The Plato Paradox and AI
Plato Paradox
I didn't study Computer Science at university. Instead I studied Classics - Latin and Greek. This meant translating texts from a variety of authors and genres, one of which was Plato.
I didn't study Computer Science at university. Instead I studied Classics - Latin and Greek. This meant translating texts from a variety of authors and genres, one of which was Plato.
Yesterday marked the launch of the latest release of MCP, the protocol championed by Anthropic and a founding member of the Agentic AI Foundation back in December 2025. MCP had been on my radar since the middle of 2025, when I integrated into a proof-of-concept project. When the last version of the spec launched in November 2025 (2025-11-25 spec) I did in-depth research on the new features. And I've been tracking the latest release for many months, particularly because of the potential of Skills over MCP. So it's great to see the launch.
Without a doubt AI has revolutionised software development and is here to stay. But we're constantly adapting to the change, and I'm reminded of the Tuckman model. However, it's worth also remembering where the pain points of the past have been beyond just writing code.
I’ve been using AI for over two years, progressing from clever auto-completion through vibe coding to agentic engineering. Recently I iterated for a couple of weeks over a prototype application. The biggest problems came down to a single word, a term I’m found myself using more than any other over the last few weeks. And it’s not exclusive to a single project. It’s a term that’s been used across coding work, research work, everything.