Draft
src/content/notes/ai-rent.mdx
Summary
Found 3 gap(s): 1 explicit marker(s), 2 soft signal(s).
Gaps
1. Stanford AI Index / Epoch AI claim — lines 43–44 (soft signal: claim without source URL)
Original text:
"Research from Stanford's AI Index and Epoch AI shows that achieving a performance level equivalent to GPT-3.5 became over 280 times cheaper between November 2022 and October 2024"
Finding:
This claim is accurate and comes from the Stanford HAI 2025 AI Index Report, which draws on Epoch AI data. Cost fell from $20.00 per million tokens (November 2022) to $0.07 per million tokens (October 2024, Gemini-1.5-Flash-8B) — a 286-fold reduction. The benchmark used is MMLU score of 64.8 (equivalent to GPT-3.5).
Sources:
- (hai.stanford.edu/redacted) (main report)
- (hai.stanford.edu/redacted) (specific chapter with the cost data)
2. Sequoia $50B/$3B figures — line 76 (explicit marker: wrong source URL)
Original text:
Sequoia estimated that the AI industry spent $50 billion on the Nvidia chips used to train advanced AI models last year, but brought in only $3 billion in revenue. [source]((www.wsj.com/redacted) that's a 17:1 ratio (check)
Finding:
The linked URL is unrelated to the Sequoia claim — it's a WSJ article about Cognition Labs (a different startup). The correct source is the September 2023 Sequoia Capital article "AI's $200B Question" by David Cahn. The $50B/$3B figures are confirmed in that report (referring to 2023 spend). The ratio $50B/$3B = 16.67:1 (not exactly 17:1, but approximately correct).
Sources:
- (sequoiacap.com/redacted) ("AI's $200B Question", David Cahn, September 2023)
- (sequoiacap.com/redacted) (June 2024 follow-up, "AI's $600B Question")
3. Matthew Effect attribution — line 80 (soft signal: claim without source)
Original text:
The Matthew Effect is the phenomenon where individuals with initial advantages accrue further success, leading to a widening gap between the "haves" and "have-nots". Popularized by sociologist Robert Merton, the term originates from a verse in the Gospel of Matthew
Finding:
Robert K. Merton coined and popularized the term in his 1968 paper "The Matthew Effect in Science" published in Science magazine. The biblical verse is Matthew 25:29: "For to every one who has will more be given, and he will have abundance; but from him who has not, even what he has will be taken away." (RSV). Merton's paper is the primary academic citation for this concept.
Sources:
Note
These are research findings only. Maggie writes the prose herself — do not interpret these as drafted replacements.
Generated by Draft Research Agent · ● 440.4K · ◷
Draft
src/content/notes/ai-rent.mdxSummary
Found 3 gap(s): 1 explicit marker(s), 2 soft signal(s).
Gaps
1. Stanford AI Index / Epoch AI claim — lines 43–44 (soft signal: claim without source URL)
Original text:
Finding:
This claim is accurate and comes from the Stanford HAI 2025 AI Index Report, which draws on Epoch AI data. Cost fell from $20.00 per million tokens (November 2022) to $0.07 per million tokens (October 2024, Gemini-1.5-Flash-8B) — a 286-fold reduction. The benchmark used is MMLU score of 64.8 (equivalent to GPT-3.5).
Sources:
2. Sequoia $50B/$3B figures — line 76 (explicit marker: wrong source URL)
Original text:
Finding:
The linked URL is unrelated to the Sequoia claim — it's a WSJ article about Cognition Labs (a different startup). The correct source is the September 2023 Sequoia Capital article "AI's $200B Question" by David Cahn. The $50B/$3B figures are confirmed in that report (referring to 2023 spend). The ratio $50B/$3B = 16.67:1 (not exactly 17:1, but approximately correct).
Sources:
3. Matthew Effect attribution — line 80 (soft signal: claim without source)
Original text:
Finding:
Robert K. Merton coined and popularized the term in his 1968 paper "The Matthew Effect in Science" published in Science magazine. The biblical verse is Matthew 25:29: "For to every one who has will more be given, and he will have abundance; but from him who has not, even what he has will be taken away." (RSV). Merton's paper is the primary academic citation for this concept.
Sources:
Note
These are research findings only. Maggie writes the prose herself — do not interpret these as drafted replacements.