🚨 NOTICIAS IA 🚨: Esto es lo que me ha sorprendido de GPT-5.2 🔥🤯

🚨 NOTICIAS IA 🚨: Esto es lo que me ha sorprendido de GPT-5.2 🔥🤯

by John Hernández

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 John Hernández 👥 725K 📅 December 15, 2025 ⏱ 33 min 👁 102K 📄 news review 🧭 2026-08-03
Available in: English (current) Français

Keywords

GPT-5.2AI benchmarksExcel generationAI speedOpenAI

Summary

In this video, John Hernández reviews the latest AI news, focusing on the release of GPT-5.2. He shares his hands-on experience, highlighting its impressive ability to generate complex Excel files with thousands of rows, but also its significant slowness, which he sees as a paradigm shift in how users interact with AI. He discusses the evolution of benchmarks, noting a dramatic jump in ARC-AGI scores. He also covers Google’s SIMA 2 agent with self-improvement capabilities, Disney’s partnership with OpenAI for Sora, the launch of ChatGPT Go in Latin America, and various other updates. He mentions the integration of Photoshop in ChatGPT, the viral ’nano banana’ feature, and the deployment of Grok in El Salvador. He also touches on the first LLM trained in space and US regulatory moves. The video includes a sponsored segment for Hostinger. Hernández concludes that GPT-5.2 is extremely capable but slow, and that users will need to adapt to a more delegated workflow.

157 words

Critical Evaluation

The video offers a comprehensive and timely overview of recent AI developments, with a particular focus on GPT-5.2. The creator’s hands-on testing provides valuable insights into the model’s capabilities and limitations, especially regarding its speed and performance on complex tasks like Excel generation. The discussion of benchmark improvements is informative, though the reliance on a single benchmark (ARC-AGI) could be seen as limited. The argumentation is generally solid, but the subjective nature of the review and the lack of independent verification of some claims (e.g., the exact number of rows in the generated Excel) slightly reduce its scientific rigor. The sources cited are mostly links to the creator’s own courses and tools, which may introduce bias. The inclusion of a sponsored segment is clearly disclosed but does not detract from the content. The title accurately reflects the content, and the video is well-structured with chapters. Overall, it is a valuable resource for staying updated on AI news, but viewers should seek additional sources for a more balanced perspective.

168 words

Title / Content Match

The title accurately reflects the content, focusing on surprising aspects of GPT-5.2 and other AI news.

Quality & Reliability

7/10

The video provides a personal hands-on review of GPT-5.2, with concrete examples and benchmarks, but relies heavily on subjective impressions and lacks in-depth verification of all claims. The creator is an established AI educator, but the content is a news roundup with a promotional segment.

Chapters

Cited Sources

Concurring Sources

Dissenting Sources

  • Community feedback on GPT-5.2 speed — Some users in the comments express dissatisfaction with the slowness, while others appreciate the quality, indicating a mixed reception.

External References

Contribution & Novelties

The video provides a practical, hands-on evaluation of GPT-5.2’s capabilities, particularly its ability to generate complex, multi-sheet Excel files with thousands of rows, and its significant slowness, which the creator argues represents a shift towards delegated AI tasks. It also highlights the rapid progress in AI benchmarks, with ARC-AGI scores jumping from 13 to 54 in a month.

Pour aller plus loin :

  • ARC-AGI benchmark — The benchmark referenced for measuring AI reasoning capabilities.
  • GPT-5.2 technical report — Likely a paper on GPT-5.2, though not explicitly cited in the video.
  • SIMA 2 from Google DeepMind — Google’s agent with self-improvement, mentioned in the video.
  • OpenAI’s Sora — Video generation model, relevant to the Disney partnership.
  • ChatGPT Go — The mobile app launched in Latin America.

125 words

Radar Profile

The radar profile shows high scores in information quantity and quality, with moderate technical depth and reliability. This indicates a well-informed and detailed news review, but with room for more rigorous sourcing and technical analysis.

Reliability 7/10

💬 Positive and enthusiastic: viewers appreciate the detailed news coverage and the creator's hands-on testing, though some express concerns about GPT-5.2's speed and a few disagree with political opinions. Overall, the sentiment is favorable, with many thanking the creator for the content.