La era post-Fable: EE.UU. apagó la mejor IA del mundo

La era post-Fable: EE.UU. apagó la mejor IA del mundo

by Codemancers - Inteligencia Artificial

Humanities, Social Sciences & Thought AI — News & Applications UYComputer scienceUYQArtificial intelligence (AI)
🎙 Codemancers - Inteligencia Artificial 👥 2K 📅 June 16, 2026 ⏱ 48 min 👁 289 📄 expert opinion 🧭 2026-08-15
Available in: English (current) Français

Keywords

FableAnthropicAI regulationknowledge graphsDeepSeek

Summary

The podcast episode discusses the recent shutdown of Anthropic’s Fable model by the US government, which the hosts interpret as a consequence of Anthropic’s earlier fear-mongering and its strained relationship with the Trump administration. They analyze the role of Amazon as an informant, the limitations of Fable, and the implications for Europe, which relies on Mistral. The second half of the episode features a technical demonstration by Eric, who built a knowledge graph of the podcast’s 70+ episodes using DeepSeek V4, costing only €0.80 and 6 prompts. He explains the differences between SQL, NoSQL, vector databases, and graph databases, and argues that knowledge graphs are superior to vector databases for certain use cases. The hosts also discuss the concept of jailbreaking AI models, the problem of statelessness in models, and the division of labor between planning models like Opus/Codex and execution models like DeepSeek.

144 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides valuable insights into the geopolitical and corporate dynamics of AI development, particularly the power of the US government to control AI models and the precarious position of Europe. The hosts argue convincingly that Anthropic’s marketing of Fable as dangerous backfired, leading to its shutdown. They also offer a practical demonstration of building a knowledge graph, which is informative for developers. However, the argumentation is often speculative, especially regarding Amazon’s motives and the ‘betrayal’ of Anthropic, with limited evidence. The hosts rely on personal anecdotes and analogies (e.g., ‘The Sopranos’) rather than concrete data, which weakens the overall rigor.

Scientific Rigor, Source Quality, Title Accuracy

The video lacks rigorous sourcing; the hosts mention events and claims without providing specific references or links. The only resource provided is a link to their own knowledge graph project. The title is somewhat sensationalist, but it does reflect the main content. The hosts do not cite any external sources, which reduces the scientific credibility. The technical explanations are simplified and may not be fully accurate, but they are presented in an accessible manner. The adequacy between title and content is good, though the title’s claim of ‘best AI’ is subjective.

207 words

Title / Content Match

The title accurately reflects the main topic: the shutdown of Fable and its implications, though the subtitle about 'best AI' is subjective.

Quality & Reliability

6/10

The video presents a mix of factual claims about the shutdown of Anthropic's Fable model and personal opinions about AI regulation and geopolitics. The hosts provide some technical explanations but rely heavily on speculation and anecdotal evidence. The project demonstration is concrete but the sources are not cited in detail.

Chapters

Cited Sources

Concurring Sources

Dissenting Sources

  • Yann LeCun's X post — The video references LeCun's comment about reaping what you sow, but the exact post is not cited.

Contribution & Novelties

The video offers a unique perspective on the Fable shutdown, framing it as a consequence of Anthropic’s own actions and the power dynamics with the US government. It also provides a practical example of building a knowledge graph with DeepSeek V4, which is a novel application. The discussion on the limitations of vector databases compared to knowledge graphs is insightful.

Pour aller plus loin :

  • Knowledge graph — Provides background on knowledge graphs and their applications.
  • Retrieval-augmented generation — Explains RAG, a technique mentioned in the video.
  • Neo4j — Official site of Neo4j, the graph database used in the demonstration.
  • DeepSeek — Official site of DeepSeek, the AI model used for the project.

113 words

Radar Profile

The radar profile shows moderate scores across all dimensions, with a slight peak in quantity of information and a dip in reliability. This suggests the video is informative but lacks rigorous sourcing and technical depth.

Reliability 5/10

💬 No comments were provided for analysis.