My AI Writing Compass
Everything I’ve Written About AI (So Far)
A roadmap through my growing collection of AI articles. Find the piece you enjoy, discover connections between ideas, and come back to see where this journey is heading next.
Modern AI Systems
The tools that are changing everything, right now
This is where I dig into what’s actually under the hood. Large language models, transformer architectures, the mechanics of how these systems generate text — the technologies reshaping how we interact with computers. I write about prompt engineering not as magic tricks but as a new kind of programming. You’ll find technical depth without the jargon overload, honest assessments of what these systems can and can’t do, and the occasional rant about overhyped capabilities. If you build with AI or just want to understand how it works, this is your starting point.
- Probably, Approximately, Shipped (2026/09/02)
- Pre-Training Gives LLMs Their Capability. Post-Training Gives Them Their Behavior. (2026/06/02)
- System Prompts, Commands & Skills: How AI Personalization Grew Up (2026/03/26)
- State of the AI: January 2026 Report (2026/01/12)
- Crafting Prompts That Click: A Framework-Driven Guide (Plus Your Ready-to-Use Prompt Generator Prompt) (2025/11/05)
- Model Choice in the GPT-5 Era: Spend Smart, Ship Fast, Defer the Fancy (2025/09/16)
- From GPT-3.5 to Now: 10 Breakthroughs that Changed the AI Game (2025/07/08)
- From Google to Perplexity: Rethinking Search in the Age of Answers (2025/05/21)
Thinking About Intelligence
The big questions that keep me up at night
What does it mean for a machine to “understand” something? When ChatGPT explains a joke, is it getting the humor or just pattern-matching? These pieces explore the fuzzy boundary between intelligence and imitation. I draw connections between how humans learn language and reason through problems versus how AI systems process information. It’s part cognitive science, part philosophy, part detective work — trying to figure out what’s really happening when we talk to machines that talk back.
- Too Smart to Be Right: Why Fluent Language Isn’t a Sign of Good Reasoning (2026/08/16)
- No Intelligence Without Illusion (2026/04/27)
- Beyond Imitation: The Reinforcement Revolution (2026/04/11)
- Explain Like I’m Human: The Reality Check We Need About AI Explainability (2025/10/28)
- Reasoning Models Are Just Internalized Prompt Engineering: How AI Learned to Question Itself (2025/10/13)
- Truth, Trust, and Hallucinations (2025/07/23)
- Modern AI Doesn’t Pass the Turing Test. It Outsmarts It. (2025/05/14)
AI & Human Work
Where the rubber meets the road
Get André Bergholz’s stories in your inbox
Join Medium for free to get updates from this writer.
This is where theory meets Monday morning. How do you actually use these tools without them taking over your creative process? What happens when entire workflows get automated overnight? I share experiments, failures, and small victories from incorporating AI into writing, design, research, and planning. You’ll find practical frameworks for working alongside AI, stories from people whose jobs have changed, and honest reflections on what collaboration with machines actually feels like day-to-day.
- An LLM Wiki for the Rest of Us (2026/11/10)
- API for Volume, MCP for Chat, CLI for Terminals (2026/09/17)
- My AI Workflow, One Year Later (2026/06/24)
- Slides Are a Build Artifact Now (2026/05/22)
- Deep Research vs. Broad Research: How AI turned Information Retrieval into Personalized Learning (2026/02/19)
- Your AI Has Assembled a Profile of You: Here’s What It Knows (2026/02/05)
- The Omnibar Pattern: A Simple Architecture for Real-World AI (2025/12/06)
- AI-First vs. AI-Native: The Distinction Your Company’s Future Depends On (2025/11/19)
- My ChatGPT System Prompt (with GPT-5 Update) (2025/08/14)
- When Intelligence Becomes Free: The Economics of AI Abundance (2025/06/25)
- How I Use AI Tools to Work Smarter (and Where They Still Fall Short) (2025/06/11)
History & Perspective
Why understanding where we’ve been matters
Before deep learning, there was symbolic AI. Before neural networks, there were expert systems. Before transformers, researchers spent decades building rule-based reasoning engines that could diagnose diseases and play chess. These pieces trace the threads connecting today’s breakthroughs to ideas from the 1960s, 70s, and 80s. Understanding AI’s past helps decode its present — and maybe avoid repeating old mistakes while we rush toward an uncertain future.
- The AI Race Has No Finish Line (2026/09/29)
- The Turing Award’s AI Story (2026/07/13)
- The Long Road to Overnight Success — Neural Nets from McCulloch-Pitts to Transformers (2026/03/12)
- The Idea That Started AI and the Researchers Who Built It (2025/12/17)
- Beyond Calculation: AlphaGo and the Birth of Machine Creativity (2025/08/04)
- Proclaimed Dead Twice: The AI Winters (2025/05/28)
What’s Next
This collection will keep growing. There’s a long queue of ideas on the desk, and the field is moving so quickly that every week reshapes the priorities. More experiments, fresh angles on emerging developments, and connections that help make sense of where this is all heading — updated here as the journey unfolds.
Each new article will link back here, and I’ll update this page regularly. Think of it as a living document, a way to navigate an increasingly complex landscape without getting lost.








