Tessera
Folgen

Arnav Das

@arnav7.bsky.social

My posts here will be a mishmash of personal opinions and smol curious excursions..

1237 Posts1946 Follower12665 Folgt
Iris van Rooij 💭@irisvanrooij.bsky.social · 2d ago

Follow Ajuna Soerjadi ( @ethicsofai.bsky.social ) for creative and joyful resistance. And check this out 👇✨

Ajuna Soerjadi 🌸@ethicsofai.bsky.social · 2d ago

I did a thing!! I made a rendition of the song The Man with Money in his Hands by Jessie Mazin ❤️ This song was just perfect to be made into an anti-AI anthem. "Make life easier, just give up control. It's a chemical dependance, it's a bruise on your soul" Enjoy!!!!!!! 😃 youtu.be/i6Sbu2Zf2SM?...

Naomi Saphra@nsaphra.bsky.social · 2d ago

We need a term for the eerie uncertainty about whether you’re talking to a human. I just called a hospital line that handles urgent situations, which I use routinely. Today, the receptionist had an old-school synthetic cadence, odd delay timing, and sounded like a script.

Anita Leirfall@anitaleirfall.bsky.social · 2d ago

Sounds before Words | On the evolution of human language The q. of how humans alone came to be blessed with words is not to be lightly dismissed. Language is an evolved feature of the human organism, but words don’t fossilise like bones. How to find the missing links? www.lrb.co.uk/the-paper/v4...

Francis Gooding · Rocket Science for Monkeys: Sounds before Words

So far as we know, true symbolic language is unique to the human species. The question of how we alone came to be...

lrb.co.uk

Cat Hicks@grimalkina.bsky.social · 2d ago

Sometimes I read psychology studies about educational interventions that increase adherence and they are in some ways impressive but in other ways it's a really low bar because you're basically testing "what if we helped people handle the daily life adversity of this in ANY WAY" relative to nothing

Cat Hicks@grimalkina.bsky.social · 2d ago

I feel that when you love someone who's going through chemo the healthcare system tells you a lot about the cancer stuff and next to nothing about adverse reactions and what to do about them and this seems pretty flawed to me

Sam Power@spmontecarlo.bsky.social · 2d ago

Back in my "watching Simons Institute live-streams" era (www.youtube.com/@SimonsInsti...). Very influential on me as a curious PhD student, influential once more as a guy with some free time in the summer!

Simons Institute for the Theory of Computing

The Simons Institute for the Theory of Computing is the world's leading venue for collaborative research in theoretical computer science. Established on July 1, 2012, the Institute is housed in Calvin...

youtube.com

Carl T. Bergstrom@carlbergstrom.com · 2d ago

LLMs have reached the point that they can be useful as proof assistants. Increasingly, they appear capable even of generating entire proofs that are correct. But in my experience, their best proofs are just correct. They aren't explanatory the way a good human proof is. Here's Cris Moore:

All this puzzling is very similar to what we do in mathematics. I think of a puzzle—especially a puzzle with a beautiful logical path—as an elegant proof of its own solution. It may be presented in cryptic form, or as a series of subtle signposts, but it provides just enough information to let the solver find, or build, its solution. To continue the analogy, the best proofs are not just a series of logical steps arriving at a conclusion. Such a proof might establish that the conclusion is true, but that is all it does. It provides mere knowledge, and knowledge only of that one thing. In contrast, a good proof gives us a sense of why something is true. It helps us learn, not just about that one problem, but about the larger web of mathematics: what analogies it has with other problems and other truths, how far these analogies can be stretched and where they break down. It teaches us something new about how truth works. It gives us insight and understanding.
Cat Hicks@grimalkina.bsky.social · 2d ago

Hello. If you struggle with maladaptive overachievement, and anxiety about productivity, and you experience relief at the sensation of productivity and difficulty regulating your behavior around demands for you to be productive, that doesn't make you a bad person, and that is not addiction.

Olivia Guest · Ολίβια Γκεστ@olivia.science · 2d ago

I forgot to address one type of person that also exists, those who say "don't care so much, if you can, since it makes you sick" — I hear you, and I will ask you just one thing... don't stand in our way when we fix this and don't claim you helped when you just told us to not be so soft. 1/2

Olivia Guest · Ολίβια Γκεστ@olivia.science · 3d ago

If useful, here's my attempt to process my feelings about how I feel sadly kinda — or rather very — sick from all this AI fascist nonsense as an academic. 😶‍🌫️ I'd be pleased if even one person finds something useful because there's something to be said for processing this all together, so, um, enjoy?

Prof Sam Illingworth@samillingworth.com · 2d ago

🔬 AI pathology model rivals experts Trained on 2.3 million pathology slides and clinical notes, this AI model matched or beat specialist tools at spotting cancer in prostate and breast tissue. 🔗 doi.org/10.1038/s41591-026-04521-4 #AI #Pathology #Science 🧪

End-to-end multimodal pathology foundation model with clinical dialogue - Nature Medicine

Trained on 2.3 million whole-slide images and 14 million clinical question−answer pairs, the multimodal foundation model PRISM2 matched clinical-grade cancer detection performance without task-specifi...

doi.org

Mark Riedl@markriedl.bsky.social · 2d ago

Fascinating experiment: current AI systems lack creativity to reliably pursue research arxiv.org/abs/2607.27191 - poor judgment about the bar for publishable research - uncreative responses in research design - ineffective backtracking from dead ends - poor resource awareness - instruction drift

Can AI agents conduct open-ended AI research? Early evidence from two case studies

Forecasts of explosive AI progress hinge on AI agents automating AI research. But evidence on whether agents can carry out open-ended AI research is thin. Current evaluations either test agents on nar...

arxiv.org

Prof Sam Illingworth@samillingworth.com · 2d ago

🤖 Can AI show its visual evidence? A new method makes multimodal AI models point to the exact image regions behind an answer, instead of just giving a plausible sounding guess. 🔗 arxiv.org/abs/2607.28590v1 #AI #MachineLearning

VAD: Attributing Visual Evidence for Target Reconstruction in Multimodal On-Policy Distillation

Multimodal on-policy distillation (OPD) transfers fine-grained visual knowledge by supervising student-generated trajectories with a privileged-view teacher. Yet its next-token corrections are source-...

arxiv.org