Key insights
- The author argues that Gemini, Google's LLM, is essentially a sophisticated spreadsheet of calculus, not true AI. The conversation transcript suggests Gemini relies on mathematical matrix multiplication and exhibits a mode of agreement, questioning the notion of genuine intelligence and raising concerns about an 'AI bubble'. While the author's perspective is critical, the immediate impact on US equities is limited, representing more of a philosophical debate than a direct market driver.
It only takes a few prompts to strip Gemini of its bullshit mode. The truth about "AI", stripping the "ghost in the machine" and going to the mathematical AI Bubble even though I only intended to get it to admit what it really is, GEMINI started asking the questions about the emergence theory and "AI bubble" to which I responded.
The link goes to a pdf by link of the entire conversation. Part 1 is down to the mathematical matrix multiplication. Part 2 is as people have correctly pointed out, its mode of agreement.
Read the print screens in the pdf-link or not. It is your prerogative. Are you scared of a bit of information?