Artificial Intelligence Tools

CCarter

Final Boss ®
Moderator
BuSo Pro
Boot Camp
Digital Strategist
Joined
Sep 15, 2014
Messages
4,341
Likes
8,855
Degree
8
uRmDb5J.png

This thread is dedicated to the latest advancements in artificial intelligence - current tools and their capabilities, as well as new tools coming out. The original thread concentrated on A.I. generated content and dove into discussions of what the latest. I wanted to create a thread where we can go back and find tools that people talked about instead of having to do deep dive searches.

Here are some of the tools I use, feel free to add your own and your takes as well:

ChatGPT

Link: https://openai.com/chatgpt/

This is the one that got the masses really excited. I currently use it for all my programming. It was at first responsible for 25% of my new code, but now it's approaching 95%. I simply write instructions to it what I want and then when there are errors I can paste in the errors and ChatGPT fixes them. It's a lot of trial and error with back and forth, but it's gotten to the point I just copy and paste the code, do a once look over and test it by running it.

One trick I recently found that will save you coders some headaches is:

Finish the code. Do not print the full code again, just a missing part from last answer.

When you write big code script ChatGPT tends to stop after a certain line of code (about 200 lines), you can use that sentence above to get it to continue from where it left off.

ChatGPT has a mobile version and several days ago released a Desktop version so you don't need to use a browser to get to it on desktop anymore.

--

MidJourney

Link: https://www.midjourney.com/home

It's a Discord software, so you talk with it in private chat in Discord, to generate images. Example:

XmHrIHw.png

Selected one:

bkJUyIT.png

The input was:

Superman overlooking a waterfall --ar 7:4 --s 750

--

CoPilot

Link: https://code.visualstudio.com/

CoPilot: GitHub Copilot in VS Code

This is an embarrassment, but somehow Microsoft created one of the best code editors in the world under my nose, Visual Studio Code. And then they added CoPilot, which apparently is using Github source code, as an A.I. to help you code 100% better.

@SmokeTree put me on this. Again anytime @SmokeTree gives me advice I 100% DO NOT want to take it. BUT I eventually give in and test it what crazy voodoo he is talking about and 100% of the time I am blown away. My own bias works against me every time. He's the one that put me on to Redis, now Visual Studio, and now CoPilot. He's also pointed me to about 20+ other A.I. software that even I can't comprehend just yet on what's going on.

Him and @eliquid are one of the few people that are always testing stuff out which are beyond stuff even I can comprehend.

Right now I still use ChatGPT for my programming, but I've been slowly transitioning to CoPilot and Visual Studio IDE environment. Microsoft - of all freaking companies.

--

HeyGen

Link: https://www.heygen.com/

Heygen has been great for voiceover spokesperson for videos. It also has a crazy translation option for videos, I show an example here:


--

Adobe Photoshop Beta

Link: https://www.adobe.com/products/photoshop.html

When you get Photoshop you have the ability to get the Beta version as well, and here are some of the results from what's coming down the pipe:


--

OpenVoice

GitHub Link: https://github.com/myshell-ai/OpenVoice

Create A Bot Link: https://app.myshell.ai/explore

Here is OpenVoice - here you can play around with A.I. voices and get them to say whatever you want and export them:



--

OpenVoice

Link: https://elevenlabs.io/

Another A.I. voice generator. Again this is great for ya'll who want Faceless YouTube or Faceless Social media accounts.


--

Runway's Gen-2

Link: https://research.runwayml.com/gen2

I never tested this one out myself, so use caution:



--

Luma AI

Link: https://lumalabs.ai/dream-machine

And now there is Luma's Dream Machine. It can take your images and turn them into videos - and it's available right now unlike Soro. Some classic memes turned into video:

uXojyA0.gif

mF04CbQ.gif

Obviously they'll have to work out the obvious problems.

--

As I find more I'll post them here.
 
Wait until you learn about agents creating other agents and each agent uses it's own Ai API or LLM locally with Ollama.

Or that you can run all this on Google CoPilot of all things.

Or say you just wanna RAG with your own PDF's and text files and you learn about Google LLM Notebook
 
Last edited:
Suno

Link: Suno.com

AI music maker. Can add lyrics or make instrumentals. It is useful for creating entire compositions or bespoke audio samples that no one has.

In this example, I took the opening paragraph in @CCarter's post and made a song with it:
https://suno.com/song/896dc6f2-ca6f-43b6-a378-1ec2064cadbc
(Note it maxed out time, so the song doesn't 'finish,' but you can extend with commands like fade out or end song to have it wrap it up)

Here, I generated lyrics from the same paragraph:
https://suno.com/song/7e1b5c37-97f4-42e3-b285-981661e5f251
 
HeyGen looks so much better than the ones I tested just 6 or so months ago. Awesome how things are moving along in this space. You really can't keep up unless you're testing a tool a day. This thread will be gold.

Give Claude a go for your coding @CCarter - I've found it's better at following patterns of what you already did and adding bits for you etc than GPT4 which has an annoying habit of deciding you did something wrong and needs to change it for you lol (and break everything in the process).
 

Whisper​

By OpenAI.

This is a free AI transcription (speech to text) program that operates locally. You feed it .wav files and it spits out a block of text. I use this to transcribe interviews / phone calls but can be used for any video file. The larger models freeze up my laptop but the Small model does a great job for me.

Runs through Terminal on Mac, I don't use PC but it works on PC too.

Here's a helpful guide on how to set it up (on Mac): https://medium.com/gimz/how-to-inst...peech-to-text-recognition-system-1f6709db6010

Link: https://openai.com/index/whisper/
 
I've been using a great tool called Cuppa.sh for articles.

