Tomnipresent

[Tomnipresent @tomnipresent]

@pikilipita Wow, been dreaming of an in-bowser version of AE for so long now. You actually did it! Hats off to you! Will be checking it out.

[Tomnipresent @tomnipresent]

@Origamiplan @LettingGo_Now @MichaelG_3D It’s certainly used in art yes. You might even call it an art style, or a decision as part of the art directing.

[Tomnipresent @tomnipresent]

@Origamiplan @LettingGo_Now @MichaelG_3D Did you skip learning about art and just go straight to seeing yourself as an art critic? 🤣 btw you’ve missed the point entirely and don’t yet understand stylised 3D rendering.

[Tomnipresent @tomnipresent]

@jiggywastaken @ThatDeadBlade The upside is the demand for hands-on creative and crafted editing will go up, I think, there will always be clients who need really refined, bespoke work from an experienced editor.

[Tomnipresent @tomnipresent]

@AustinBauwens @Oddernod @SchoolOfMotion To chime in, from an editors perspective, cutting on or to the beat serves a *purpose* in an edit. beyond just flow and music - in the viewer’s eye. Ideally you want to be sure of that purpose within the context of the whole edit rather than give off something unintentionally.

[Tomnipresent @tomnipresent]

@eyedesyn “Less chance” of getting sued 😂 oh cool great thx Adobe 👌

[Tomnipresent @tomnipresent]

@JamesLee03 Hey @BlacePlugins have a plugin called mask promoter which uses the ‘segment anything’ model to (somewhat) auto roto sections. Gives you mattes etc. Might be what you’re looking for. Used it and it worked for my use case.

[Tomnipresent @tomnipresent]

@MartinNebelong @krlosdice @FuChowChinaTown Think they just mean a workflow that… works. Something usable in production. Rather than just content for ai chat.

[Tomnipresent @tomnipresent]

Been waiting for this! 🙌

[Winbush @JonathanWinbush]

UNREAL MOTION DESIGN IS OFFICIAL HERE!
https://t.co/VRQhD6WkVk

[Tomnipresent @tomnipresent]

@levelsio If that was the case tho it would be generating 3D data, ie point cloud or geometry, but it’s not, it’s ‘just’ generating 2D pixels a la midjourney and DALL-E etc. I think they’ve confirmed as much in the paper.

[Tomnipresent @tomnipresent]

@nickfloats This is really interesting, the training data is very suspect. Besides openai being vague about specifics of what data the models are trained on. This to me looks like the same data set as midjourney (ie everything they could get their hands on).

[Tomnipresent @tomnipresent]

@Oddernod I’ve always quite liked that about your tweets!

[Tomnipresent @tomnipresent]

@AIandDesign @Mr_AllenT Depth information can be used in a few ways; it’s essentially replicating distance from camera - to add some camera movement where there wasn’t any before, to create a faux 3D space / perspective, to add FX based on distance like depth of field

[Tomnipresent @tomnipresent]

@blizaine @Mr_AllenT Blace Plugins already made something like this for video if that’s what you’re after? @BlacePlugins https://t.co/CUSQY1hPd4

[Tomnipresent @tomnipresent]

@USCTrojansFTFO @Rahll Do you license work that gets used in the deliverables for your client? Stock footage, images, audio etc that are used in the end product?

[Tomnipresent @tomnipresent]

@nickfloats @dustinthedad Is this just to move as fast as possible away from image generation now they’ve been rumbled about the copyright infringement?

[Tomnipresent @tomnipresent]

@visarga @NecroKuma3 @GaryMarcus @theinformation Completely different. An artist getting inspiration from art, or a writer getting inspiration from reading, is vastly different than a company using unlicensed work within their software.

[Tomnipresent @tomnipresent]

@aifurryart @artemlucis @JonLamArt Where did the training data come from?

[Tomnipresent @tomnipresent]

@kaihenthoiwane @Oddernod I think they made it when they generated the image. But what it means to have “made it” is maybe more the questionable area.

[Tomnipresent @tomnipresent]

@that1detectiv3 @GhillieYT Why? They’re doing just fine by all accounts.