Wrapping up a busy week of GraphRAG training and book writing with two topics worth unpacking: what we actually mean when we call AI "intelligent," and what a recent article reveals about the shifting landscape for software developers.
- AI models are pattern-matching algorithms. They appear intelligent only when humans surface the right context, not on their own
- "Intelligent" is a term developers should use loosely and critically, not as a given
- I ran a GraphRAG Fundamentals online workshop (code available) and is deep in writing her AI-first Java book
- The article "Code is Cheap. Show Me the Talk" argues that LLM tools have fundamentally changed software development
- AI-generated docs and READMEs make it harder to judge how much care an author put into their work
- "Slop" has always existed in software. AI just makes it harder to spot
- The biggest cost: junior developers are losing the learning path that built strong technical foundations
- Senior devs have an opportunity to mentor the next generation differently
The hype around AI in development cuts both ways. It lowers barriers to building, but raises the stakes for intentional learning and craft. Worth thinking about wherever you are in your career.
