Off Topic Discussions

[Key-J]

So,

Unless you have been living under a rock you have most likely heard of the AI craze that is taking over the world.

No, not Skynet - put down the weapons people.

I'm talking about the tools like ChatGPT, MidJourney, DALL-E, Google BARD and more.

I am curious to ask what everyone on the Red Team thinks about these AI tools.

• Have you used them?
• Did you like them?
• What do you think the future with them will look like?

I've experimented a bit with chatGPT and visual AI generators like DALL-E and Midjourney. AI technologies like these are becoming increasingly sophisticated, and it's exciting to see how they can be leveraged in a variety of different ways. I mean looking at the integration of chatGPT into Bing - these things are going to become much more mainstream.

AI chatbots like ChatGPT are really revolutionizing the way we interact with computers and access information. By allowing us to converse with a machine in a conversational way. If that makes sense.

It is also impressive and scary at the same time how powerful these tools can be. I played around with some image generation tools and created some of the following images: 1) A generic poster I thought of. 2) D&D Character based on the description @AFKG-Curtis gave me. 3) D&D Character based on the description of my own.

I know there are even more tools out there being created. Video generation, Music generation, Code generation and more. My job is already talking about how we are going to try and integrate options like chatGPT to assist in certain areas - and it really sounds impressive.

I truly believe that the AI train has left the station, and we need to get on board. It is all going to be about how we can take advantage of these new amazing tools.

[Axxemann]

I haven't knowingly used them, and I'm dead set against them. I prefer human interaction with knowledgeable people.

[ajlueke]

You can play around with "Stable Diffusion" and train it with your own images.  It will come up with some pretty amazing artificial selfies.

It does create some image rights issues.  Like if I have a bunch of pictures with a friend in them, I could train an AI and then make artificial images of that person.  Is that right to do without that persons consent?  etc. etc.  Still fun to play around with though.

[mengelag]

I'm honestly really scared of the narrative that is used to base the algorithms on. If you start with bias on certain ideas the AI will follow that bias. For example, the output could be more Liberal leaning or more Conservative or WOKE leaning. That's a problem from the ground up I think when it comes to AI. You just don't know where the information is coming from and if it's reliable. They've proved already that ChatGPT makes quite a few mistakes. The whole thing is scary to me. Hopefully, I'll be long gone by the time it becomes Skynet....

[Sam_AMD]

I'm torn between loving and "disliking" AI. Part of me believes in and supports AI. It is the future. I remember one of my software development professors saying that the best engineers are those that work themselves out of the job ("...as we strive to simplify our world with technology, we will eventually find ways to automate processes to the point we will automate ourselves!") I believe this has been happening, now more than ever. I believe AI will benefit everyone in the long term.

What scares me is losing control at every level (from personal beliefs and relationships to tools and cars for example). Do you remember the 1982 movie Blade Runner? I may be pushing my imagination too far, but I believe we will get there one day. 

@Key-J  you said "I know there are even more tools out there being created. Video generation, Music generation, Code generation and more." I agree...but, well, why not go further and say "Human generation" 

[ScornfulWizard]

The AI will take over our world eventually by creeping into all the possible spheres and aspects of our lives, pass the point of no return sometime later, and finally will wipe the humanity off the face of the Earth. I'd love to live to see myself die to the AI in the most dystopian way possible.

[BigAl01]

I suspect we are doomed as a species, but I could be wrong.

[Justifier]

I think the have the potential to be extremely helpful, and like anything that has that potential they will also be abused substantially along the way by the people who make the AI themselves

I think the biggest draw to them right now is how many people are trying to limit creative expression via legal methods

Artists

Musicians

Engineers

Legislators (Law Fields)

Streamers, YouTubers, Content Creators, Enthusiasts wanting to make wallpaper art for pictures or backdrops, or thumbnails, or background music, heck groups of kids who want to start a band and want their own personal band cover but no connections with artists or personal talent themselves, and further no money to pay them if they did

All of them are limited in one way or another from any possible litany of barriers: overprotective artists who think they are owed the world for their work or who think the world doesn't deserves their work, law, copyright, etc

If you've ever heard of "The Death Effect", I've always personally thought it was attributable purely to the absolute abrasiveness of artists. It's not that people do not like/appreciate those people's work, it's that their personalities or attitudes of possessiveness/protectiveness toward their work is so bad/annoying that when people come into contact with them it completely spoils the perspective of whatever beautiful thing they've managed to create

I'd also say that AI in itself is *EXTREMELY* likely to have to go through that same process, since the people and groups behind it seem to have the same personalities as artists -- the people and companies who made it will have to pass away before it can actually flourish and enrich society, and until that happens it will be regarded with hesitancy, and possibly with disgust

[mengelag]

You make a really good point here @Justifier , AI combined with Mass Control, that's a wicked combination...

[Vynski]

I watched a Marine training video concerning the droid reconnaissance robots.  The marines won!

One grunt dressed up as a tree and walked right up to the robot without being detected.  Two others decided to each get into a large cardboard box.  They didn't walk they ran up to the droids undetected, snicker and laughing along the way.  

So my opinion is that AI has a long way to go to out-smart a Marine. 

A Navy Vietnam VET.

[BigAl01]

I did like those dancing robots in the Super Bowl commercial last night.

[Vynski]

I didn't even watch it.  I did get a kick out of that training video though.

[tredeuce]

My thoughts are that tools are only as useful as the person wielding them. I also think that Artificial Intelligence should never be confused with Artificial Intelligent. I think the availability of AI tools to the masses is another chance for pioneers and innovators to make their mark in the world and/or seek their fortune.

