Comment l'IA est devenue plus chère que ceux qu'elle remplace

Comment l'IA est devenue plus chère que ceux qu'elle remplace

by Yassine Sdiri

Humanities, Social Sciences & Thought AI & Society UBDigital and Information technology: general topicsUBJDigital and information technologies: social and ethical aspects
🎙 Yassine Sdiri 👥 273K 📅 August 30, 2026 ⏱ 18 min 👁 335 📄 expert opinion 🧭 2026-08-30
Available in: English (current) Français

Keywords

AI costJevons paradoxAI agentsreturn on investmententerprise AI

Summary

The video discusses why AI costs are skyrocketing despite falling per-unit prices, using the Jevons paradox as a framework. It cites examples like Microsoft and Uber exhausting AI budgets, and explains that increased efficiency leads to increased consumption. The video highlights the shift from simple Q&A to autonomous AI agents, which consume far more resources due to iterative processing and context accumulation. It also addresses the difficulty of measuring AI ROI, with only 1 in 10 companies reporting measurable returns, and the high rate of project abandonment (42% in 2025). The speaker, an AI consultant, argues that the real cost lies in poor implementation and lack of training, not the technology itself. He contrasts failed projects with successful ones like Carrefour and Orange, which invest heavily in training. The video concludes that the most expensive word is ‘almost’ (à peu près), as near-success often leads to hidden costs and project failures, and that companies that rush to adopt AI without understanding it are the ones that lose.

167 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides a valuable perspective on the hidden costs of AI adoption, particularly the Jevons paradox and the shift to agentic AI. The argument is well-structured, moving from specific examples to a general principle, and then to practical implications. However, the argumentation relies heavily on anecdotal evidence and the speaker’s own consulting experience, which may introduce bias. The video does not provide a balanced view of the benefits of AI, focusing almost exclusively on costs and failures. The use of the Jevons paradox is insightful but not fully explored, and the claim that AI agents consume 1000x more than simple queries is presented without a clear source.

Scientific Rigor, Source Quality, Title Accuracy

The video cites several studies and reports, including one from ‘Lut’ (likely a mispronunciation of ‘Gartner’) on ROI, and a Stanford study on AI agent costs. However, these references are not detailed enough to verify, and no direct links are provided in the description. The video also mentions a study on AI agent autonomy (METR) and a statistic about 42% of AI projects being abandoned, but again without specific citations. The title is accurate and not misleading. The video includes a promotional segment for the speaker’s AI training school, which is clearly marked as such and does not detract from the content.

225 words

Title / Content Match

The title accurately reflects the video's central thesis that AI costs can exceed the costs of the human labor it replaces, though the video also discusses broader issues of implementation and measurement.

Quality & Reliability

6/10

The video presents a coherent argument with some references to studies and reports, but lacks detailed citations and relies heavily on anecdotal evidence and personal experience. The central claims about cost overruns and the Jevons paradox are plausible but not rigorously substantiated.

Chapters

Cited Sources

Concurring Sources

  • Gartner: Only 10% of AI projects deliver measurable ROI — The video cites a consulting firm (likely Gartner) stating that only 1 in 10 companies report measurable ROI from AI, which aligns with this report.
  • METR: Measuring AI agent autonomy — The video references an independent organization (METR) that measures AI agent autonomy, which is consistent with this source.

Dissenting Sources

  • McKinsey: The state of AI in 2025 — McKinsey reports that many companies see significant value from AI, with a majority reporting revenue increases, which contrasts with the video's emphasis on failures and cost overruns.

Contribution & Novelties

The video offers a clear and accessible explanation of the Jevons paradox applied to AI, and highlights the often-overlooked cost drivers of agentic AI, such as context accumulation and unpredictable pricing. It also emphasizes the importance of training and implementation methodology over raw technology. The video’s main contribution is to reframe the AI cost debate from a simple price-per-token issue to a broader economic and organizational challenge.

Pour aller plus loin :

  • Jevons paradox — The economic principle that increased efficiency leads to increased consumption, which is central to the video’s argument.
  • AI agent — The concept of autonomous AI systems that perform tasks without direct human supervision, which the video identifies as a major cost driver.
  • Return on investment (ROI) — The metric used to evaluate the profitability of AI investments, which the video argues is difficult to measure accurately.

141 words

Radar Profile

The radar profile shows moderate scores across all dimensions, with slightly higher scores for information quantity and technical level, reflecting the video's informative but not deeply technical nature. The lower scores for information quality and reliability indicate that while the video is engaging, it lacks rigorous sourcing and may overstate certain claims.

Reliability 5/10

💬 No comments were provided for analysis.