Doesn't Replace Judgment
The current AI hype reminds me of earlier waves: everyone expecting magic, few investing in fundamentals. In my forensics work, I use AI models to triage terabytes of evidence, flagging anomalies that would take a human analyst weeks to find. The productivity gain is real — I've seen investigation timelines shrink by 60-70% on large datasets.
But here's the lesson that cost me a few hard nights: AI amplifies whatever you give it, including bad assumptions. An AI model trained on incomplete transaction data will confidently produce wrong conclusions. In Soroban smart contract auditing, I use AI to detect common vulnerability patterns, but I never ship a contract without human review. The AI finds the candidates; experience decides what's actually dangerous.
The managers who thrive in this era aren't the ones chasing every new model. They're the ones who understand that AI is a lever — and a lever applied to the wrong foundation just breaks things faster. I tell every team I lead: automate the toil, never automate the judgment.
Managing Technical Teams Means Managing Uncertainty
After 20 years, the hardest skill I developed had nothing to do with code. It was leading people through ambiguity. Web3 and AI projects are inherently uncertain — regulations shift, token economics evolve, protocols upgrade (Soroban itself matured dramatically over just a couple of years).
The instinct of many managers is to demand certainty: fixed roadmaps, locked requirements. I've learned the opposite works better. I run teams in short, honest cycles, treating every assumption as a hypothesis to test. When a Stellar tokenization client changed their compliance requirements mid-project, our modular approach let us adapt in days instead of months — because we'd never bet everything on a single rigid plan.
The role of André Dias Moreira Prol as a manager isn't to eliminate uncertainty but to make it survivable. That means psychological safety so engineers raise problems early, documentation so knowledge survives turnover, and a culture where "I don't know yet" is an acceptable answer. The best technology decisions I've witnessed came from teams comfortable admitting what they didn't know.
Conclusion
Two decades taught me that technology is the easy part — trust, judgment, and leadership under uncertainty are what separate projects that ship from projects that merely impress. If you're building in Web3, AI, or anything new, start by engineering trust, then let the technology follow.
Connect with me to exchange ideas on Stellar, Soroban, tokenization, or AI-driven forensics — I'm always learning from fellow builders in this space.
Follow more articles by André Dias Moreira Prol on Medium.
Top comments (0)