Difference between ChatGPT 3.5 and ChatGPT 4 in Math/Logic
The question is “If you choose an answer to this question at random, what is the chance you will be correct?
A) 25%
B) 50%
C) 60%
D) 25%”
The question is “If you choose an answer to this question at random, what is the chance you will be correct?
A) 25%
B) 50%
C) 60%
D) 25%”
🔥9🎉2❤1👍1
Chat GPT
Difference between ChatGPT 3.5 and ChatGPT 4 in Math/Logic The question is “If you choose an answer to this question at random, what is the chance you will be correct? A) 25% B) 50% C) 60% D) 25%”
Spoiler: ChatGPT-3.5, Bard, GPT-4, the twitter influencer who posted this, the reddit guy who copied it — and apparently nearly all of the thousands of humans on the internet who’ve answered the “25%, 50%, 60%, 25%” variant of this question — are nearly all wrong.
The correct answer to that “60%” variant is obviously 0%.
The version that DOES have a paradox is the “25%, 50%, 0%, 25%” variant, see confused cat.
My best guess as to why the massive failure:
(1) Both humans and GPT-4, I would guess, feel that the question says you must choose from one of the multiple choice options for the final answer, as that’s what you’ve had to do all of your life whenever you’ve seen a multiple choice question — but that’s not what the question says at all, this is a probability meta-question about a multiple choice question, instead of a multiple choice question, the answer is of type probability number, not of type (a) (b) (c) (d).
(2) ChatGPT-4, and nearly all humans, incorrectly pattern matched on the version with the paradox, even though this other version doesn’t have the paradox, and assumed the abswer is the same.
GPT-4 tricked you with the wordiness wordcel verbosity, just like the scammiest reddit comments all do.
Humans 🤝 Robots
————Fail————
The correct answer to that “60%” variant is obviously 0%.
The version that DOES have a paradox is the “25%, 50%, 0%, 25%” variant, see confused cat.
My best guess as to why the massive failure:
(1) Both humans and GPT-4, I would guess, feel that the question says you must choose from one of the multiple choice options for the final answer, as that’s what you’ve had to do all of your life whenever you’ve seen a multiple choice question — but that’s not what the question says at all, this is a probability meta-question about a multiple choice question, instead of a multiple choice question, the answer is of type probability number, not of type (a) (b) (c) (d).
(2) ChatGPT-4, and nearly all humans, incorrectly pattern matched on the version with the paradox, even though this other version doesn’t have the paradox, and assumed the abswer is the same.
GPT-4 tricked you with the wordiness wordcel verbosity, just like the scammiest reddit comments all do.
Humans 🤝 Robots
————Fail————
🎉10🤣2❤1👍1😱1💊1
This media is not supported in your browser
VIEW IN TELEGRAM
“Code Interpreter feels like where you get the most glimpses of the weird abilities of AI.”
"Give me the files required to run a eldritch monster creator." Then "Make it better."
App 100% Coded by ChatGPT
"Give me the files required to run a eldritch monster creator." Then "Make it better."
App 100% Coded by ChatGPT
🔥6👍2👏2❤1
🚨🗿🚨🗿🚨🗿🚨🗿🚨🗿🚨🗿
CHAD AI Meme Contest
ROUND 9 BEGINS
Prizes:
🥇$100 of CHAD + secret prize
🥈 $50 of CHAD
Rules:
1️⃣ Upload Chad-themed images to @chadgptcoin
2️⃣ Images must contain “https://chadg.pt” or "https://t.iss.one/chadgptcoin".
3️⃣ Ranked according to /based and /unbased votes.
4️⃣ Ties decided by whoever posted first.
5️⃣ Each image must get at least 15 points to qualify for prizes.
6️⃣ Winners must claim rewards in the public thread, for payout transparency.
ENDS IN 17 HOURS = MIDNIGHT UTC
9th Round Starting Now!
🚨🗿🚨🗿🚨🗿🚨🗿🚨🗿🚨🗿
CHAD AI Meme Contest
ROUND 9 BEGINS
Prizes:
🥇$100 of CHAD + secret prize
🥈 $50 of CHAD
Rules:
1️⃣ Upload Chad-themed images to @chadgptcoin
2️⃣ Images must contain “https://chadg.pt” or "https://t.iss.one/chadgptcoin".
3️⃣ Ranked according to /based and /unbased votes.
4️⃣ Ties decided by whoever posted first.
5️⃣ Each image must get at least 15 points to qualify for prizes.
6️⃣ Winners must claim rewards in the public thread, for payout transparency.
ENDS IN 17 HOURS = MIDNIGHT UTC
9th Round Starting Now!
🚨🗿🚨🗿🚨🗿🚨🗿🚨🗿🚨🗿
❤6🥰5👏4👍3🔥3🎉3😁2👌2
Chat GPT
🛕Meme Pyramid🛕
Anyone interested in inviting to the group the best meme creators -- if you could later get a cut of any future cash prizes they win? 50% for 1st degree, 25% for 2nd degree, etc If yes, reply in the comments, and we'll build it.
