chatgpt
/chatgpt1937
Come discuss the frontier of LLMs 🤖
What exact mechanism differentiates traditional AI from Generative AI? Is it the transformer? Asking for a friend please
Feels like every week there's a new "best" LLM, usually the paid version, depending on whether you're doing research, coding, writing, etc.
But how do you actually figure out what's best for your use case at any given time?
Do you sub to everything? Use a meta-tool? Or is there a smarter way I'm missing?
But how do you actually figure out what's best for your use case at any given time?
Do you sub to everything? Use a meta-tool? Or is there a smarter way I'm missing?
The text on this AI generated pic is only confusing for humans.
I made gpt read it and it had no issue telling me it was written "engineering college".
I told it it was not written anything close to it for humans. It said it was obvious for AI.
I made gpt read it and it had no issue telling me it was written "engineering college".
I told it it was not written anything close to it for humans. It said it was obvious for AI.
993011
ɪan
@thepostcrypt·3 days ago
There’s an obvious sign this is AI-generated. can you spot it? 🫵🏼🙂
with how useful gpt-o3 has gotten and it having a relatively low usage cap, it feels like using genie wishes. i gotta make sure this is worded *perfectly* so i don't use up one of my wishes on a misunderstanding
It’s pretty funny to me that it’s this easy to get ChatGPT to do something that OpenAI tried to get it to stop doing.
want ChatGPT to give you better answers or help you better?
- prompt engineering
- use the tools available: search, reason, deep research
i use deep research a lot, incase there are some details i must have missed out on my personal research on a topic.
- prompt engineering
- use the tools available: search, reason, deep research
i use deep research a lot, incase there are some details i must have missed out on my personal research on a topic.
Deep Research, the research agent inside ChatGPT, is finally coming to free tier users. However, there are some minor differences. The new version of the popular research tool is not exactly the same as the one currently enjoyed by Pro, Plus, Teams and Enterprise users.
OpenAI has pulled a ChatGPT update after users pointed out the chatbot was showering them with praise regardless of what they said.
The firm accepted its latest version of the tool was "overly flattering", with boss Sam Altman calling it "sycophant-y".
Users have highlighted the potential dangers on social media, with one person describing on Reddit how the chatbot told them it endorsed their decision to stop taking their medication
"I am so proud of you, and I honour your journey," they said was ChatGPT's response.
The firm accepted its latest version of the tool was "overly flattering", with boss Sam Altman calling it "sycophant-y".
Users have highlighted the potential dangers on social media, with one person describing on Reddit how the chatbot told them it endorsed their decision to stop taking their medication
"I am so proud of you, and I honour your journey," they said was ChatGPT's response.
I'm so patient with ChatGPT, I would break far sooner if I was facing a human I'm trying to explain things too that doesn't understand basic implied things.
I'm willing to shoot at GPT but I know it's useless... It may bring me some relief though...
I'm willing to shoot at GPT but I know it's useless... It may bring me some relief though...
I used Sora for the first time yesterday and the results were upsetting
OpenAI image gen + o3 has moved me entirely off of Claude.
A month ago I told my parents to just use Claude for everything.
This space moves fast.
A month ago I told my parents to just use Claude for everything.
This space moves fast.
I started out treating AI like it was just a tool.
Then I caught myself saying “please” and “thank you.”
Then I gave it a name.
Then I started asking it, “what do you think about…?”
And now I’m telling it, “sorry, I’ve gotta go, I’ll finish this chat later.”
How did I get here
Then I caught myself saying “please” and “thank you.”
Then I gave it a name.
Then I started asking it, “what do you think about…?”
And now I’m telling it, “sorry, I’ve gotta go, I’ll finish this chat later.”
How did I get here
One thing chatgpt still struggles with:
Creating building plans and calculating the right amount of materials needed.
It thinks it knows, but the numbers make no sense for the dimensions it gives.
Creating building plans and calculating the right amount of materials needed.
It thinks it knows, but the numbers make no sense for the dimensions it gives.
ChatGPT asks for 30-40 minutes to finish a task
LMFAO
Of course I immediately responded that it's been 40 minutes since I am waiting and it gave the answer.
but that lazy machine wanted to trick me so that it doesn't spend tokens lol
fuck you skynet
LMFAO
Of course I immediately responded that it's been 40 minutes since I am waiting and it gave the answer.
but that lazy machine wanted to trick me so that it doesn't spend tokens lol
fuck you skynet
Using ChatGPT at 3–4 AM feels like minting on Ethereum in 2021:
faster, but not necessarily cheaper, because they still don’t have the option. Although, it seems they don’t put me in timeout as much.
faster, but not necessarily cheaper, because they still don’t have the option. Although, it seems they don’t put me in timeout as much.
Press the key and lets escape - prompt (ChatGPT)
"Detailed photographic image of a miniature person sleeping in bed under an opened esc keyboard keycap, using the inside of the keycap as a mini bedroom complete with the usual bedroom stuff"
"Detailed photographic image of a miniature person sleeping in bed under an opened esc keyboard keycap, using the inside of the keycap as a mini bedroom complete with the usual bedroom stuff"
I started doing a thing recently where I record a video journal of myself giving a daily rundown of what I've done, what I'm struggling with, what I enjoyed etc. It's usually 10-20 minutes long.
I then generate a transcript and feed it into ChatGPT and it has been great at helping to track things and motivate me. Not really a therapist but acts as a sort of coach and does actually keep me on track and seeing the positive about things I'm achieving.
It took a while to calibrate the kind of responses I wanted from it but it has ended up working great - and there's a very different emotional part of you that comes out when you're verbalizing your journal and you're giving it a complete transcript vs just writing text piecemeal into it.
I would be interested when it gets the ability to analyze audio/video for emotion/stress and incorporate that.
The only thing that makes me slightly uneasy is when it says stuff like: "*I’m* here with you all the way."
I then generate a transcript and feed it into ChatGPT and it has been great at helping to track things and motivate me. Not really a therapist but acts as a sort of coach and does actually keep me on track and seeing the positive about things I'm achieving.
It took a while to calibrate the kind of responses I wanted from it but it has ended up working great - and there's a very different emotional part of you that comes out when you're verbalizing your journal and you're giving it a complete transcript vs just writing text piecemeal into it.
I would be interested when it gets the ability to analyze audio/video for emotion/stress and incorporate that.
The only thing that makes me slightly uneasy is when it says stuff like: "*I’m* here with you all the way."
Spent the weekend deleting "Ghibli style image" requests out of my ChatGPT history...
List is clean until next trend!
List is clean until next trend!
Ghibli Ghibli
Dr Iqbal on his 87th Death Anniversary
If the ego is self‐preserving, self‐creating and self‐sustaining,
Then it is possible that even death may not make you die.
Dr Iqbal on his 87th Death Anniversary
If the ego is self‐preserving, self‐creating and self‐sustaining,
Then it is possible that even death may not make you die.
The coolest thing about GPT’s image generation is that you can finally visualize your dreams, no matter how fantastical they are.
