Skip Navigation

User banner
Posts
0
Comments
29
Joined
2 yr. ago

  • What happens if you take a picture of the screen with your phone? Not ideal in terms of quality but I'm willing to bet it would work as a stupid easy bypass.

    What I'd like to see is an image format that is digitally signed by default so that modifications can be flagged and the source can be verified. Yeah people could modify them for malicious use cases (I don't think it will be possible to ever prevent that unfortunately) but at least we'd know that it wasn't the original. This doesn't work for cases where people want to be anonymous but for things like social media selfies and posts where people are visible it could be helpful in preventing people spreading deepfakes and claiming they're real.

  • Ah yes those powerless companies that have simply no choice but to destroy our biosphere.

    It's not like they can use the money that they're currently using to lobby against environmental regulation and oversight for heavier oversight, regulations and gasp* penalties instead... Yes it's literally impossible to take a principled stance at the expense of literally any amount of profit. Besides, who needs clean air and livable temperatures anyway?

    No it must be the consumer who must wade through the sea of greenwashing and propaganda spread by companies who have no problem outright lying about their carbon emissions to everyone's faces. Because it's just so easy for everyone to know the complete supply chain of every product they buy. Simple really.

    Edit: After looking through your comment history I can't help but think what you said was meant as more of a call to action for people rather than stepping to bat for multinationals as if they are helpless victims with no agency. Apologies if my comment came across as mean spirited I've just been on edge recently due to certain undisclosed life events. I'll keep my base comment as is for archival though.

  • Voice typing long messages is IMO faster than using a phone keyboard, even with gesture typing.

  • With how things are going, who's to say that a year from now google catching you running an ad-blocker at all won't cause them to ban your entire google account?

  • Pure "prompt → image" with nothing in between I absolutely agree. It's lazy and ripe for abuse by copyright trolls. That being said there's a lot more in the world of AI assisted art than what most people are aware of.

    Determining where the legal lines will be drawn is going to be a monumental task but I think there's value in allowing authors to retain copyright on AI assisted works. I also can't see the free open source models not going the way of restricting training data to public domain works like Adobe did with Firefly if that becomes a legal issue.

  • My pie in the sky hope is that copyright somehow becomes less stringent after all of this.

    Don't get me wrong I want protections for creators and support reasonable copyright (life of the author +25 years with the possibility of a 15 year extension) but letting a company lord over an IP for damn near a century isn't ideal for anyone.

  • So, what you're saying is maybe one more lane in space then? /s

    FWIW, I totally agree. It's just that I've lost hope in people after disposable single-use Li-Po battery banks. Who knows, maybe I'm wrong and the public will start to push more for public transit rather than buy the next new shiny.

    I try, but it does feel kind of hopeless.

  • Hey, if someone can actually do it in a way that:

    • Satisfies public safety requirements
    • Satisfies all legal requirements
    • Restricts usage to trained pilots or autopilot
    • Runs on electricity or other low/zero emission fuel source
    • Lasts an appreciable amount of time before recharge

    Then it may free up public roads enough to push pedestrian and bike usage while offering people a large incentive to go electric. I just don't think the technology for personal automobile-aircraft hybrids is there yet.

    I'd love to be proven wrong.

  • I just kind of assumed that they, as well as anyone in the space was doing that already.

    Whether that means that we all collectively have ownership over the outputs of these models if they're trained on content that we produced over the years is another thing. As someone who uses AI tools a fair bit I would be totally fine with generated content being public domain unless a threshold for human intervention is met.

    That threshold is where the messy legal work lies.

  • This is a list of all the open source software I have come across and use frequently to semi frequently. There will likely be some overlap with stuff everyone has already posted.

    Photography and Image manipulation

    • Darktable → RAW photo processing
    • GIMP→ Photoshop alternative
    • Krita → Digital painting (have only used it a bit, but I hear it's good)
    • Inkscape → Vector Graphics
    • Automatic1111 → Diffusion model AI toolkit (mostly Stable Diffusion but also has extensions for other diffusion based models like OpenAI's Shap-E)

    3D modeling and Printing

    • Blender → 3D Modeling, sculpting, 2D animation, compositing all rolled together (simply one of the best examples of FOSS)
    • Meshroom → Photogrammetry
    • PrusaSlicer → 3D printing slicer based on Slic3r

    Video editing and Processing

    • Kdenlive → Genuinely good video editor
    • FFMPEG → Command line media toolkit (very complex but also works on android through Termux)
    • Instant NeRF → Neural Radiance Fields, think photoscan to a 3D representation (not meant to make 3D meshes unfortunately)

    Misc

    • Calibre → E-book management
    • Serge → Self hosted Local LLM's made a bit easier to deal with
    • Firefox → Web browser

    FOSS I'm excited for

    • DragGAN → Manipulate images by intuitively dragging, more on this here and here (official code being released this month but there are already projects based on the paper with working examples)
    • CoDi → "Composable Diffusion" Any2Any conversion Txt2Vid, Vid2Audio, Audio+Txt2Img, whatever
    • Neuralangelo → Promises to be NeRF's for 3D models (don't know if it will be FOSS but I'm hopeful)