iiinova — where your digital self keeps evolving

Reposted by Thomas Brox Røst

Daniel Pietersen @pietersender.bsky.social · 20/09/2026

A rare photograph of Reviewer Two

Reposted by Thomas Brox Røst

Emma Vogel @efvogel.bsky.social · 17/09/2026

I love visuals that bring data to life, so I was very excited when I was asked to create animations for the @uitnorgesarktiske.bsky.social Museum showing the transatlantic scale of humpback whale migrations! This is just a subset of the data I am working with for @bioprotectproject.bsky.social🐋🌍🧪🌊

16125

Reposted by Thomas Brox Røst

Margaret Mitchell @mmitchell.bsky.social · 06/07/2026

If you haven't read this piece on language use in around AI, you definitely should. Great work from @emilymbender.bsky.social , @nannainie.bsky.social and Peter Zukerman, with solutions for what to call things we backoff to just anthropomorphising. firstmonday.org/ojs/index.ph...

Table 3: Suggestions for de-anthropomorphizing language.
Category, Alternative Strategies, Examples.

Cognizer & Products of cognition: Instead of language that locates thinking in an algorithm, describe the algorithm as performing calculations or other algorithmic operations and the people using it doing the thinking. Examples:
artificial intelligence → probabilistic automation

hybrid intelligence → augmented human intelligence

image recognition → image labeling

speech recognition → automatic transcription

the model shows bias → the model reflects bias

model mistakes → model errors

chatbots are good at … → chatbots are good for ….

912644

Reposted by Thomas Brox Røst

Nina Willburger @drnwillburger.bsky.social · 13/06/2026

Just 789 more signatures to hit 20,000! A significant threshold - come on, we can do it!

511266

Reposted by Thomas Brox Røst

Margaret Mitchell @mmitchell.bsky.social · 02/02/2026

Without her work, the “intricately accurate” navigation and timing of GPS would not have been possible. ❤️

35020

Reposted by Thomas Brox Røst

Gergely Orosz @gergely.pragmaticengineer.com · 21/01/2026

How is the world's largest storage system, AWS S3 built? Mai-Lan Tomsen Bukovec, who has been heading up S3 for 13 years (!) S3's scale is something else (tens of millions of hard drives, eleven 9s of durability (!!) and heavy usage of formal methods, amongst others.) (cont'd)

2547

Reposted by Thomas Brox Røst

The Journal of Immaterial Science @immaterialscience.bsky.social · 25/07/2025

We turned medicinal chemistry into a card game, which takes 3-10 years to play. PDF: jabde.com/wp-content/u...

01510

Reposted by Thomas Brox Røst

Melissa Chan @melissakchan.bsky.social · 02/07/2025

Me, to every young journalist I ever meet.

09715

Reposted by Thomas Brox Røst

Harvard T.H. Chan School of Public Health @hsph.harvard.edu · 16/05/2025

With nearly all of Harvard Chan School’s direct federal grants terminated, we are relying on philanthropy to power our research and support our educational programs. Every gift, regardless of size, advances our vision of health, dignity, and justice for every human. Support our work: hsph.me/whygive

13274161

Reposted by Thomas Brox Røst

Hetan Shah @hetanshah.bsky.social · 03/05/2025

‘It offers ASMR in a world of ADHD.’ So many great lines in this piece on snooker economist.com/britain/2025...

economist.com

The strange success of snooker

Immigration, agglomeration and amorality keep the sport going

2164

Reposted by Thomas Brox Røst

Tom Gauld @tomgauld.bsky.social · 28/04/2025

My latest @newscientist.com cartoon

Title:
SOME GRAMMATICAL VOICES FOR USE IN SCIENTIFIC WRITING

ACTIVE VOICE
e.g.
Our team collected samples and then we tested them.

CLICKBAIT VOICE
e.g.
We collected some samples...
You won't believe what happened next!

PASSIVE VOICE
e.g.
Samples were collected and tested.

HAIKU VOICE
e.g.
Quiet science lab.
Workers arrive with samples.
The testing begins.

PASSIVE-AGGRESSIVE VOICE
e.g. We did all the collecting and testing.
No need to thank us.
Just doing our job.

CONSPIRACY VOICE
e.g.
Mysterious "Samples" were harvested and covertly "tested" by so-called scientists.

7250041893

Reposted by Thomas Brox Røst

MIT Press @mitpress.bsky.social · 10/04/2025

Before the invention of the gummed envelope in the 1830s, almost all letter writers had to use letterlocking. "Letterlocking" provides an archival history of the practice, including sources so readers can learn how to make locked letters for their own correspondence: mitpress.mit.edu/978026204927...

