The 6 reasons your work is hard — and which ones AI is automating this year + prompts to build your map

Source: Nate’s Newsletter Date: 2026-02-23 URL: https://natesnewsletter.substack.com/p/gemini-31-pro-broke-every-benchmark

Summary

Nate argues that “which AI should I use” is the wrong question — understanding what makes your work genuinely difficult should precede tool selection. He introduces a six-axis framework for diagnosing work difficulty, with different axes being automated on different timelines. The Gemini 3.1 Pro framing reveals a strategic signal: Google “underpriced its best reasoning model on purpose,” suggesting competitive positioning around accessibility rather than raw capability benchmarks.

Implications

• Enterprise adoption thread. The “evaluate whether AI output is actually good” framing identifies evaluation skill as the core competitive differentiator — not model selection. Organizations that can’t assess output quality will make poor model choices regardless of which framework they use to select them.
• AI economics thread. Google’s deliberate underpricing of its best reasoning model signals a market-share strategy over margin optimization — willing to sacrifice near-term revenue to win adoption at scale. This is the classic platform play: get the usage, monetize later. It pressures OpenAI and Anthropic’s pricing positions.
• Watch: Whether the six-axis work-difficulty framework becomes a widely adopted diagnostic tool, and whether Google’s underpricing strategy succeeds in shifting Gemini’s market position relative to GPT-5 and Claude.