I already read the articles your wrote on the topic, and it was a delight to watch you deliver the content in this format! 🥰 Thank you for creating it! I love the silky flowy background, and how perfectly blending the color of your shirt is with it! Some of those semicircles have identical color. 🤩
I like the analogy that ‘unlearning’ is like removing sugar from a cake once baked. Instead of trying to ‘learn’ Linked-in content with the nightmare of requiring unlearning/relearning everything if someone deletes a post which contributed to the model... why not use something more like Retrieval Augmented Generation (RAG) with the Retrieval only based on the content currently approved to be ‘in’ the training set..this is easy to manage and no ‘unlearning’ required. I guess RAG does not generalise as well, as writes content on what considers to be the most relevant linked-in posts/documents.
💯 Thank you Cassie. Here's the message we get in Canada when clicking the one shot opt out link. I'll be sure to share this video and the link you provided. "Setting not available" "This setting is not available in your location because we are not training generative AI for content creation using member data from your region. Learn more"
I suspect the 10% who choose to stay "In", are missing the point which is what they're planning on doing with it (content generation) which competes with the user themselves. You're essentially 'opting in' to training your own replacement. I can't wrap my head around anyone willingly wanting to train their own replacement. Hence, my only take is they're missing the nuance of the fine print.
I don't support all automation features, but if training on user data makes machine learning tools capable of replacing creatives, won't that be a problem we'll have to find a resolution for? Might as well escalate the issue by providing it good data and seeing the full potential we can get machine learning to, so we can start identifying the appropriate solution to the labour market.
Hi Cassie! Big fan of your content🙏🏻 I wanted to know if you think its fair to copy content from other creators on RUclips and make shorts out of the podcasts for example and start earning from those videos? AI is significantly used to achieve this. Is this an acceptable approach according to you?
One thing is what is said, other is what is done. The user can not see the code. Your Data is a sequence of 1 and 0. Humans, in an unconsciousness, storage all your inputs since your born and before. Unlearn is not the correct word
LinkedIn was a great platform until AI scientists came along and told us that it's okay that our artwork can be used for free to train models. Now this platform has become unbearable for artists. Artists may not have a deep understanding of AI models, but scientists clearly understand nothing about art...
Linked-In AI will learn how to make content from you and replace you eventually, so all needed from you is to just buy a premium. AI HR will recruit AI employees🤣
I already read the articles your wrote on the topic, and it was a delight to watch you deliver the content in this format! 🥰 Thank you for creating it!
I love the silky flowy background, and how perfectly blending the color of your shirt is with it! Some of those semicircles have identical color. 🤩
I saw ur post....and thats when I understood the implications and consequences of that opt in.....thanks for that post
I like the analogy that ‘unlearning’ is like removing sugar from a cake once baked. Instead of trying to ‘learn’ Linked-in content with the nightmare of requiring unlearning/relearning everything if someone deletes a post which contributed to the model... why not use something more like Retrieval Augmented Generation (RAG) with the Retrieval only based on the content currently approved to be ‘in’ the training set..this is easy to manage and no ‘unlearning’ required. I guess RAG does not generalise as well, as writes content on what considers to be the most relevant linked-in posts/documents.
💯 Thank you Cassie. Here's the message we get in Canada when clicking the one shot opt out link. I'll be sure to share this video and the link you provided.
"Setting not available"
"This setting is not available in your location because we are not training generative AI for content creation using member data from your region.
Learn more"
Vile. They are far from the only ones doing this but they need to be held to account
I’m confused. Why is it okay for OpenAI or Anthropic to scrape material from LinkedIn, but not for LinkedIn to do it themselves?
Hey Cassie, Thanks for the heads up!
Thank you Cassie. Love your content. We live in "interesting times" .
The rumors were true. People still use LinkedIn
I suspect the 10% who choose to stay "In", are missing the point which is what they're planning on doing with it (content generation) which competes with the user themselves. You're essentially 'opting in' to training your own replacement.
I can't wrap my head around anyone willingly wanting to train their own replacement. Hence, my only take is they're missing the nuance of the fine print.
I don't support all automation features, but if training on user data makes machine learning tools capable of replacing creatives, won't that be a problem we'll have to find a resolution for? Might as well escalate the issue by providing it good data and seeing the full potential we can get machine learning to, so we can start identifying the appropriate solution to the labour market.
Hi Cassie! Big fan of your content🙏🏻
I wanted to know if you think its fair to copy content from other creators on RUclips and make shorts out of the podcasts for example and start earning from those videos? AI is significantly used to achieve this. Is this an acceptable approach according to you?
While I'm not a fan, exact copies are not as bad as generative clones at scale. RUclips is still honoring its creators but we'll see what happens.
One thing is what is said, other is what is done. The user can not see the code. Your Data is a sequence of 1 and 0. Humans, in an unconsciousness, storage all your inputs since your born and before. Unlearn is not the correct word
We never stop learning, as well as we are always changing. Memory is an expansion phenomenon. Bergson
LinkedIn was a great platform until AI scientists came along and told us that it's okay that our artwork can be used for free to train models. Now this platform has become unbearable for artists. Artists may not have a deep understanding of AI models, but scientists clearly understand nothing about art...
Linked-In AI will learn how to make content from you and replace you eventually, so all needed from you is to just buy a premium. AI HR will recruit AI employees🤣
Как говорится, добро пожаловать в Будущее 🤔
Considdering I'm subversive AF an much of my Linked in content is Anti-Ai and Anti-Big Corprate I'm neither as surprised nor as Mad as I could be 😎☕💋