brandon69

should’ve recused himself

did you see the article that says he was actually getting his ass ate?

Critical support for the russian catgirl who managed to reproduce the superconductor in her kitchen

100+ million users are using this thing for a narrow set of use cases? That’s an incredible claim that shouldn’t be believed without evidence

is this why my lk99coin isn't taking off?

sorry to cancel you but anti-worker is anti-worker, even if singular

🤝

whereas once you filter out the objectively-wrong, you're left with tried-and-true Marx-style criticisms that still apply, which I feel like is good news. instead of looking like luddites we can be like "welcome to the conversation, noob" to people considering income loss for the first time bc of AI

Right, income loss as the main criticism makes sense to me. But people get surprised by what AI can do... over and over again. SO as one of maybe a dozen total ai-knowing Left guys, I feel like I have to warn people against arguments around "AI can't _", bc you'll just have instant egg on your face

you're right, i'll never goes

Every new model is competitive with GPT-3.5 on some benchmarks (always different ones) but trails behind GPT-4 (who did all the scoring).

Still unpublished/unreviewed, & AI/ML papers are usually bad (so 2 grains of salt), but this is one of the only studies I've seen on the subject, and they claim GPT beat humans at creativity: https://t.co/OG7SLPddUG
I feel like it's not *super* surprising because GPT beats humans at other stuff too

Respectfully, hard-disagree on this. They *can* plagiarize, bc they are general-purpose and currently mostly trained on human-generated data (today), and humans can ask them to plagiarize.

But where's the evidence that they can't create, or that plagiarism even makes up most/much of their ability?

Provided to YouTube by Arts MusicIt's Okay to Feel Sad · Daniel Tiger's NeighborhoodDaniel Tiger's Neighborhood: Big Feelings℗ 2014 Fred Rogers Company under... It's Okay to Feel Sad

jfc imagine how hard this would go, if it was the hook on a diss track https://youtu.be/3vbCjUB0tXc

Devastating that the X.com web app is a lil better

Tbf she’s had *every* job

Does bluesky’s protocol just expose a Worst Posts list* for us or something**? Like can we just cut out the middlemen and do what we’re all here to do?

*it’s not just my profile

**stop saying it’s my profile

Aight time to learn how to use blocklists

Think of all the people that have felt shame and embarrassment for needing to sign their name as “X”, and whether they’ll feel welcome here having seen your posts today.

scanning for bad words…. none detected! have a good one

COPY AND SHARE!!!

I DO NOT GIVE BLUESKY OR ANY ENTITIES ASSOCIATED WITH BLUESKY PERMISSION TO USE MY POSTS, BOTH PAST AND FUTURE. THE CONTENT OF THIS ACCOUNT IS PRIVATE AND CONFIDENTIAL INFORMATION. THE VIOLATION OF PRIVACY CAN BE PUNISHED BY LAW (UCC 1-308- 1 1 308-103 AND THE BOFA STATUTE).

while true
do
llm "go nuts in bash. you're in the driver's seat now. i'll turn off the monitor" -m gpt5 | bash -e
echo aw maybe next time
done

Following | What's Hot | Popular With Friends | Popular With Hot Friends

we were butts to nuts, tell you what

By default, it could have placeholder alt text that’s like a nice message for blind folks, like “Sorry I didn’t include alt text, but I hope you have a splendid day. Cheers.”

Not to mention, adding alt text to your images helps in training AI image models. And eventually, that data makes it into AI accessibility tools that use the same data in reverse, obviating the need to write alt text in the future, hopefully

If you QT this you have to 1v1 chatgpt+ on math before you laugh