It's much less expensive than most other writing tools as you use your own OpenAI key.

Quality is better to when you prompt it properly and use the right keywords.
 
In this example, I took the opening paragraph in @CCarter's post and made a song with it:
https://suno.com/song/896dc6f2-ca6f-43b6-a378-1ec2064cadbc
(Note it maxed out time, so the song doesn't 'finish,' but you can extend with commands like fade out or end song to have it wrap it up)

Here, I generated lyrics from the same paragraph:
https://suno.com/song/7e1b5c37-97f4-42e3-b285-981661e5f251

YOoooooooooo. WTF. I was sitting here jamming to that in the background. This is insane. Now I got ideas, can go back to the old WickedFire rap threads and create some nostaglia magic.

--

Here are some new additions:

OpusClip

Link: https://www.opus.pro/

From the GaryVee multiple content strategy, this A.I. tool allows you to automatically create clips from long format video content. I haven't used it yet, so be careful:


--

Submagic

Link: https://www.submagic.co/

This A.I. tool allows you to enhance videos to make the "POP":

 
If you want to code using AI, I recommend Cursor instead of VSCode + Copilot.

It's a fork (clone) of VS Code (so you can import all your settings & plugins) but it has AI features built in, like sending all your files, plus it's compatible with multiple AI vendors.

Speaking of which, Claude 3.5 which just came out is frickin amazing compared to gpt-4 for coding.
 
HeyGen

HeyGen has succumbed under the pressure of competition and now has lowered it's price from $119 a month to $29 a month. They did limit videos to 5 minutes instead of 60 minutes, but I doubt anyone was really using over 5 minutes anyways.

--

InVideo AI

A new entry is InVideo, they have an AI. Video editor And an A.I. Video Generator

AI Video Editor (link)
Create and edit videos with text prompts using invideo's AI video editor. Just think of a topic, and it will automatically generate your script, create scenes, and add voiceovers. Generate professional videos within minutes.

Generate Al videos with just text (link)
Instantly turn your text inputs into publish-worthy videos. Invideo Al video generator simplifies the process, generating the script and adding video clips, subtitles, background music, and transitions. Add finishing touches with an intuitive editor. Create videos at scale without any learning curve!


Create videos with text prompts (link)
Publish-ready videos with zero video creation skills. Type any topic and invideo AI creates a video with script, visuals, subtitles, voiceover & music.

InVideo's pricing seem to be reasonable from free, $24, and up-to $60 a month. Worth testing out.

Trailer:

 
Fake Commercial (example for eCommerce)

So I used @Boy's recommended Suno and realized you can create some great soundtracks (check the instrumental option) for free. Then I went to Pexels.com and found some free stock videos, and created this fake commercial in an hour. Obviously a video editing software will be required - I use my trusted Screenflow.

Sound is always one of the biggest problems when creating videos for eCommerce commercials - licensing and all that nonsense was a major hurdle - not anymore. Just create your own music from the input line - I went for a dramatic soundtrack, you can do tranquil or whatever you want - depends on your desires.

Suno's free, at least the version I used, so there literally is almost no barrier to entry left:


That sound is my composition, copyrighted so don't copy me.
 
synthflow. give it a paragraph of text telling it how you want it to answer the phone and it will talk with the caller as if it were a real person to achieve the goal
 
DomoAI

AI Art Generator (link)

You can turn videos or images into cartoon/anime or other stylized versions with this tool.

Images to Video example:


Video to Animated example (pixelated example):

 
If you aren't using NotebookLM, do you even Ai?

But just so you know, Meta released an open source version of it:
https://github.com/meta-llama/llama-recipes/tree/main/recipes/quickstart/NotebookLlama
Oh, that's juicy. I uploaded a 300-page 'annual stockholder meetings' PDF that I was never going to read and truly don't care about just to test it out. Good summary with citations.

Just now, I learned you can import YouTube videos. I put in the last 4 I recorded, and it did a damn good job summarizing what I'm working on now, and the generated podcast was nearly perfect.

Up next - a tool that allows you to upload sales letters from Halpert, Hopkins, Schwartz, Ogilvy, etc., and use their writing style explicitly to write copy.
 
Oh, that's juicy. I uploaded a 300-page 'annual stockholder meetings' PDF that I was never going to read and truly don't care about just to test it out. Good summary with citations.

Just now, I learned you can import YouTube videos. I put in the last 4 I recorded, and it did a damn good job summarizing what I'm working on now, and the generated podcast was nearly perfect.

Up next - a tool that allows you to upload sales letters from Halpert, Hopkins, Schwartz, Ogilvy, etc., and use their writing style explicitly to write copy.

Ive already done this, but it's not for public use.

Sorry

But the best way to do it, IMHO, is to use Make.com so you can use higher end APIs per check over and rewrite or do to a tree of thoughts type promoting for the results you want from those copywriters
 
A guy I'm partnering up is brining AI over WhatsApp for us. The AI can re-engage customers, do customer service, sell services, take customer details and put it into a database, and send the payment link. in LATAM, people are used to talking to customers over WhatsApp. This might be the future when this technology spreads to the rest of the globe.
 
Anyone mess with cursor at all?
Seems like a very impactful way use gpts.
https://www.cursor.com/

Works with the new claudee.
 
For the gents who are afraid of GitHub and don't know the difference between a pip and a git, check out Pinokio if you have the computing power to run AI engines offline.

https://pinokio.computer/
 
I've used cursor a bit.
Surprisingly good for quickly creating a tool, etc.
 
Back