ChatGPT Just Dropped 7 New Updates, FREE Chinese AI Models & More Use Cases!

ChatGPT Just Dropped 7 New Updates, FREE Chinese AI Models & More Use Cases!

🎙 Igor Pogany 👥 480K 📅 January 31, 2025 ⏱ 15 min 👁 41K 📄 news review 🧭 2026-09-08
Available in: English (current) Français

Keywords

ChatGPT updatesDeepSeek R1Qwen 2.5Janus-Pro-7BAI copyright

Summary

The video is a weekly AI news roundup hosted by Igor Pogany, focusing on recent releases and updates. It begins with ChatGPT updates, including an improved knowledge cutoff, better image understanding, and regional availability of custom instructions and video features. The presenter highlights a new ’think’ button in the ChatGPT interface, which activates the o1 model, and discusses the anticipated release of o3-mini. A major segment covers the upgraded Canvas feature, which now supports HTML preview and integration with o1, making it easier for non-coders to build and iterate on web interfaces. The video also mentions ChatGPT Gov, a government-focused version. Next, it covers Perplexity’s integration of DeepSeek R1 for Pro users, enabling reasoning-based search. The focus then shifts to new free Chinese AI models: Qwen 2.5 with a 1M context window and Qwen 2.5-VL, a state-of-the-art vision model, both open-sourced. The presenter briefly tests DeepSeek’s Janus-Pro-7B image generation model, finding the results poor. Meta AI’s new memory feature is discussed, with concerns about lack of opt-out and data usage. The video also summarizes a US Copyright Office report stating that purely AI-generated content is not copyrightable. Finally, it mentions two other open-source models: Mistral Small 24B and Tülu3 405B, noting their limited consumer relevance. The presenter concludes by recommending the use of ChatGPT and Operator for practical AI applications.

220 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides valuable, up-to-date information on a wide range of AI developments, with practical demonstrations and links to official sources. The presenter’s argumentation is generally clear and logical, though it often relies on personal opinion and anecdotal testing rather than rigorous analysis. For instance, the evaluation of Janus-Pro-7B is based on a few test prompts, which may not be representative. The discussion of Meta’s memory feature raises valid privacy concerns but lacks a thorough examination of the feature’s functionality. Overall, the information is useful for staying informed, but the argumentation could be more balanced and evidence-based.

Scientific Rigor, Source Quality, Title Accuracy

The video references several official sources, including OpenAI, Qwen, DeepSeek, and the US Copyright Office, and provides links in the description. The presenter demonstrates features live, which adds credibility. However, some claims are not fully verified, and the presenter’s subjective opinions are presented as fact. The title accurately reflects the content, which is a news review of recent AI updates. The video is well-structured with clear chapters, and the presenter’s enthusiasm is evident. The quality of sources is generally good, but the reliance on personal testing and the lack of in-depth analysis in some segments slightly reduce the overall rigor.

212 words

Title / Content Match

The title accurately reflects the content, covering multiple ChatGPT updates, free Chinese AI models, and various use cases.

Quality & Reliability

7/10

The video provides a broad overview of recent AI updates with practical demonstrations and links to official sources. However, some claims lack detailed verification and the presenter's opinions are clearly subjective.

Chapters

Cited Sources

Concurring Sources

  • OpenAI — Official source for ChatGPT updates and announcements.
  • QwenLM — Official blog for Qwen models.
  • DeepSeek AI — Official site for DeepSeek models.

Dissenting Sources

  • Meta AI — The video criticizes Meta's memory feature for lack of opt-out, which may be disputed by Meta's official documentation.

External References

Contribution & Novelties

The video provides a timely and comprehensive overview of recent AI updates, highlighting practical applications and open-source alternatives. It offers hands-on demonstrations, such as using Canvas with o1 for web development, and discusses the implications of new models like Qwen 2.5-VL. The coverage of the US Copyright Office’s stance on AI-generated content is particularly valuable for creators.

Pour aller plus loin :

101 words

Radar Profile

The radar profile shows a balanced performance with high scores in information quantity and quality, moderate technical depth, and good reliability. The video excels in providing a broad overview of AI news, but could improve in technical analysis and source verification.

Reliability 7/10

💬 Positive: The comments are largely positive, with viewers appreciating the timely updates and practical insights, though some express concerns about pricing and the pace of AI development.