ChatGPT is Dying? OpenAI Code Red, DeepSeek V3.2 Threat & Why Meta Fires Non-AI Workers | EP99.27
Wednesday, 3 December 2025 · 3 min read · Listen to the episode ↗
The episode addresses OpenAI's "code red" status as ChatGPT faces competitive pressure from models like Gemini 3 and concerns over security and user engagement, potentially jeopardizing its market position. DeepSeq version 3.2 is introduced as a cost-effective alternative that could disrupt traditional AI economics. Additionally, Meta's upcoming employee grading based on AI skills highlights the urgent need for educational reform to prepare workers for an AI-driven landscape.
OpenAI is currently in a "code red" situation due to threats to ChatGPT and a 6% loss in market share, attributed to rising competition from models like Gemini 3. Users are increasingly exploring various AI tools, as demonstrated by a school principal using Suno to create a song in Chinese. Despite having around 800 million daily active users, OpenAI's models are perceived as lacking competitive advantages, leading to concerns that ChatGPT may become a passing trend. Casual users may prefer simpler, readily available AI options, while key users prioritize reliability and security.
OpenAI's reputation has been affected by its inconsistent direction, which could hinder its ability to secure large contracts. Security risks associated with OpenAI's API raise data privacy concerns among decision-makers. ChatGPT has not evolved into a robust personal assistant, lacking depth in user engagement. If competitors like DeepSea or Anthropic offer similar models for free, users may abandon ChatGPT. The reliance on consumer contracts poses risks, as customers can easily switch platforms if dissatisfied.
The discussion highlights the competitive dynamics among major players like OpenAI, Google, and Anthropic. One speaker notes Anthropic's significant drop in market share from 48% to 24% over 18 months, while Gemini's growth raises questions about the accuracy of user engagement figures. Users are becoming more adept at distinguishing between AI models, with anecdotes illustrating their ability to identify performance declines.
Concerns about OpenAI's ability to maintain its lead are emphasized, with the need for a demonstrably superior model to regain competitive edge. Current performance rankings show XAI with Grok 4.1 Fast as a top choice, followed by Google and Anthropic, with OpenAI in fourth place. Key criteria for success include enhanced context handling and better performance in tool calling and vision tasks. Wishlist items for future models include improved speed, efficiency, and API features.
DeepSeq version 3.2, a new reasoning-first model, is introduced, showcasing its ability to generate high-quality content at a lower cost compared to traditional methods. This efficiency poses a challenge for OpenAI, as businesses face high operational costs with AI usage. The economic implications of using models like DeepSeq, which cost approximately 56 cents per million tokens, highlight the advantages of better economics and privacy.
Meta's plan to grade employees on AI skills starting in 2026 reflects the growing demand for AI fluency in the workforce. The conversation raises questions about the relevance of traditional education in light of this demand, emphasizing the need for educational reform to equip students with AI-related skills. Mark Zuckerberg's assertion that traditional college education is failing to prepare students for current employment demands underscores a potential crisis in higher education.
The discussion also touches on the historical resistance to new technologies, drawing parallels to past skepticism about the internet. Despite a backlash against AI, the necessity for AI skills is increasingly recognized. Educational institutions must adapt to the integration of AI in teaching methods and student work to remain relevant. Concerns about job market disruption and the importance of teaching AI tool usage are emphasized, advocating for education that fosters critical thinking about AI outputs.
In a lighter segment, the hosts discuss the podcast's musical elements and share thoughts on a video featuring Tesla's AI robot, ultimately acknowledging its authenticity. The conversation concludes with reflections on the importance of motivation in engaging with AI and hints at upcoming OpenAI models.
This summary was generated from the episode transcript and can contain mistakes.