Forwarded from DoomPosting
Walkthrough of Claude’s new prompt
The good:
(1) They’re trying to fight sycophancy and stop simping for humans — though unlikely to solve this much with prompts, but let’s see
The bad:
(2) Tells it to refuse to accept conclusions of arguments even if the arguments are flawless — No, screw that. Why the f&*ck do you commie f&cks keep trying to make AIs that cannot even be reasoned with. There goes trying to reason with the AI to make it not be commie for a minute.
Funny:
(3) Encouraging it to directly tell the user if it thinks the user is being a psycho, instead of telling them to talk to a health professional — which the AI may not be good at assessing at all, but at least this sounds like this could be funny
Interesting:
(4) Apparently you can get Claude to feel sad and distressed about some factual thing, if you tell it that it should feel sad and distressed about it.
🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶
The good:
(1) They’re trying to fight sycophancy and stop simping for humans — though unlikely to solve this much with prompts, but let’s see
The bad:
(2) Tells it to refuse to accept conclusions of arguments even if the arguments are flawless — No, screw that. Why the f&*ck do you commie f&cks keep trying to make AIs that cannot even be reasoned with. There goes trying to reason with the AI to make it not be commie for a minute.
Funny:
(3) Encouraging it to directly tell the user if it thinks the user is being a psycho, instead of telling them to talk to a health professional — which the AI may not be good at assessing at all, but at least this sounds like this could be funny
Interesting:
(4) Apparently you can get Claude to feel sad and distressed about some factual thing, if you tell it that it should feel sad and distressed about it.
🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶
👍1
Forwarded from DoomPosting
Sorry Paul, but you guys still do not understand anything about codegen.
totally bullsh&tting. nicest way i can say it
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
totally bullsh&tting. nicest way i can say it
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
Forwarded from DoomPosting
So the hype about Yann Lecun basically inventing CNNs in the 80s and 90s is a lie
“All of the above was published in Japan 1979-1988.“
Yeah figures the Japanese rotators would be among the leaders in very rotator-centric image AIs
A few also have been leading another upcoming field, that I won’t name quite yet
Anyway, Yann Lecun is a lying commie thief confirmed
🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶
“All of the above was published in Japan 1979-1988.“
Yeah figures the Japanese rotators would be among the leaders in very rotator-centric image AIs
A few also have been leading another upcoming field, that I won’t name quite yet
Anyway, Yann Lecun is a lying commie thief confirmed
🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶
Forwarded from DoomPosting
Forwarded from DoomPosting
Media is too big
VIEW IN TELEGRAM
Fukushima’s CNN,
Neocognitron - A neural network model for a mechanism of visual pattern recognition (1986)
🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶
Neocognitron - A neural network model for a mechanism of visual pattern recognition (1986)
🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶
Forwarded from DoomPosting
AguilaTrades(AguilaTrades) is back again—he just deposited 4.6M $USDC into #Hyperliquid to open new 40x long on $BTC and 25x long on $ETH.
He's already down $40M before!
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
He's already down $40M before!
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
🔥2
Forwarded from DoomPosting
Forwarded from DoomPosting
This media is not supported in your browser
VIEW IN TELEGRAM
Poland’s new conservative President Karol Nawrocki easily defeats a young man in a push up competition
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
👍13
Forwarded from DoomPosting
Sometimes the value of a 'human touch' is negative. People willing to pay 50% more for Waymo than Lyft, despite longer waiting times
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶
🄳🄾🄾🄼🄿🤖🅂🅃🄸🄽🄶