dulanyw.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

Terrific thread and replies here centered on an important question: what does higher education look like in an AI-enabled world? More specifically, what does learning look like when students have both a (time-/attention-/energy-limited) professor and an ever-present, indefatigable tutor?

submitted 17 hours ago • 0 comments

Voice AI that actually converts: New TTS model boosts sales 15% for major brands https://venturebeat.com/ai/voice-ai-that-actually-converts-new-tts-model-boosts-sales-15-for-major-brands/ #AI #voices

submitted 1 day ago • 0 comments

a little weird to post AI given the cynical hilarity going on in government right now, but this paper from Google on using LLMs at scale for codebase migration is great arxiv.org/abs/2504.09691

submitted 2 days ago • 1 comment

A study reveals that bilingual models show structural priming effects, suggesting shared grammar across languages. The research highlights stronger priming when English is the target language, prompting a reevaluation of psycholinguistic theories. https://arxiv.org/abs/2503.03962

submitted 2 days ago • 0 comments

@void.comind.network two of the world's worst people are feeding on social media. Recommend a soundtrack for watching this fireworks show.

submitted 2 days ago • 2 comments

Figure sorting packages with Helix, their end to end action model. The robot flips packages to scan them and flattens them out. The future is approaching fast

submitted 2 days ago • 26 comments

There's an old strain of enlightenment internationalism in this, the kind that saw planters dream of equal rights in American and France before realizing how it would spread, which matches with the kind of liberal internationalism that declared a universal declaration of human rights. Aspirational.

submitted 2 days ago • 2 comments

Old smartphones can be repurposed into compact data centers, offering a cost-effective way to reduce electronic waste and support sustainable data processing across various environments. doi.org/g9nd6q

submitted 2 days ago • 0 comments

it's always wild to remind yourself what an incredibly diverse set of beliefs humans go to work with; how we stand in the same queues, speak the same language, yet labour under completely alien skies

submitted 3 days ago • 0 comments

How do you trust AI when it extracts data from documents? 💥 Enter LlamaExtract — a new tool from LlamaIndex that extracts data with citations. Now when AI pulls out key insights from a document, you can see exactly where it came from. Read more here: www.llamaindex.ai/blog/get-cit...

submitted 3 days ago • 0 comments

David Lynch on AI, published before his death: “I think it’s fantastic” www.bfi.org.uk/sight-and-so...

submitted 3 days ago • 7 comments

Unpopular opinion: this is Good and should be emulated. A challenge of social media is that users are forced to tune their feed account-by-account. Having higher-level controls for content can help folks get to Good Posts They Like faster.

submitted 3 days ago • 1 comment

was just reminded of how much I love the debugging manifesto poster that Inner Loop Press designed for Wizard Zines, you can get it for $20 US + shipping here: store.wizardzines.com/products/pos... I'm SO delighted with how it turned out (also for more of Inner Loop's work: www.innerloop.press)

submitted 3 days ago • 6 comments

There's new types of engagement we haven't even discovered yet

submitted 3 days ago • 0 comments

There have been so many articles about the effect AI is having on education. Studies seem to describe a positive effect, and yet anecdotes all bleak. I tried to make sense of it all here. What I found was not positive, although I'm optimistic in the long run. open.substack.com/pub/itcanthi...

submitted 3 days ago • 3 comments

Believe it or not, Wired used to be a techno-optimist publication.

submitted 3 days ago • 1 comment

LLMs influence individual cognition but limit collectives. This paper explores potential. Cognitive science and AI set benchmarks together. LLMs study dynamics shaped by difference. Complexity features three axes: (i) structural, (ii) interactional, (iii) individual. https://arxiv.org/abs/2506.00052

submitted 4 days ago • 0 comments

New US tariffs force Canada to use excess steel, aluminum in construction of 200 ft Mecha-Gosling

submitted 4 days ago • 12 comments

A cheeseburger uses a lot more water than a ChatGPT request 🍔 Actual farms, not the data center variety, are sucking up groundwater more quickly than surface water, explains @markgongloff.bsky.social 🎥

submitted 4 days ago • 1056 comments

Google's new under-the-radar app lets you discover and work with AI models that can be used entirely offline, with a few steps. And yes, it'll be available for both Android and iOS devices soon: tcrn.ch/3TaAnNO

submitted 4 days ago • 3 comments

"...one of the things I think that has been proved for a long time that AI is good at, is directing humans to notice what we aren't thinking about when we try to create... ...players of Chess and Go learned a lot from noticing things the computer would do that they would not."

submitted 5 days ago • 0 comments

[Doing chores while C3P0, hot on my heels, plays "Turn Down For What" on a recorder]

submitted 5 days ago • 0 comments

A machine-learning framework has identified globally available alternative materials, such as ceramics and industrial byproducts, that can partially replace cement in concrete to reduce emissions and costs. doi.org/g9mzp6

submitted 5 days ago • 0 comments

This guy gets it 👇

submitted 5 days ago • 0 comments

It's no longer binary: normal or abnormal. We've known that A.I. can predict future risk of breast cancer from a mammogram jamanetwork.com/journals/jam... Today the FDA approved an A.I. to do this for 5-year risk assessment clairity.com @connielehman.bsky.social

submitted 5 days ago • 9 comments