"Letterlocking: The Hidden History of the Letter" on a white background. The cover is black featuring an image of an unsealed locked letter. Text denotes the authors are Jana Dambrogio and Daniel Starza Smith with the Unlocking History Research Group.

0439

Reposted by Thomas Brox Røst

The New Yorker @newyorker.com · 09/04/2025

Today’s Daily Cartoon, by Lisa Rothstein. #NewYorkerCartoons nyer.cm/QAnwGsj

665578787

Reposted by Thomas Brox Røst

Katie Mack @astrokatie.com · 07/04/2025

Check out tonight's livestream of Derek Muller (Veritasium on YouTube) giving the @perimeterinstitute.ca Public Lecture on the topic "How Will AI Change Education?" It starts at 7pm Eastern! www.youtube.com/live/PYeBGyX...

youtube.com

Veritasium: How Will AI Change Education? Derek Muller on the Future of Learning.

YouTube video by Perimeter Institute for Theoretical Physics

1029232

Reposted by Thomas Brox Røst

Ketan Joshi @ketanjoshi.co · 29/03/2025

What a heartbreaking photograph - the impact of satellite constellations on the night sky. Taken above the Pinnacles in Nambung National Park, Western Australia. Credit: Joshua Rozells

Photograph of the Sky with streaks of light cutting across it in patterns from satellites

391636635

Reposted by Thomas Brox Røst

David Woodruff @dmwoodruff.bsky.social · 27/03/2025

A very old Soviet joke, from an especially dark time: Foxes are fleeing the USSR in droves. Q: Why are you running away? Fox: The Soviets passed a new law that they’re going to arrest all camels. Q: But you’re foxes! Fox: Yeah, why don’t *you* try proving to the NKVD that you’re not a camel.

110136364874

Reposted by Thomas Brox Røst

Mark Cuban @mcuban.bsky.social · 22/03/2025

From 2010 to 2016 (latest data I have ), NIH research contributed to EVERY drug approved by the FDA

708319688473

Reposted by Thomas Brox Røst

Ludwig Boltzmann Institute for Digital Health and Prevention @lbi-dhp.bsky.social · 20/03/2025

📢 Excited to share our paper "The Last JITAI?" was accepted to #CHI2025! We explored how LLMs can transform Just-in-Time Adaptive Interventions. GPT-4 outperformed humans in JITAI generation signaling a shift toward Generative Context-Aware Computing! #DigitalHealth #ScienceForSociety

142

Reposted by Thomas Brox Røst

Stephanie Carvin @stephaniecarvin.bsky.social · 25/02/2025

Hard to imagine how counterproductive this would be. It would dramatically increase threats to the US, given shared role in continental/cyber defence. But given many of Canada’s key contributions relate to Russia, & you don’t see Russia as a threat anymore… A reckless and dumb trade tactic.

1621265

Reposted by Thomas Brox Røst

Alvin @alviebarr.bsky.social · 21/02/2025

You can create synthetic data with a single prompt! We present a zero-shot approach to synthetic data generation which can outperform a GAN in preserving correlations, means, 95% CI and privacy of real-world data (without access to data) Link: arxiv.org/abs/2502.14523

083

Reposted by Thomas Brox Røst

Alvin @alviebarr.bsky.social · 14/02/2025

Super excited to share our latest preprint: We generate synthetic neurosurgical data with an LLM (using a single prompt) and train an ML model on it to predict real-world neurosurgical outcomes Read more: arxiv.org/pdf/2502.09566 @emreszg.bsky.social

0124

Reposted by Thomas Brox Røst

Zach Weinersmith @zachweinersmith.bsky.social · 13/02/2025

Yeats talking smack again though perhaps not anticipating that in the future nobody would care about poetry unless it was uplifting short slogans with line breaks.

5638

Reposted by Thomas Brox Røst

Eric Topol @erictopol.bsky.social · 04/02/2025

New: The largest medical A.I. randomized controlled trial yet performed, enrolling >100,000 women undergoing mammography screening The use of AI led to 29% higher detection of cancer, no increase of false positives, and reduced workload compared with radiologists w/o AI thelancet.com/journals/lan...

thelancet.com

Screening performance and characteristics of breast cancer detected in the Mammography Screening with Artificial Intelligence trial (MASAI): a randomised, controlled, parallel-group, non-inferiority, ...

The findings suggest that AI contributes to the early detection of clinically relevant breast cancer and reduces screen-reading workload without increasing false positives.

