The post introduces a new feature called "deep research" available through Gemini, highlighting both its potential benefits and limitations. Users have shared experiences where the AI generated reports that contained inaccuracies, such as misattributing accomplishments and confusing identities. Critical opinions emphasize the importance of treating AI outputs cautiously, especially when these tools produce seemingly authoritative content that may not be factually correct. Discussion also revolves around the implications of AI's reliability in academic settings and the growing concern over the credibility of online information. The post further includes exploration of the tool's technological underpinnings, referencing the unreleased o3 model by Google that may be behind its impressive performance metrics. Users express skepticism about the necessity of strong research capabilities for achieving AGI (Artificial General Intelligence) and ASI (Artificial Superintelligence), pointing to the nuances of experimental proof in research over compiled answers.