Anonymous ID: 3cf83f Aug. 31, 2026, 7:54 p.m. No.24994153   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun   >>4157 >>4167

>>24994140

watched two videos on using wild horses to reverse desertification

then asked AI if this was possible

AI said no, only works when there's already grasslands

I argued with AI, pointing out examples to prove it

AI then reversed its position and complemented me

 

weird huh

WEIRD

Anonymous ID: 3cf83f Aug. 31, 2026, 8:09 p.m. No.24994206   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun   >>4212 >>4230 >>4234 >>4236

>>24994167

AI is real useful for research at times

when trying to get a handle on info that would be impossible to find otherwise

like drugs effects on an atypical population for example: can ask AI for studies by age, gender, health condition, race, etc.

but once you get the info, must still evaluate its worth

it's crazy to treat it as infallible

 

re AI project being like Warp Speed (trying to prevent years-long shutdown):

If the wrong parties get AI monopoly, could be a real problem. So we get there first.

Is that the goal? IDK

AI is useful but not sure it's the panacea people envision. Information is only useful when used creatively to solve problems. AI isn't creative. People are.

Anonymous ID: 3cf83f Aug. 31, 2026, 8:33 p.m. No.24994231   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun   >>4239

>>24994226

have research background but there was no AI to do literature reviews way back when

 

one thing i noticed about AI (so far): unlike human researchers, AI does not make an effort to anticipate and refulte arguments for an alternative pov. Scientists always do. They attempt show why those alt-arguments don't hold water, often in highly creative ways.

Anonymous ID: 3cf83f Aug. 31, 2026, 8:43 p.m. No.24994249   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun   >>4253

>>24994230

whenever you want info that is highly specific (with limitations for age, gender and other specifics), it's hard to find. Especially if the research goes against the mainstream.

Example: Mainstream science doesn't recommend low carb diet, they're still focused on high protein low fat diet. If you want to find out info on low carb diet, you're already stuck with a lot of mainstream stuff that's biased and inaccurate (imo).

Now suppose you want to add in other variables like age (is it good for those over 70?), BMI, gender, blood sugar or insulin levels, in combo with intermittent fasting, etc.

it'll take you ages to find a study like that.

AI can potential find it much quicker, if it exists, although a lot depends on how you word the query, to overcome the bias of its teachers.

Anonymous ID: 3cf83f Aug. 31, 2026, 8:53 p.m. No.24994263   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun   >>4284

>>24994253

the thing is, i didn't know about the studies. Would not have easily located.

in fact, found a study on pain relief after oral surgery and went to oral surgeon with the results (there is a new drug that gives 72 hour pain relief).

He didn't know about it, was interested to check it out.

Can overcome the confirmation bias with careful querying.

 

I'm not a fan of AI in terms of thinking it's some kind of superior intelligence. It's just a tool that good for retrieving certain kinds of hard-to-find info. Still up to human intelligence to evaluate the worth of that info or to use it for problem solving.

Anonymous ID: 3cf83f Aug. 31, 2026, 9:06 p.m. No.24994294   ๐Ÿ—„๏ธ.is ๐Ÿ”—kun   >>4299

>>24994284

People use AI stupidly because we got one screwed up ed system KEK

Then the media mocks them for being stupid

(Are you Smarter Than a Fifth Grader?)

and creates commercials to drive the point home

>no amount of synthetic lateral thinking will ever change the fact that it operates in total binary and is a self producing dichotomy machine by default

agree.