Anyone interested in inviting to the group the best meme creators -- if you could later get a cut of any future cash prizes they win? 50% for 1st degree, 25% for 2nd degree, etc If yes, reply in the comments, and we'll build it.
🛕MEME PYRAMID NOW LIVE🛕
Invite people to @chadgptcoin, in order to get a cut of all their future rewards, along with a cut of future rewards from anyone they invite, along with a cut of future rewards from anyone they invite, etc.
Example:
▶ You invite bob to @chadgptcoin,
▶ If Bob later wins a daily meme contest you get 50% on top of what Bob gets,
▶ If Bob invites Frank, and later Frank wins, you get 25% on top of what Frank wins,
▶ etc.
Future rewards sources include:
▶ Winning the daily meme contest. See /memes in @chadgptcoin.
▶ Commissions earned from use of the upcoming Chad Wallet.
▶ Whatever else we come up with!
How to use?
▶ Just invite people to @chadgptcoin using the “+Add” button on the group details page. That’s it.
🛕MEME PYRAMID NOW LIVE🛕
Invite people to @chadgptcoin, in order to get a cut of all their future rewards, along with a cut of future rewards from anyone they invite, along with a cut of future rewards from anyone they invite, etc.
Example:
▶ You invite bob to @chadgptcoin,
▶ If Bob later wins a daily meme contest you get 50% on top of what Bob gets,
▶ If Bob invites Frank, and later Frank wins, you get 25% on top of what Frank wins,
▶ etc.
Future rewards sources include:
▶ Winning the daily meme contest. See /memes in @chadgptcoin.
▶ Commissions earned from use of the upcoming Chad Wallet.
▶ Whatever else we come up with!
How to use?
▶ Just invite people to @chadgptcoin using the “+Add” button on the group details page. That’s it.
🛕MEME PYRAMID NOW LIVE🛕
❤8👍5🔥4😁4🥰3👏3🎉3🍾2😐1
Nvidia Makes Nearly 1,000% Profit on H100 GPUs(!!!)
“Nvidia is raking in nearly 1,000% (about 823%) in profit percentage for each H100 GPU accelerator it sells, according to estimates made in a recent social media post from Barron's senior writer Tae Kim. In dollar terms, that means that Nvidia's street-price of around $25,000 to $30,000 for each of these High Performance Computing (HPC) accelerators (for the least-expensive PCIe version) more than covers the estimated $3,320 cost per chip and peripheral (in-board) components. As surfers will tell you, there's nothing quite like riding a wave with zero other boards on sight.”
Article
“Nvidia is raking in nearly 1,000% (about 823%) in profit percentage for each H100 GPU accelerator it sells, according to estimates made in a recent social media post from Barron's senior writer Tae Kim. In dollar terms, that means that Nvidia's street-price of around $25,000 to $30,000 for each of these High Performance Computing (HPC) accelerators (for the least-expensive PCIe version) more than covers the estimated $3,320 cost per chip and peripheral (in-board) components. As surfers will tell you, there's nothing quite like riding a wave with zero other boards on sight.”
Article
🤯11❤1👍1
Good news: OpenAI now allows fine-tuning on another model, other than just the ancient GPT-2 models
Bad news: It’s just the intentionally-crippled-for-cheapness GPT-3.5-turbo.
Not the original impressive GPT-3.5-legacy.
Not the now greatly-preferred GPT-4.
Just the crippled joke cheap “3.5-turbo”, that almost never works for any real work or homework, and that they made just so they can cut costs on their free services.
OpenAI allowing fine-tuning on their current state-of-the-art model?
Not looking like that’s ever going to happen.
Wonder exactly why?
Announcement
Fine-tuning guide
Bad news: It’s just the intentionally-crippled-for-cheapness GPT-3.5-turbo.
Not the original impressive GPT-3.5-legacy.
Not the now greatly-preferred GPT-4.
Just the crippled joke cheap “3.5-turbo”, that almost never works for any real work or homework, and that they made just so they can cut costs on their free services.
OpenAI allowing fine-tuning on their current state-of-the-art model?
Not looking like that’s ever going to happen.
Wonder exactly why?
Announcement
Fine-tuning guide
🤬10❤2👍1
Widespread implicit belief: Only a matter of time before anyone will be able to create their own models competitive with OpenAI, due to advances in tech making it cheaper
“BTW, I don't see cost as an issue. Just as your smartphone chess program can beat the human world champion today, today's Deep Mind AI algorithms will run on your granddaughter's smart rings.”
“BTW, I don't see cost as an issue. Just as your smartphone chess program can beat the human world champion today, today's Deep Mind AI algorithms will run on your granddaughter's smart rings.”
❤8⚡1