iiinova — where your digital self keeps evolving

Bluesky operates on a completely opposite paradigm the AT Protocol. It is structurally decentralized, open, and public by design. Anyone can plug straight into the Bluesky "Firehose" a live, unencrypted data stream broadcasting every single public post, image, metadata link, and user relationship

000

When Elon Musk took over Twitter, he slammed the door on the open API, walled off the platform behind a mandatory login screen, and implemented aggressive rate limits specifically to stop external tech companies from scraping the timeline for free.

110

It seems that you fundamentally misunderstand the process of training an AI model you must give it input data that is extremely well detailed and in this particular case of profile pictures and Banners those wouldn't be very detailed at all so it doesn't make any sense to use those to train models

300

openai cant really train on files purely on ingest for use in pre-existing models its only possible for them to train a new one off them its possible they could harvest anything submitted but just dont use those tools on linktree?

000

its trivial for a tos update to say ingest all data from your private clouds to training or even your desktop ala microsoft recall/agentic OS features

000

the biggest threat to artists no one is even CONSIDERING right now is your cloud storage solutions if you use google drive, adobe cc, onedrive(or a windows desktop as its a onedrive based system in 11 now not local anymore) those files are already basically exclusively licenced to those providers

110

you have to have legal clause like this to enable the capacity to share data with openai purely for generative functions to work think image in to a blur algorithm. but openai is hosting it same idea

000

yeah so they mean you give us a photo to upload to openai and edit for your needs on demand. not this file will be used to train models. and DALLE is outdated. why would they use a legacy model?

300

hell agentic features will be intergrated into OSes too look at recall for christ sake its trivial to add clauses to harvest your files from your pc under microsoft or even apples control not to mention your desktop in 11 is onedrive now. not local

000

these file storage companies already have significantly more legal rights to possess and manipulate data on their servers. if you store copies or originals on a cloud its *not your cloud* remember this

100

yes openai abused artists and scraped the web but instead of panicing about these little things like this we should be working on regulations to prevent harvesting of data for genai training in the future as this WILL be a problem but not in sites like linktree think mega google drive onedrive

100

a good thread on the linktree/gen ai topic

000

while i understand the worry and concerns about these things we need to be accurate in our understanding not to mention yall should seriously consider where you store your files anyway if you use drives like google or azure/onedrive etc those companies trivially could edit their tos to train on it

000

functionally at a systematic level the difference between generative ai and say a algorithm that dynamically adds blur to a image are no different. they both are preprpgrammed mathematical functions that execute a specific purpose

100

this same exact moral panic of "artists are under attack and truth and reality are up for question" when photoshop launched.. and yet also photoshop has similar legal text in its agreements as it uses your photo as input on request to its firefly ai models. (its not learning anything from this)

210

fundementally DALL-E isnt even used anymore its a legacy model and its not possible to feed images into it to train it without editing the entire model itself during its creation of weights and such the actual burning of all its "dataset" into the model or baking in after that its static

000

"a worldwide, non-exclusive, royalty-free, sublicensable license to host, store, use, display, reproduce, modify, adapt, publish, and distribute your content for the purpose of training our machine learning models."

000

it states they use it as input if you use their generative features did you read the "eg uploading a profile photo and editing it with our ai tool" they dont legally have the rights to claim your uploaded images unless it clearly states such in the tos like as follows

100

why does trump love mcdonalds so much? well because his name is in their name

000

oh my god i know why trump likes mcdonalds so much

100

i must say this specific instance of the change is questionably worded but it doesnt scream they are feeding your images rather its "if you use our edit tools it can create a edit of your uploaded image"

000

if they did wish to "train" a model on your content youd need this kinda blarb "a worldwide, non-exclusive, royalty-free, sublicensable license to host, store, use, display, reproduce, modify, adapt, publish, and distribute your content for the purpose of training our machine learning models."

000

this is false. (fear uncertainty doubt) they claim the tos update lets linktree harvest your art and other information you submit but all the update actually claims is we are adding dalle for image and banners to generate them off user input (i.e a prompt) not anything more

221

if they were doing this itd have to state something like "a worldwide, non-exclusive, royalty-free, sublicensable license to host, store, use, display, reproduce, modify, adapt, publish, and distribute your content for the purpose of training our machine learning models."

010

it clearly states "we will use the input data to process and generate NEW images" (not verbatim) but yeah where does it say anywhere in the tos update "we reserve the right for training and content analysis and reserve a exclusive license to any and all content uploaded to linktree"

000

i hate how people kneejerk and assume its scraping your data when its not even outlined like can you people read? or is it just " OH THIS PERSON SAID THIS SO IT MUST BE TRUE! WHO EVER WOULD LIE ON THE INTERNET?"

100

bro its FUD they dont say ANYTHING about training or using your content its purely legal shit for them to have generative features by dalle intergrated

310

Toni Morton @tonimorton.bsky.social · 20/05/2026

its literally ai generated banners and pfps not stealing your art you upload? where does it say we will take licence to your works?

100

al capone would have killed for this

000

why can you see the con here but not with tom?

000

the billionare spending 130+ million on a campaign seeking massive returns lying through his teeth with populist retoric and appeals to the more progressive base while hiding his intentions and the aspect of his ad campaign being purely for saturation and availability bias is astonishing

000

i cant wait till he discards us like a used napkin

100

idealizing a politician is just like really thinking the stripper likes you...

200

effectively killing the free tier access to LLMs and moating it for billionares also cutting up the energy monopolies in ca? nah more like selling its scraps to private equity

100

does styer really? or is it a case of strongman saviour complex when he says "ive been there in the rooms i am your best shot at fighting these people, only i know the way forward" (not verbatim but roughly) his brother is also pushing open ai to start a ai regulatory agency that will tax tokens

200

ditching a candidate with a proven legislative track record (Porter) for a billionaire who is saying the right words but avoiding the same hard stances. Steyer gets a pass because he bought the megaphone..

100

@hasanabirepeater.bsky.social hasan is supporting styer softly yet i promise you when styer fucks california he will pretend like he never believed in him.. lmao

100

this is a active campaign as of today wed April 22 2026

000

and they also only run on first execution of the app itself seemingly not the installer?

000

all are following similar signs to different payloads

100