The Cost Reality
Author/Source: Chris Blattman, claudeblattman.com
Key Ideas¶
- The full AI productivity stack costs approximately $140/month (~$1,700/year): Claude Max ($100), ChatGPT Plus ($20), Wispr Flow ($10), Granola ($10).
- For professors with research budgets, this is equivalent to about 7 RA hours/month at $20/hour -- easily justified if it saves more time than that.
- A minimum viable setup for graduate students costs ~$20/month: Claude Pro plus free built-in dictation and Zoom transcription.
- Conservative ROI: 5 hours/month saved at $40/hour = $200/month recovered. Realistic with a configured system: 10-15 hours/month.
- There is an honest upfront time investment to learn the tools and build the system; ROI is real but not instant.
- Important data policy distinction: Claude Code (API) conversations are not used for model training; Claude.ai (browser) is a consumer product with a different policy.
- A recommended four-week ramp-up sequence starts with just the chatbot and adds tools progressively.
Summary¶
This article provides a frank assessment of what an AI productivity workflow costs, who should pay for it, and how to think about return on investment. The full stack the author uses -- Claude Max for the primary AI tool and Claude Code, ChatGPT Plus for deep research and web search, Wispr Flow for dictation, and Granola for meeting transcription -- runs about $140/month. The article breaks down who should adopt the full stack (professors with research budgets, organizations and NGOs) versus who should start smaller (graduate students, early-career researchers).
The ROI framework is straightforward: if the system saves 5+ hours/month of administrative work and your time is worth $40+/hour, the math works. The author notes that a configured system realistically saves 10-15 hours/month, but emphasizes the honest caveat that reaching that level requires a significant upfront time investment. The article also flags an important distinction between Claude Code's API data policy (not used for training) and Claude.ai's consumer policy, which matters for researchers handling sensitive data.
Relevance to Economics Research¶
The cost analysis maps directly to academic budget decisions. At $1,700/year, the subscription cost is equivalent to a modest research expense that can often be charged to grants. The comparison to RA hours provides a concrete benchmark that any PI can evaluate. The minimum viable setup at $20/month makes the entry point accessible even for unfunded graduate students. The data policy distinction is particularly important for economists working with confidential data (CRSP, Compustat, administrative records) who need to understand what information is sent to AI providers.