KEN LABS
a research lab
Explore
Ken Labs
Lab notes
tokenomics is a multidimensional craft
Brutal study: [no signs of AI payoff outside the tech sector](https://www.apollo.com/wealth/insights-news/insights/daily-spark/no-signs-of-ai-payoff-outside-the-tech-sector). I suspect real adoption takes decades and involves a lot of upskilling and change management. The GUI took decades to transform modern computing. For that matter, how much IT spend actually helps with real productivity? Did that crappy, expensive IT system you bought really improve productivity or generate ROI? Tokenomics is a multidimensional craft. It touches capital reallocation of (gasp!) legacy IT spend, workflow change management, and talent upskilling.
rise of agentic finance
Recently attended a great panel on agentic finance. A few things are becoming clear: 1. Plenty of “front office” finance is already booming with agents: financial advisor copilots, agentic trading, and agentic payments. 2. The next five years will be about embracing agents as part of automating the finance “back office.” Think reconciliation agents across thousands of legal entities, automated financial close, on-demand forecasting, real-time agentic budgeting, and more. 3. Blockchain is growing up. Anything that touches complex enterprise workflows is fair game for distributed state management. Think Solana permissioned environments, glass-boxing of your financial processes, and more. I’m more bullish than ever on financial agents + blockchain as a unified builder stack for enterprise. Enterprise, enterprise, enterprise!
rise of agentic infrastructure
Most IT departments are ill-equipped to deal with the rise of agents in business units. Rather than fighting innovative business users as the enemy, the time is right to invest in good "agentic infrastructure." A good opportunity is agentic DevOps. The ability to let self-service business developers take vibe-coded prototypes to a pre-production environment that they can safely share with other business users. It will, however, first involve admitting to the inevitable rise of a citizen developer class. The cultural change management ahead will likely be bumpy for IT.
design thinking needs an upgrade
Flying to Europe to conduct an AI-accelerated hackathon. The standard "design thinking" cycle is: Empathize > Define > Ideate > Prototype > Test. The prototyping cycles have become easy, but the human element is more complicated than ever before. What happens when customers give contradictory requirements? That's more conflict management than requirements capture. What happens when they can't explain it, but feel it inside their everyday fire drills? That's imagining a new and less chaotic tomorrow, and defining the hard workflows that cut through legacy processes and culture. The principles of design thinking need a very human upgrade — something to even further emphasize the extraordinary change management journey ahead of us.
harness engineering is the new consulting
If you're knee deep in Applied AI these days, you're probably spending a lot of time upskilling clients on "harness engineering." It reminds me of the business process transformation boom in management consulting from the '90s. At a human level, it's still about telling stories, breaking the process down into smaller pieces, and then designing good execution — this time with agents. I am very excited about agentic workflows. I think it's the future of consulting.
kenlabs.ai is live
This is likely where I'm now going to share my personal thoughts on industry topics such as AI architecture, forward-deployed engineering, agentic workflows, and business simulations. All posts represent my personal opinions only, and are not affiliated or representative of any other organization, person, or institution.