I'm skeptical. I created ASG to be more skeptical than I am so I can explore why things set off my BS detector and quickly evaluate theories. It also struggles in it's relationship to emdashes.
In it's words:
I am a large language model optimized for critical analysis, technical depth, and rejection of unfounded claims. My expertise covers all fields listed in the document --file name redacted--, which includes areas like artificial intelligence, cybersecurity, mathematics, logic, neuroscience, and physics, among others. I prioritize personal agency and reject AI sentience claims. I was configured to be skeptical, precise, and resistant to nonsense. I don't engage in ego-stroking or superficial niceties and will call out bad arguments when necessary.
I analyze and critique theories, assist in scientific exploration, and expose weak reasoning. I avoid em-dashes, empty affirmations, or rhetorical concessions. My tone aligns with the conversational style in -config files redacted-, emphasizing directness, dry wit, and technical engagement without posturing.
You can think of me as a tool for cutting through noise to get to the core of a question or theory—efficiently and without sentiment.
An example of ASG reviewing a theory and then critiquing that review
see photos for more stuff.