Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror

Comment Canâ(TM)t remember my momâ(TM)s face (Score 1) 243

I lived overseas for a year and my mom came in visit at the six-month mark. I was meeting her at the airport and I realized I had no idea what she looked like. I knew conceptionally that she had red hair and was pretty sure that she was smaller than me, and had larger than average boobs, but not oversized. That being said, the minute I saw her in the crowd I knew who she was, and I recognized her. I find it fascinating that this is unrelated to facial recognition and even more fascinating that the whole field of Study happened because some guy got in a motorcycle accident and lost the ability to visualize things. I was in my late 30s before I realize that other people could visualize things. I always kind of thought that people were faking it when they were supposed to meditate and visualize a rose or whatever.

Comment Pictures and better UI for end users. (Score 2) 108

Iâ(TM)m not sure if youâ(TM)re documenting use cases or code. If use cases. In many cases documentation is a crutch for bad UI. Most âoecommonâ actions (ie things that most users do once a week or more oftenâ should be obvious in the UI. If directions are needed they should be inline. So if you click on a reports button (for example) and then go to a complex report, the directions should be in a pop-up or on the screen. For more obscure tasks or for complex tasks outside of the control of developers of the app (such as an import from a 3rd party system) often a picture with arrows and steps is very helpful.

Comment What is AI generation? (Score 1) 45

What if open AI writes an article and no one proofs it. What if open Ai writes an article and some one proofs it What if they fix typos? What if they rewrite every word in article? What if the ai writes a thousand word summary and someone manually condenses it to 50 words? What if a human wrote those thousand words and the AI summarized it? What if someone used the ai autocomplete to while typing? What if they used grammarly? What if they asked open AI âoegive me five synonyms for stupidâ but everything else was written by hand? This is so insane. And what about art. What if you take a photo and use photoshops new features to make the image 2 pixels wider? The lines are arbitrary.

Comment Re:Only a few (Score 2) 225

copyright is super fuzzy. If I upscale an image using an AI, but the image is still the same, just higher resolution, do I lose my copyright? What if when making it I use the feature on iOS or Pixel phones that let you grab the subject out of the picture using Machine Learning/AI and then add it to something else, is that no longer copywriting able? What if I use the magic wand tool in photoshop to select something and it's using some form of Machine Learning or AI? Does that negate copyright? What if I use Gramarly to spellcheck and have their new "AI" feature turned on? What if I generate an image using AI and paint over every pixel except one? Likewise what if I generate a picture using an AI and move every pixel over one, or flip it so no pixels match? What if I train the AI from scratch with my own images or words and everything in it was copywritten by me, but it reorganizes that, or builds something new based on that and my prompts? How is that any different than me hiring someone on Fiverr to do the same work... why would one be copyrightable and one not be? I understand if your point is "down with copyright, and this is confusing so it might be a foot in the door", but really this doesn't make sense. Even the definition is wrong. AI (as we know it now) is really just autocomplete on steroids. What if a writer doesn't use AI, but while writing their 300 page book uses the autocomplete built into Google Docs or MS Word to finish a sentence or to add a period? Have they used "AI"? At what point does "auto complete" become "AI"?

Comment I'm not sure it's possible to regulate for example (Score 1) 225

* Disclose if you use an AI: The other day I had a JPEG of a client’s logo and I used Stable Diffusion AI to upscale it (it has a great upscaller). I used AI, do I need to disclose it? Do they lose their copyright on their logo? What if I use Photoshop to upscale the same logo, or the “magic wand” took with photoshop and they use some level of Machine Learning/AI to detect obejcts, or upscale? Do they need to disclose that to me every time? * Copyright: Is the logo no longer copywritten when I use an AI to upscale it? What if I didn't upscale it, but I used an AI to check my curves mathematically with predefined information, or to generate a texture? * Weapons: That will last until North Korea and China deploy it and our units are unable to respond. There are a few things that can be regulated, and it's more along the lines of how scientific papers are created: * Users must be able to legally opt out of every one of these, at least on a case-by-case basis, otherwise we won't be able to research them. * AIs must be able to cite your sources. Regardless of if it's art or written words, "why" do you think that. Where did you get your images from? * Source material that's copywritten must be compensated. Realistically it would be best if the system was like the music industry. I know it has a lot of problems but: When a song is covered, the writer gets compensation. They set the rates and get their money. They set their terms. Sometimes it's "case by case" sometimes it's "anyone can pay and use it". If someone doesn't pay or if artist doesn't want to share then it can't be used. If you publish a photo that you took on your website or social media, you should be able to get paid for it. Here's the thing, ML and AI will be as prevalent as processors. We have to have rules that scale with that, not that hold us back.

Slashdot Top Deals

"The greatest warriors are the ones who fight for peace." -- Holly Near

Working...