Did Claude make a minor factual error when he discussed the VP presiding over an impeachment trial, seemingly in the context of the president being impeached?
There's a potential similar issue with discussing the President's discussions with the VP about his confirmation hearing (what confirmation hearing?), but you could take this as applying to a nominee to fill a vacancy in the Vice Presidency (e.g. Gerald Ford in the brief window between Agnew resigning and Ford being confirmed).
What was the corpus of documents/materials you provided to Claude for use during the initial set of questions you posed? At the very end, it appears you uploaded a copy of Trump's Jan 6 speech in connection with a follow up question. But presumably there was an initial set of documents Claude had access to - the SCT opinion, party briefing, etc.
Would you mind sharing how you primed Claude? I'm really interested in best practices for use of Claude in these contexts.
"Every judge presiding over a Trump case faces the challenging task of setting aside her views—positive or negative—about Trump. Some judges succeed better than others. Again, Claude doesn’t have that problem."
Is this actually true? Isn't any LLM going to be influenced by the valence of comment in its corpus about the person being discussed, which are disproportionately likely to be significant for Trump?
With the deliberate and well considered presidential immunity in the constitution, The Supreme Court stepping into overriding the constitutional intended failure to provide that protection for the president against the evidence of the deliberation of the subjection by the founders places the onus squarely on the Justices to proved as completely understandable position as a 5 year old's rationale for a choice between Sunny and Melon. If they are the arbiter of law, when they create law out of whole cloth, they are under the highest of obligation to create such law based upon sound rationale which must also include sound reason for creating such law.
Were such product as we have been given by the Supreme Court given to them by a lower court, who can foresee and result other than remand to the lower court for reconsideration? An arbitrary and irrational decision between Sunny and Melon is easily discerned by a 5 year old, and should be as promptly and unequivocally rejected by our legal community when it is designed to create more questions than it answers.
This is an excellent illustration of how Claude can be used to parse complicated sets of murky issues. Claude’s speed and clarity amazes me. Thank you, Adam, for this demonstration.
Did Claude make a minor factual error when he discussed the VP presiding over an impeachment trial, seemingly in the context of the president being impeached?
Good eye. The VP usually presides over impeachment trials, but the Chief Justice presides over impeachment trials of the President.
There's a potential similar issue with discussing the President's discussions with the VP about his confirmation hearing (what confirmation hearing?), but you could take this as applying to a nominee to fill a vacancy in the Vice Presidency (e.g. Gerald Ford in the brief window between Agnew resigning and Ford being confirmed).
What was the corpus of documents/materials you provided to Claude for use during the initial set of questions you posed? At the very end, it appears you uploaded a copy of Trump's Jan 6 speech in connection with a follow up question. But presumably there was an initial set of documents Claude had access to - the SCT opinion, party briefing, etc.
Would you mind sharing how you primed Claude? I'm really interested in best practices for use of Claude in these contexts.
"Every judge presiding over a Trump case faces the challenging task of setting aside her views—positive or negative—about Trump. Some judges succeed better than others. Again, Claude doesn’t have that problem."
Is this actually true? Isn't any LLM going to be influenced by the valence of comment in its corpus about the person being discussed, which are disproportionately likely to be significant for Trump?
With the deliberate and well considered presidential immunity in the constitution, The Supreme Court stepping into overriding the constitutional intended failure to provide that protection for the president against the evidence of the deliberation of the subjection by the founders places the onus squarely on the Justices to proved as completely understandable position as a 5 year old's rationale for a choice between Sunny and Melon. If they are the arbiter of law, when they create law out of whole cloth, they are under the highest of obligation to create such law based upon sound rationale which must also include sound reason for creating such law.
Were such product as we have been given by the Supreme Court given to them by a lower court, who can foresee and result other than remand to the lower court for reconsideration? An arbitrary and irrational decision between Sunny and Melon is easily discerned by a 5 year old, and should be as promptly and unequivocally rejected by our legal community when it is designed to create more questions than it answers.
This is an excellent illustration of how Claude can be used to parse complicated sets of murky issues. Claude’s speed and clarity amazes me. Thank you, Adam, for this demonstration.