The official ChatGPT thread - the next big thing?

[YOhio]

This is a really cool tool to play around with. You can use all the different language models to compare similar prompts.

https://nat.dev

[YOhio]

This is a really cool tool to play around with. You can use all the different language models to compare similar prompts.

https://nat.dev

Here's an example:

Screenshot 2023-03-10 at 1.45.04 PM.png

[Jeff Lebowski]

I like the bottom two.

[tooblue]

Didn't know where else to put this, but this thread seemed relevant ...



Chinese firm invents lockdown-inspired kissing machine for remote lovers

https://nationalpost.com/news/world/...emote-lovers-2

china-kissing-device-.jpg?quality=90&strip=all&w=1128&h=846&type=webp.jpg

[All-American]

Didn't know where else to put this, but this thread seemed relevant ...



Chinese firm invents lockdown-inspired kissing machine for remote lovers

https://nationalpost.com/news/world/...emote-lovers-2

china-kissing-device-.jpg?quality=90&strip=all&w=1128&h=846&type=webp.jpg

One thousand percent chance this device is used for ulterior purposes.

[YOhio]

I’m excited about the new plug-in features.

Embedded Tweet

[SteelBlue]

Y-O, you are the best I've seen at the AI image generation. Have you messed around with Bing's new ability to generate with Dall-E? I'm having fun with it, and especially like that I don't seem to run out of attempts for free. It doesn't touch some of the others you're using but it's still impressive. Just need to get better at prompts because when I get something close to what I like, I haven't figure out how to make Bing recreate that exact image with suggested changes. So I feel like I have to get it right at first, or I'm going to be dealing with a totally new image.

[smokymountainrain]

One thousand percent chance this device is used for ulterior purposes.

[YOhio]

Y-O, you are the best I've seen at the AI image generation. Have you messed around with Bing's new ability to generate with Dall-E? I'm having fun with it, and especially like that I don't seem to run out of attempts for free. It doesn't touch some of the others you're using but it's still impressive. Just need to get better at prompts because when I get something close to what I like, I haven't figure out how to make Bing recreate that exact image with suggested changes. So I feel like I have to get it right at first, or I'm going to be dealing with a totally new image.

There are a couple resources I suggest.

Prompt Hero allows you to search for images specific to the platform you're using. Then you can modify the prompt as appropriate for the image you're trying to recreate.
https://prompthero.com

I've also created a Twitter list of people who have some good AI insights. If you see any images with the Alt tag you can select that and see the prompts they use.
https://twitter.com/i/lists/1635321511522476034?s=20

I have messed around a little with Bing, but haven't had much success. I'm trying to recreate an image of David Patten riding a mule through the Tennessee forest when he saw Cain and I can't get it right on any of the platforms. I'll post the Bing image in the other thread.

[BigPiney]

I have messed around a little with Bing, but haven't had much success. I'm trying to recreate an image of David Patten riding a mule through the Tennessee forest when he saw Cain and I can't get it right on any of the platforms. I'll post the Bing image in the other thread.

That will be awesome.

[SteelBlue]

There are a couple resources I suggest.

Prompt Hero allows you to search for images specific to the platform you're using. Then you can modify the prompt as appropriate for the image you're trying to recreate.
https://prompthero.com

I've also created a Twitter list of people who have some good AI insights. If you see any images with the Alt tag you can select that and see the prompts they use.
https://twitter.com/i/lists/1635321511522476034?s=20

I have messed around a little with Bing, but haven't had much success. I'm trying to recreate an image of David Patten riding a mule through the Tennessee forest when he saw Cain and I can't get it right on any of the platforms. I'll post the Bing image in the other thread.

Thank you! And the David Patten image is a phenomenal idea. Can't wait.

[Jeff Lebowski]

I got an invite from Google to be an early user/tester of Bard. I tried it out today. Not bad, but nowhere near as good as ChatGPT.

1. Writing letters, emails, essays, etc based on prompts. - Speed is good, but results are just OK. More terse than ChatGPT. In one case I asked it to write a letter of recommendation and gave it some information and it refused to do it (ChatGPT knocked this one out of the park). I asked it to write a church talk and the result was almost identical to the ChatGPT version. Like, almost word for word - that is weird.

2. Programming - terrible. Didn't even attempt to give a response. Can't do it yet.

3. Interactive research - I asked it to explain the random forest method of machine learning. Then I used a series of follow up questions to explore in more detail ("Can you give me an example?", etc). It did pretty well here. I would expect that from a google product.

4. Math and engineering problems. - this is one area where it seemed to do way better than ChatGPT. It handled complex problems and nailed the answer on the first try. This part was impressive.

[Uncle Ted]

A professor says he's stunned that ChatGPT went from a D grade on his economics test to an A in just 3 months

• The progress that ChatGPT made in an exam in just three months stunned an economics professor.
• Bryan Caplan of George Mason University said the chatbot got a D on his economics test in January.
• He tried again with the GPT-4 update last week and its score improved to an A.

[...]
ChatGPT's responses underwhelmed him so much that Caplan bet an AI model wouldn't score an A on six out of seven of his exams before 2029.

But when ChatGPT-4 debuted, its progress stunned Caplan. It scored 73% on the same midterm test, equivalent to an A and among the best scores in his class.

ChatGPT's paywalled upgrade sought to fix some of the early issues with the beta version, GPT-3.5. This purportedly included making ChatGPT 40% more likely to return accurate responses, as well as making it able to handle more nuanced instructions.

For Caplan, the improvements were obvious. The bot gave clear answers to his questions, understanding principles it previously struggled with. It also scored perfect marks explaining and evaluating concepts that economists like Paul Krugman have championed.

"The only thing I can say is it just seems a lot better," Caplan said.

https://www.businessinsider.com/econ...-months-2023-3

LOL... I wouldn't be surprised if ChatGPT is taking bets with dumb econ professors now.

[Uncle Ted]

Bing AI has some major mental problems... Microsoft needs to shut this thing down:

No wonder Bill Gates says AI will attack us.

[YOhio]

I got an invite from Google to be an early user/tester of Bard. I tried it out today. Not bad, but nowhere near as good as ChatGPT.

1. Writing letters, emails, essays, etc based on prompts. - Speed is good, but results are just OK. More terse than ChatGPT. In one case I asked it to write a letter of recommendation and gave it some information and it refused to do it (ChatGPT knocked this one out of the park). I asked it to write a church talk and the result was almost identical to the ChatGPT version. Like, almost word for word - that is weird.

2. Programming - terrible. Didn't even attempt to give a response. Can't do it yet.

3. Interactive research - I asked it to explain the random forest method of machine learning. Then I used a series of follow up questions to explore in more detail ("Can you give me an example?", etc). It did pretty well here. I would expect that from a google product.

4. Math and engineering problems. - this is one area where it seemed to do way better than ChatGPT. It handled complex problems and nailed the answer on the first try. This part was impressive.

Interesting. I used ChatGPT to build an arduino sketch and it was perfect. Haven’t tried with Bing or Bard.