Here’s How I Pick the Right AI for Jobs that Matter: My 5 Prompts + a ChatGPT 5.2 vs. Claude Opus 4.5 vs. Gemini 3 Deep Dive

Source: Nate’s Newsletter Date: 2025-12-15 URL: https://natesnewsletter.substack.com/p/grab-the-5-prompts-i-use-to-discover

Summary

Nate argues that asking “which AI model is best” is the wrong question — model capability is too granular and task-specific for blanket comparisons to hold. The real skill is identifying your specific work context and testing models against concrete criteria: spreadsheet handling, coding throughput, output style. He runs this framework across ChatGPT 5.2, Claude Opus 4.5, and Gemini 3.

Implications

Vendor positioning thread. Benchmark-based model comparisons flatten meaningfully different capability surfaces. As models diverge on specialized tasks while converging on general ones, the “which is best” narrative becomes primarily a marketing artifact rather than a useful enterprise signal.

AI economics thread. Cost-benefit varies dramatically by workload type — a model that’s cheap and good enough for drafting may be wrong for code generation or data analysis. Organizations running undifferentiated model choices are likely misallocating inference spend.

Agent product strategy thread. If model selection requires task-specific evaluation, agent orchestration layers that assume a single best model will underperform against those that route by task type.

Watch: Whether enterprise procurement moves toward task-based model evaluation frameworks or continues relying on vendor-published benchmarks through 2026.