Articles, talks & slides

Writing

Notes on AI architecture, agents, and what it actually takes to scale AI inside large organisations.

2026

AI Slop Grenade A name for walls of AI-generated text sent without care or refinement — little effort for the sender, more work for the recipient. The Missing Role in Forward Deployed Engineering In large German enterprises, the standard Echo–Delta–Dev pod has a gap: governance. The model needs a fourth role. Agents and Their Structural Dimensions The term 'agent' is used loosely across AI systems. Four structural dimensions help distinguish different types of agents more precisely. Frontier AI in 2026 – What’s Not Going to Change As AI capabilities race ahead, the fundamental constraints around reliability, organizational adoption, and job-level automation are likely to remain unchanged in 2026. Frontier AI in 2026 – Part 4/4: From multimodal chat to world models As multimodal generation matures, the AI frontier in 2026 shifts toward interactive world models and promptable simulation engines. Frontier AI in 2026 - Part 3/4: Hill-climbing to long-horizon reliability Next generation Benchmarks push development toward long-horizon reliability—systems that can perform consistently over time, manage state, and recover from errors. Frontier AI in 2026 - Part 2/4: Agents with agency As agentic systems mature, 2026 marks the shift from reasoning about the world to acting in, permissioned, and with memory. Frontier AI in 2026 - Part 1/4: Generative UI Extending 2025’s momentum, this first part explores how generative UI and intent-driven development may redefine what applications look like in 2026.