A Deep Dive into OpenAI Deep Research: Can You Feel The AGI? (Also Why Is Everything “Deep” Now FFS)

Source: Nate’s Newsletter Date: 2025-02-04 URL: https://natesnewsletter.substack.com/p/a-deep-dive-into-openai-deep-research

Summary

A comparative review of OpenAI Deep Research, DeepSeek R1 with Internet Search, and Google Deep Research, using a real research question (AI automation impact on industries by 2030) as the test case. Published February 2025, the piece also pokes at the tech industry’s “deep” naming trend while examining whether these tools can handle genuinely complex research synthesis.

Implications

Agent-product positioning thread. Deep Research products represent a specific category of agentic behavior — long-horizon, multi-step retrieval with synthesis — that is distinct from single-turn chat. The February 2025 comparative moment is when this category became a recognizable product type rather than a feature.

AI economics thread. The irony Nate flags — humans asking AI to analyze AI-driven job displacement — captures the recursive nature of the current moment. The research tools analyzing this question are themselves part of the displacement wave they’re describing.

Watch: Whether deep research capabilities commoditize (every major model vendor now ships some form) or whether differentiation emerges at the synthesis quality layer — and whether that matters to buyers who mostly use these tools for the first-draft research they’d otherwise outsource.