Insights on the future of finance operations.
Perspectives from the Segmental team on applied AI, enterprise workflows, institutional memory, and the evolution of finance workflows.
Archive
09 articles · JAN–MAR 2026
The Coding Agents' Generalization Problem
In 2026, coding agents aren't just autocomplete tools; they are teammates.
Why Big Tech can't win the Enterprise (and why Startups shouldn't touch Consumer)
There is a structural flaw in how Big Tech innovates, and it is the single greatest advantage for Enterprise Startups today.
Coding Agents are exhaustive, not smart. They write code like they get paid by the token.
I just spent 4 days debugging a random SegFault. The code was written by an Agent. The diagnosis was driven by an Agent.
LLMs are Semantic Giants, but Epistemic Infants. That is why your Agents are failing.
We need to have an honest engineering conversation about what a Large Language Model actually is, versus what we want it to be.
There are two types of AI Agents. Both are failing in the Enterprise for the same reason.
We need to stop using the word "Agent" as a catch-all.
The Tale of Two AIs: Why the "Demo" Never Matches the Reality
We are currently witnessing a massive schism in the AI community. There are now two distinct worlds of AI enthusiasts, and they are drifting further apart every day.
The $100 Million Mistake: Why Generic LLMs and Agents Crash in the Enterprise
I recently witnessed the AI hype train crash into the wall of enterprise reality.
Why Enterprise Agents are stuck at "Co-pilot" (and won't reach "Auto-pilot" anytime soon).
We are being sold a vision of the Agentic Future where AI autonomously navigates our workflows, fixes our bugs, and closes our tickets.
AGI won't be achieved by scaling compute. It will be achieved by structuring the "Tribal Nuance" we are too lazy to capture.
We are entering the "trough of disillusionment" with Enterprise AI. The easy demos are over. The wrapper startups are churning.