Introducing Claude 4—here’s where it shines and how to use it alongside ChatGPT and Gemini

Source: Nate’s Newsletter Date: 2025-05-23 URL: https://natesnewsletter.substack.com/p/introducing-claude-4heres-where-it

Summary

Nate’s Newsletter covers early access impressions of Claude 4, framing it as distinctly valuable for “power users” in ways that ChatGPT and Gemini are not. The piece promises concrete breakdowns of coding and writing performance after deeper testing, with the overall thesis that model selection now meaningfully matters at the high end — all frontier models are capable, but the capability gap is real and task-dependent. Full analysis is behind a paywall; the public portion establishes the framing without detailing the specific comparisons.

Implications

• Feeds the model differentiation thread — if the Claude 4 framing holds up (specialised value for heavy users rather than general parity), it creates a stable segmentation argument that cuts against “all models converge” positions.
• Relevant to any team choosing a primary model for agentic coding work: if Claude 4 is genuinely differentiated on long-horizon coding tasks, the switching cost case gets stronger at rollout.
• Watch for follow-up posts from Nate with actual benchmark tasks — that’s where the “immediate practical applications” claim will be testable.