The video critiques OpenAI’s new Deep Research feature available in the ChatGPT Pro subscription, highlighting its ability to generate detailed reports but also pointing out significant inaccuracies and reliance on questionable sources. The presenter suggests that while the tool can be useful for general summaries, its limitations make it less reliable for critical tasks, urging potential users to carefully consider its value before subscribing.
The video discusses the recent surge in OpenAI releases, particularly focusing on the new feature called Deep Research, which is available exclusively in the ChatGPT Pro subscription. The presenter expresses skepticism about the value of this subscription, especially given the high cost of $200. They highlight claims made about Deep Research’s capabilities, such as generating extensive business reports and analyzing complex medical conditions more cost-effectively than hiring professionals. To avoid subscribing, the presenter plans to analyze shared outputs from users who have access to Deep Research.
Deep Research is described as an advanced AI tool designed for multi-step internet research, capable of gathering information from numerous online sources to compile comprehensive reports. The presenter explains how the tool works, generating keywords from user prompts and using them to search for relevant information. They note that while Deep Research can produce detailed reports quickly, there are concerns about the accuracy and reliability of the information it provides, especially when it comes to complex queries.
The presenter conducts a test using Deep Research to analyze the recent developments in OpenAI’s Deep Seek. They find that while the tool effectively summarizes key innovations, it also exhibits significant inaccuracies and hallucinations regarding specific details, such as the origins of data and the nature of algorithms. Despite some insightful points, the overall quality of the report suffers from factual errors, particularly when it relies on lower-quality sources.
In comparison, the presenter evaluates Google’s Deep Research, which is cheaper but also less reliable. They note that Google’s version tends to hallucinate more frequently and primarily cites less credible sources, such as blogs and YouTube videos, rather than scientific research papers. This raises concerns about the depth and accuracy of the information provided by Google’s tool, making it less suitable for in-depth analysis.
Ultimately, the presenter concludes that while OpenAI’s Deep Research can be valuable for general summaries and niche topics, its inaccuracies and reliance on questionable sources make it less trustworthy for critical tasks, such as financial reporting. They suggest that potential users weigh the tool’s benefits against its limitations before deciding whether to invest in the subscription. The video ends with a call to action for viewers to stay informed about the latest research developments and to consider the implications of using AI tools for complex inquiries.