Author | Lightnews

Reposted by The Flaky Wanderer

ellie lockhart (she/they) @eleanor.lockhart.contact · 11h

sincerely people are massively deluding themselves if they think copyright is going to be a problem for the AI industry long term

most of the suits that can get people paid have already happened. at this point there's so much synthetic data & licensing that it's a dead issue basically

6 7 50

The Flaky Wanderer @flakywanderer.bsky.social · 11h

*Uses crude approximations of second-order methods using modified Newton Schulz iteration*

Them:

An orange colored 3d rendered cartoon bird, with overlaid text "Listen here, you little shit"

2

The Flaky Wanderer @flakywanderer.bsky.social · 12h

They've become Portable Network Graphics?

1 3

The Flaky Wanderer @flakywanderer.bsky.social · 1d

Poke

The Flaky Wanderer @flakywanderer.bsky.social · 1d

You get them by happenstance if you are as good at doing random sampling of the populace as any competent poller

1 1

Reposted by The Flaky Wanderer

fffanatic06 @fffanatic06.bsky.social · 1d

a really easy way to tell that BlueSky's anti-AI circlejerk doesn't have any actual ethical issues with AI is that they're defending Zitron working with DoNotPay.

Liz Fong-Jones (方禮真) @lizthegrey.com · 3d

what the fuck?!!!! @kathryntewson.bsky.social, get a load of this, Ed Zitron reps DoNotPay. www.wired.com/story/ai-pr-...

Ed Zitron Gets Paid to Love AI. He Also Gets Paid to Hate AI

He’s one of the loudest voices of the AI haters—even as he does PR for AI companies. Either way, Ed Zitron has your attention.

www.wired.com

3 4 16

Reposted by The Flaky Wanderer

Isaiah Bishop @isaiahbishop.bsky.social · 1d

Once again asking people to use their brains and not their identities to understand content. Its possible to have better AI critique than ed zitron or whatever

Adverb @adverb.bsky.social · 1d

Visit their website and tell me you don't feel slimy just looking at it: nomi.ai

They are not "using their own data" in any sense -- their image model knows ghibli well. Their language model knows flickr. They are either using pretrained models that use web data, APIs, or their own web data.

1 2 18

Reposted by The Flaky Wanderer

blights u @bitesu.kawaii.social · 2d

the luddites were not pro workers rights they were not a labor movement they were artisans and workshop owners and guild members who were financially threatened by industry because factories could produce textiles far cheaper than their poorly paid apprentices stop with this ahistorical nonsense

Jo Miles @jomiles.bsky.social · 2d

The Luddites weren't anti-progress. They weren't anti-technology. They were pro worker rights. And it's convenient for the tech barons (in any age) to make us believe those are the same thing.

7 54 290

The Flaky Wanderer @flakywanderer.bsky.social · 2d

And does it have a decay rate for contextual distance (time, different person, different thread, etc)?

The Flaky Wanderer @flakywanderer.bsky.social · 2d

How is "I've said it before" defined? Embedding similarity?

1

Reposted by The Flaky Wanderer

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Me reconnecting with nature (going to the graveyard and cooking the dead corpses to denature their proteins for fungi to more easily access, then reburying them in bare soil)

1 1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

I was thinking someone with a staff

1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

You can probably take that as a lower bound

1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Stockfish uses a NNUE as an evaluator for positions, and training it (with a training set, not self play) takes 2 weeks on a single computer

1 1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Even failures can be highly productive

Reposted by The Flaky Wanderer

conputer dipshit @davidcrespo.bsky.social · 2d

oh no how will people quote that extremely shitty MIT study now

Ethan Mollick @emollick.bsky.social · 2d

New data on the corporate ROI from generative AI from a large-scale tracking survey by my colleagues at Wharton.

They found that 75% already have a positive return on investment from AI, less than 5% negative. Also 46% of businesses leaders use AI daily. knowledge.wharton.upenn.edu/special-repo...

2 2 10

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Me reconnecting with nature (going to the graveyard and cooking the dead corpses to denature their proteins for fungi to more easily access, then reburying them in bare soil)

1 1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Ah right

1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Or only be able to downzone if they can find another local community that is willing to upzone an equal amount (perhaps in exchange for payment in kind)? Or would that cause issues?

1

The Flaky Wanderer @flakywanderer.bsky.social · 2d

Do you have a reference implementation to cross reference at least

1 2

The Flaky Wanderer @flakywanderer.bsky.social · 3d

Chemical reduction of acetic acid, then repeated distillation steps to try to (only mostly successfully) get rid of side product stank

youtu.be/QzP3vx8XadU?...

Turning children's glue into drinkable alcohol

YouTube video by NileRed

youtu.be

1

The Flaky Wanderer @flakywanderer.bsky.social · 3d

Food for an unpleasant bacterium is what it'll turn into

3

The Flaky Wanderer @flakywanderer.bsky.social · 3d

(as in divergence measures that aren't KL divergence)

The Flaky Wanderer @flakywanderer.bsky.social · 3d

Have people tried using non-KL divergence constraints to see if they allow for nudging the model in "deeper" ways?

2 2

Reposted by The Flaky Wanderer

dame @dame.is · 3d

ed zitron fans trying to suddenly defend his usage of and advocacy for AI is hilarious — extreme cognitive dissonance, these folks are the ones who will add me to an “AI Boosters” list just for saying AI might not be all bad

ed is quite literally an AI booster, he gets paid to boost it! i don’t!

dame @dame.is · 3d

Ed Zitron is not a journalist, he is an influencer manipulating peoples’ emotions and ignorance for personal profit.

2 3 35