Wow that’s amazing! Can you please share your recipe? Thanks

Is it? Facebook was launched in 2004 and this changed the way people communicate and advertise because of collective feedback and the rise of e-commerce. In 1999 people would read a newspaper and that impression would dissipate. There is no evidence I have seen that suggests this has no effect.

Terrible if true.

Morning all. #FriYay

Otto Loewi (1873–1961): Dreamer and Nobel laureate

To that point, I’ll add that acetylcholine came from a dream pmc.ncbi.nlm.nih.gov/articles/PMC...

Calling NeurIPS 2026 researchers! 🚀

The CFPs for the evaluations & datasets track are officially out, and we want to support you if you're working on a submission.

When you build your benchmark on Kaggle Benchmarks, you get complimentary compute, top model access, and exposure to 30M+ users.

Yes and I wouldn’t want someone using it in a harmful way, which is why best practices are a good idea. Equally if someone is going to deliberately use code in a harmful manner, then I would prefer the focus to be on that individual as opposed to AI

Absolutely & I agree, I was asking whether or not this could lead to more restrictions on developers and if anyone had considered that.

I have an R package that I developed and it’s in a private GitHub repository because I have a manuscript waiting for feedback from other authors. Quite frankly, this is a blocker to progress- a necessary one, but an outcome of collaboration.

I’m not complaining, I’m asking if anyone has thought about what would happen if the “best-practices” were monetized and what that would mean to the open source community.

That’s really good then, but when I read “best-practices” I’m naturally inclined to think about whether training is involved and then there is the question of how much training.

A scenario that firewalls apps that can be used only after training & then that gets monetized and then turns into software that can only be accessed with a grant and an agreeable PI, for instance

I’m asking about potential blockers to progress that this could raise and if there is a risk strategy for that, but model building is great also and yes a good way to pass time.

My friends in physics used to get 1 year postdocs and if you have to wait 6 months for the data, that’s a problem.

Would you prevent what happened with academic data access, having to have a “PI” sign off (often) multiple forms before someone qualified to do the work but with a different job title & actually doing the work can even begin? There are pros and cons to this system, but is this where you are heading?

This @anshulkundaje.bsky.social, Kyle Fahr and William Greenleaf paper is finally out, we've been following it for a while in its preprint form, phenomenal work.

www.nature.com/articles/s41...

This paper is brilliant, thank you for writing it.

Spectral requirements for cooperation www.biorxiv.org/content/10.64898/2026.04...

Sharing the non paywalled study referenced www.nature.com/articles/s41...

1 week left to apply for a fully funded PhD position in the UK on developing autonomous beamlines!

www.gre.ac.uk/docs/rep/com...

#chemsky #AcademicSky #PhDChat #MaterialsScience #MachineLearning #PhysicsInformedAI

I'm thrilled to announce I'm joining Purdue University as a tenured professor this August! Purdue holds a special place in my heart: its undergraduate program prepared this Indiana farm girl…... I'm thrilled to announce I'm joining Purdue University as a tenured professor this August! Purdue holds a special place in my heart: its undergraduate program prepared this Indiana farm girl remarkab...

I'm moving to Purdue University as a tenured professor this summer! Excited to hire postdocs there and get a lab up and running with @runxishen.bsky.social as co-lead.

More details: www.linkedin.com/posts/anneca...

The Association for Computing Machinery (ACM) Prize in Computing has gone to Matei Zaharia for his "visionary development of distributed data systems and computing infrastructure" that has supported large-scale machine learning, analytics and AI. 1/2

GitHub - pachterlab/edgePython: Repository for edgePython which is a Python implementation of the Bioconductor edgeR package for differential analysis of genomics count data. Repository for edgePython which is a Python implementation of the Bioconductor edgeR package for differential analysis of genomics count data. - pachterlab/edgePython

Happy to see multiple optimizations to the edgePython codebase, with more on the way: github.com/pachterlab/e...

I have already experimented with porting to Rust, and have it mostly done (with some extra speedup); useful to complete an end-to-end single-cell Rust ecosystem.

Oh look at this: @pnas.org has a Bluesky share button! 🙌🥳 :success-kid.jpg:

**And it comes before the LinkedIn one** 😁😎

If you're at @bspparasitology.bsky.social #BSP2026 next week, come say hello 😊👋 I'll be around with my @microbiologysociety.org #Microbial #Genomics Editor-in-Chief hat on, to talk all things publishing for the community 🤓 💻🧬 #MicroSky #parasites 🧪 1/2

Black and white photo of Hedwig Kohn in her laboratory. She wears a puffy white blouse with long sleeves, and a long black skirt with a high waist. Her hair appears to be pulled up in a bun, though it's hard to see clearly with dark hair against the dark background of the photo. Her right hand rests on a lab bench with various pieces of equipment set upon it.

Physicist Hedwig Kohn was born #OTD in 1887. Known for her work on flame spectroscopy and black body emissions, she was one of only three women teaching university physics in Germany at the outbreak of WWII. 🧪 ⚛️ 👩‍🔬

Image: Brenda Winnewisser / Jewish Women’s Archive @jwaonline.bsky.social

Welcome back to #StatsSunday — your statistical fix, pulled from this week’s stories.⁠ Read the story behind this stat here: reasonstobecheerful.world/the-spark-food-donations...

That’s nice. I had to leave my postdoc earlier than expected because of personal reasons, at the time I didn’t want to. It was difficult but I think if you’re doing something productive & interesting then it’s all worth it. There is a lot of knowledge transfer now between academia & industry also.

From Heard to Lived Opinions: Simulating Opinion Dynamics with Grounded LLM Agents in Economic Environments Opinion dynamics (OD) studies how individual opinions evolve and generate collective patterns such as consensus and polarization. While recent work explores OD using populations of LLM-based agents fo...

After my PhD on systems engineering AI (incl behaviour trees) I thought I'd now revolutionise ABM, but in fact learnt we want the simplest agents possible cf Ockham's razor. So what should I make of "Grounded LLM Agents"? arxiv.org/abs/2603.26701 @psmaldino.bsky.social @alexanderjstewart.bsky.social

This is the kind of paper I always looked forward to reviewing when I was a postdoc