Research

Dispatches from the front lines of agentic AI

Every engagement is a laboratory. Our research program is how Taller thinks out loud, how Frontier Engineers stay at the edge, and how the firm closes the gap between what AI promises and what production requires.

8th August 2026Steering the Talking Probability MachineA primer arguing that language model agents are stochastic systems, and reliable AI engineering is a control problem solved through feedback loops and correction mechanisms.PDF15th July 2026From Player Types to Professional Reward Ecologies: A Bartle-Informed Open-Science Validation Agenda for Software-Role ArchetypesA research agenda for mapping Bartle’s game player types onto software engineering roles.PDF23rd June 2026The Compressed Apprenticeship: AI, Task Recomposition, and the Economics of Human JudgmentHow AI is reshaping early-career engineering: which tasks compress, which expand, and where human judgment becomes more valuable.PDF27th April 2026Toward a Science of Intent: Closure Gaps and Delegation Envelopes for Open-World AI AgentsA framework for defining what AI agents are allowed to decide on their own and where they need to stop and ask.arXiv6th April 2026Coupled Control, Structured Memory, and Verifiable Action in Agentic AI (SCRAT)Using squirrel ecology as a comparative model, the paper proposes a formal framework for agentic AI that couples control, memory, and verifiable action to operate reliably under partial observability.arXiv17th March 2026Orchestrating Human-AI Software Delivery: A Retrospective Longitudinal Field Study of Three Software Modernization ProgramsA field study measuring what actually happened when three software modernization programs ran with hybrid human-agent teams.arXiv3rd March 2026The AI Identity GapAn analysis of 7,758 firms showing that AI capability is being built faster than organizations are redesigned around it, locating the real bottleneck not in technology but in unchanged workflows, decision rights, and operating models.PDF1st February 2026The Hybrid Workforce: How AI Agents Will Reshape IT StaffingResearch on IT staffing finds executives expecting roughly 30% faster delivery and higher margins from hybrid teams of humans and AI agents, examining the use cases already delivering impact alongside the barriers still slowing adoption.PDF

Newsletter

Stay at the frontier.

Research and perspectives on agentic AI from the Taller team, delivered to your inbox.