[52aje48d]

David Chalmers, a professor at NYU, spoke on November 28th, 2022, at one of the most prestigious AI gatherings, Neural Information Processing Systems (NeurIPS — an earlier public version of the talk is here). He would not, however, discuss his machine learning studies. Chalmers is not an AI scholar; rather, he is a philosopher. Instead, he had been asked to talk about consciousness, a topic that most serious AI researchers had previously avoided. AI awareness is currently a mainstream topic: AI awareness was recently referred to as "the last word" by The New York Times.

People have been imagining what it would be like to build artificial beings that think and feel exactly like us for hundreds of years. In some respects, artificial intelligence (AI) is a continuation of that pattern. But is consciousness a legitimate issue or a red herring? Should it really influence the way we approach AI?

[Sam_AMD]

Wow! Great point! Thank you for sharing. I'm gonna look up David Chalmers. 

And welcome to Red Team @52aje48d !

[52aje48d]

thanks for more information 

[Slow_Jams]

I don't have experience with ChatGPT yet, but I am dissapointed that my superior google search skills will soon be obsoleced by this new technology. 😂

I have been playing around with the image generators.  A bit for fun but also as to help plant the seeds for ingenuity.  As i've aged my imagination for new ideas sometimes stalls based on my education and experience.  Sometimes its nice to have a fresh concept in design that may be completely improbably but also gets the mind moving.  I'm hoping to achieve similar success with ChatGPT once I have the time to begin experimenting.

[joebangles]

AI has some amazing potential. There's a video from boyinaband with some scary implications.

Supposedly, ChatGPT has already passed a law exam.

I've heard (about 3rd hand) that a high school kid tried to submit an AI generated term-paper, but was detected when the teacher decided that the "tone" didn't sound like the student's previous work, and all the footnote references were made up

There's a lawsuit against microsoft's co-pilot (copilotlitigation?) -- microsoft seeded the code generation engine using copyrighted code stored on their github "free" repository, but did not attribute back copyright to the original authors. (if any action results from this, there's a possibility that microsoft will need to either re-seed and/or re-train their copilot engine to retain attribution. That could be VERY expensive. OTOH, github probably has a clause in their free hosting contract that makes you give them permission to use any code you post. It'd be a typical Microsoft move. There's reasons NOT to use Office 365!)

I've heard that the artistic programs are under suit as well.

AI may be quite useful, but it should be kept under human curation... there will be plenty of motivations NOT to, however. Elon Musk has cast warnings about the dangers of AI, but has offered to take over use of AI. What? Elon wants a monopoly on playing with fire.

Oh, yes, and Team Green are all in on machine learning.

@Key-J did you get anything out of the art experimentation that you discarded? This would be where human curation is potentially giving the AI feedback on what people like and do not like.

Call me cautiously optimistic on AI as a whole

[Key-J]

---

@joebangles wrote:

@Key-J did you get anything out of the art experimentation that you discarded? This would be where human curation is potentially giving the AI feedback on what people like and do not like.

---

I felt like I did get something out of it, for example with the prompts that I gave it for mine and Curtis' DnD characters, they were pretty spot on of what we were thinking. I haven't used it too much other than that. It really was just something I used to mess around on. However, since then I have been following a lot of AI "expert" accounts on Twitter, and it is really amazing what they can come up with. Learning how to prompt is going to be such an incredibly useful skill for people.

[rdenmon]

I just think that AI will be used in bad ways. I just think of the Boston Dynamics robot dog and robot person with AI...Very scary potential. 

[larrydarren]

Ask it some serious questions, like how to cure cancer, the meaning of life.

Mainly just to guage where it's at. Haha, you never know.

[red5]

One of the members of my company's Board of Directors keeps bugging me about AI.  I have a trial account for ChatGPT and it's underwhelming.  For example, if I ask it a question about part prices online it comes back and tells me it can't do that.  !!!  LOL, that's helpful.  People just need to realize that on the backend it's still all about programming and available data. 

[Key-J]

We have something similar going on here. Not sure of your industry, but if you can take a look at integrating ChatGPT4's API locally into your network, and have it source scan all of your website as an example.

Then show it to your Board of Directors and ask them to ask anything to the system about the most minute detail, or complex item on your website and see what response it gives.

This is where I've seen some testing, and it is really impressive.

[red5]

I have a new inventory/data analyst coming on board next month. He was an intern with us previously. We've talked about it a bit.  We may go down the rabbit hole once he starts.  We're in the parts business. Predominantly for heavy trucks, mining, cranes, etc.  We use a chat tool on our portal store called Tawkto, and that's something our board member was interested in replacing with AI.  Personally I think we have many other issues to be looking at, but I'll do my due diligence. I'm certainly open to anything that brings value to the company.

[Sam_AMD]

I played around with DALL-E 2 and to my surprise, it worked. I asked it to create an image of a PC case with a GPU installed. I needed this image for one of the "How-to" articles I've been posting in our PC Building How-to Articles repository. 

This is what it produced:

Not bad, I'm sure it would have been better if I spent the time to describe the image in more detail (with more descriptors). 

Anyways, did you all read the book "1984"...know what I mean???

[Key-J]

Not bad at all. You should give Midjourney a try next time - unfortunately, it is Discord integrated, which I know some people might not enjoy.

The quality of images compared to DALL-E is incredible.

[red5]

No, but I had the Van Halen album 1984. Is that close enough? 😎  Couldn't resist