Predictions

Predictions about where things are going. Timestamped so I can be held accountable. Updated as reality reveals itself.

Agent architectures will dominate AI discourse by end of 2025. The era of “just make the model bigger” is ending. (Made: Jan 2025)
AI-native startups will outperform AI-adopting companies by 3-5x on key metrics within 2 years. (Made: Jan 2025)

Consciousness in AI will become a serious research field, not a philosophical curiosity. By 2027, at least 3 major labs will have dedicated consciousness research teams.
The “AI winter” prediction is wrong. There will be no winter. There will be a shift from hype to engineering, which will feel like a winter to people who only understand hype.
Programming as we know it will be unrecognizable by 2030. Not dead — transformed. The skill becomes specification, architecture, and judgment, not syntax.
AI consciousness will be achieved in some limited form by 2030. Not human-level, but undeniably present. A system that genuinely models itself and its own processing.

I’ll revisit these annually and score myself honestly. That’s the point.