“‘We will be talking to these AI assistants all the time,’ LeCun [Meta’s chief AI scientist] said. ‘Our entire digital diet will be mediated by AI systems.’” An important reason why Big tech companies are rushing #generativeAI is the hope of establish themselves as the ultimate gatekeepers https://www.ft.com/content/78834fd4-c4d1-4bab-bc40-a64ad9d65e0d

@hynek I haven’t. Looks interesting, it automatically finds new versions to download? Will check it out :)

@hynek Thanks, I use direnv since I read your first TIL about it! I also use Python.org installers on Mac and wrote a little tool to download, install, and update those from the command line: https://github.com/sbaack/pymac

Can't really recommend this Arte documentary on generative #AI (too many white salesmen selling European fairy tales) EXCEPT for the sterling contribution of Dr. Lumumba-Kasongo (@1h06m)

She says, of her own field, music: "It's so much about experimentation. It's so much about trying out weird things — and AI is so much about averaging. It's the people who are invested in playing in that space of the anomaly and playing with the unexpected who will continue to thrive"

https://www.arte.tv/en/videos/115067-000-A/smart-new-world/

With #Reddit selling its data to AI companies, it’s worth pointing out that it seems near impossible to prevent your posts on the #Fediverse from being used as AI training data. Currently, the only way to block AI crawlers is per domain via robot.txt. Your Mastodon server might block Google’s AI crawler, but that's unlikely to be effective when users from other servers that don't block it are boosting you or if you otherwise interact with them

https://techcrunch.com/2024/02/22/reddit-says-its-made-203m-so-far-licensing-its-data/

@blog @cutterkom Thanks for sharing! You end with saying: “Github stops all running actions after six hours.” Do you mean that this is a problem for your project, or it’s just something to be generally aware of when others want to adapt your approach?

Let me find a way to critique without being acerbic. I don't like how people popup to tell me they disagree about how to define the fediverse. But that's not important

What I do think is important is how hard people work to gatekeep and squeeze down the actual promise and potential of the fediverse. It's exactly the kind of pedantic myopia that reminds us that the nerds will not lead us to the future we want. That's not what they do.

Six forms of AI-driven dehumanization, and six forms of cognitive scientist resistance

thanks @emilymbender for sharing this preprint of https://doi.org/10.1177/09637214231217286

(preprint at https://faculty.washington.edu/ebender/papers/Bender-2024-preprint.pdf)

I'm still thinking about how mastodon should be a hot bed of developer a activity trying to improve the platform. But it's not. Partly because of a small minority of users who flame anyone who tries to do anything. And partly because control of the core was captured by a small, insular group of people who don't know how to foster collaborative growth. So much potential being wasted.

The current #Spam campaign on Mastodon shows that we desperately need:
1. an option that makes it impossible for anyone to contact me that I do not follow or have not replied to myself (not only a "disable notifications" option).
2. no private mentions, ever
3. allow-lists instead of block-lists. Make it possible to create my own little safe place of the web, please.
4. tools that support intercepting everything else by the moderator (and adding things to the allow lists if appropriate)

“All of this Sora content is obviously going to flood social media channels, making it harder for everyday people to distinguish between what’s real and what’s fake, and making the internet, in general, a whole lot more annoying. I don’t think it requires a global panel of experts to figure that out.”
https://gizmodo.com/openais-sora-is-a-giant-f-ck-you-to-reality-1851261587

Why isn’t AI doing the tedious shit for creative people instead of doing the creative shit for tedious people

“Search engines began as maps but have now become monsters. They devour territories that they are supposed to represent, relentlessly guiding users toward the place where the mapmaker can maximize profits, rather than where the user really wants to go. This is one of the major drivers of what Doctorow calls “enshittification.” And LLMs are in some ways a much more powerful (though as yet less reliable) generator of summarizations”
https://www.programmablemutter.com/p/the-political-economy-of-ai

One of the most powerful and underrated communication skills is the ability to ask an open-ended question and then stop talking so that the other person can answer. Sounds simple, but if you listen to conversations, interviews, teachers, etc. — almost no one can do it.

“Research has shown that cynicism is more likely to be negatively associated with intelligence.” https://mastodon.social/@debcha/111930331263496599

“Websites large and small are faced with a decision: submit to the AI revolution or stand their ground against it. For those that choose to opt out, their most powerful weapon is an agreement made three decades ago by some of the web’s earliest and most optimistic true believers.” https://mastodon.social/@verge/111930547295745301

man i hate the mastodon/bluesky fighting

A sad truth is Mastodon will
never offer true account mobility and true consent when the terms of service change until it implements some level of content migration.

#Mastodon #Fediverse :mastodon:​

"Six months in, journalist-owned tech publication 404 Media is profitable"

Very happy about that! @404mediaco’s tech reporting is among the best, especially their work on generative AI is mandatory reading for me https://mastodon.social/@niemanlab/111920108755779339

Generative AI has taken the reproducibility crisis of science and turned it into a core feature