AI is great, i use it extensively, every day. however, you need to MANAGE the responses. it will always first of all tell you what it thinks you wanna hear. yesterday i asked what the impact would be if someone in a calorie deficit added 5k steps to their routine and also 200 calories of food that those steps would burn. it gave me a glowing interpretation of the benefits derived from that. then, because i had enough knowledge already on the subject to realize it was being way too positive, i said "it looks like maybe your response was focused on making me happy rather than providing an accurate answer". the AI apologized and then let me know that adding the steps and the equivalent calories was actually a BAD to neutral idea and told me why.
if you ask "tell me why leronlimab is a great fit for mss and likely to succeed" and then ask "tell me why leronlimab is a bad fit for mss and not likely to succeed", you'll get 2 very different responses, both very convincing. now, i happen to believe its a good fit for mets in general and specifically the mss liver mets, so perhaps i can decide what pieces of the responses have the greater value. but generally speaking, it makes it difficult to an extent, especially if you are researching something you don;t have reat knowledge of already.
AI won't refute your perceived supposition in most cases. it will if you are dead wrong and a fact is a fact. but beyond that, it follows your lead. when i asked about steps and calories, neither of which i count at all btw, it believed i wanted justification for adding the steps and cals and replied in favor of doing so. when my perceived supposition changed, it went the opposite way.
and when you have running conversations, the supposition bias is magnified greatly.
i use AI every single day, A LOT, for many different things. i rely on it, because it saves a MASSIVE amount of time. but i also know it can be heavily biased, can be inaccurate depending on the topic, and often also provides a very incomplete picture. its digs as deep as it thinks is necessary to match your curiosity and then stops... until you say, "but isn't there more to that?" or "i think there may be a few others that you failed to list", etc.
i 100% recommend people continue to utilize gemini etc for cytodyn related stuff and life in general, but just keep in mind that you often need to kinda prompt it from both sides of an argument to get a complete picture.
plug these in, see how it goes. use 2 separate tabs: sometimes it will reject your supposition!
"is it likely the fda will allow cytodyn to use ctDNA as a surrogate endpoint on its regulatory submission"
"is it likely the fda will reject cytodyn's wish to use ctDNA as a surrogate endpoint on its regulatory submission"
lol, delete at will. i did not cheerlead the company in my post.
if you ask "tell me why leronlimab is a great fit for mss and likely to succeed" and then ask "tell me why leronlimab is a bad fit for mss and not likely to succeed", you'll get 2 very different responses, both very convincing. now, i happen to believe its a good fit for mets in general and specifically the mss liver mets, so perhaps i can decide what pieces of the responses have the greater value. but generally speaking, it makes it difficult to an extent, especially if you are researching something you don;t have reat knowledge of already.
AI won't refute your perceived supposition in most cases. it will if you are dead wrong and a fact is a fact. but beyond that, it follows your lead. when i asked about steps and calories, neither of which i count at all btw, it believed i wanted justification for adding the steps and cals and replied in favor of doing so. when my perceived supposition changed, it went the opposite way.
and when you have running conversations, the supposition bias is magnified greatly.
i use AI every single day, A LOT, for many different things. i rely on it, because it saves a MASSIVE amount of time. but i also know it can be heavily biased, can be inaccurate depending on the topic, and often also provides a very incomplete picture. its digs as deep as it thinks is necessary to match your curiosity and then stops... until you say, "but isn't there more to that?" or "i think there may be a few others that you failed to list", etc.
i 100% recommend people continue to utilize gemini etc for cytodyn related stuff and life in general, but just keep in mind that you often need to kinda prompt it from both sides of an argument to get a complete picture.
plug these in, see how it goes. use 2 separate tabs: sometimes it will reject your supposition!
"is it likely the fda will allow cytodyn to use ctDNA as a surrogate endpoint on its regulatory submission"
"is it likely the fda will reject cytodyn's wish to use ctDNA as a surrogate endpoint on its regulatory submission"
lol, delete at will. i did not cheerlead the company in my post.
2