381348346

Thomas Brox Røst @brox1974.bsky.social · 03/02/2025

“The number of smart founders is not a function of the amount of money available.” ... "LPs of American VC firms have made compound annual returns of 1.45% in the past three years". www.economist.com/business/202...

economist.com

Has Sequoia Capital outgrown its business model?

Venture capital’s hardiest perennial gets back to its roots

010

Reposted by Thomas Brox Røst

404 Media @404media.co · 30/01/2025

Archivists are working to identify and save thousands of datasets disappearing from Data.gov 🔗 www.404media.co/archivists-w...

81151322

Reposted by Thomas Brox Røst

Alex Hern @hern.bsky.social · 31/01/2025

i like asking deepseek questions which look like logical puzzles but are actually just nonsense and watching it go insane

1121647

Reposted by Thomas Brox Røst

Dave Karpf @davekarpf.bsky.social · 28/01/2025

This is a golden opportunity for cancer researchers to seize the potential of Web3. Launch a memecoin to fund your study. It's The Future!

10847

Thomas Brox Røst @brox1974.bsky.social · 24/01/2025

"Imagine, for example, that you keep your books clumped together, leaving empty space on the far right of the shelf. Then, if you add a book by Isabel Allende to your collection, you might have to move every book on the shelf to make room for it." www.quantamagazine.org/new-book-sor...

quantamagazine.org

New Book-Sorting Algorithm Almost Reaches Perfection | Quanta Magazine

The library sorting problem is used across computer science for organizing far more than just books. A new solution is less than a page-width away from the theoretical ideal.

010

Reposted by Thomas Brox Røst

The Onion @theonion.com · 17/01/2025

Schlubs From U.S., China Meet In Lowest-Level Talks theonion.com/schlubs...

373448292

Thomas Brox Røst @brox1974.bsky.social · 17/01/2025

"[F]ine-tuning LLMs with medical data can expose the training data used during learning to adversarial attacks. This issue is particularly important as medical data contain sensitive and identifiable patient data." ai.nejm.org/doi/full/10....

ai.nejm.org

Fine-Tuning LLMs with Medical Data: Can Safety Be Ensured?

Developing large-scale language models (LLMs) for health care requires fine-tuning with health care domain data suitable for downstream tasks. However, fine-tuning LLMs with medical data can expose...

010

Thomas Brox Røst @brox1974.bsky.social · 17/01/2025

"Minimal research has been published to understand the net benefit on efficiency and cognitive burden when physicians’ efforts are shifted from generating new content to reviewing content generated by LLMs." ai.nejm.org/doi/full/10....

ai.nejm.org

The Burden of Reviewing LLM-Generated Content

Large language models (LLMs) in health care aim to reduce the cognitive and administrative burden for health care professionals. Yet the accuracy and completeness of LLM-generated material is still...

000

Thomas Brox Røst @brox1974.bsky.social · 17/01/2025

"Patient perspective is important not only for patient-facing applications but for provider-facing applications, especially those intended to support shared decision-making." www.thieme-connect.de/products/ejo...

thieme-connect.de

Nothing for Me or About Me, Without Me: Codesign of Clinical Decision Support

Thieme E-Books & E-Journals

000

Reposted by Thomas Brox Røst

Matthew Facciani @matthewfacciani.bsky.social · 06/01/2025

New study finds that frequent use of AI tools encourages offloading cognitive tasks and reduces critical thinking. Higher AI usage correlated with lower critical thinking skills, especially in younger users. www.mdpi.com/2075-4698/15...

mdpi.com

AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking

The proliferation of artificial intelligence (AI) tools has transformed numerous aspects of daily life, yet its impact on critical thinking remains underexplored. This study investigates the relations...

681373689

Reposted by Thomas Brox Røst

Simon Willison @simonwillison.net · 04/01/2025

Mini thread here answering a question about how I manage to produce so much content on my blog

31189

Reposted by Thomas Brox Røst

Birmingham 81 @birmingham81.bsky.social · 30/12/2024

To celebrate Jeff Lynne's birthday, here's my brother's dog doing his impression of the ELO frontman

919118

Reposted by Thomas Brox Røst

Philip Mai @philipmai.com · 26/12/2024

But, of course, it is. Russia says it's using bitcoin to evade sanctions and sidestep traditional restrictions imposed by governments and institutions. www.axios.com/2024/12/25/r...

axios.com

Russia says it's using bitcoin to evade sanctions

The Kremlin recently passed a new crypto framework that allows international payments in crypto.

34558170

Thomas Brox Røst @brox1974.bsky.social · 23/12/2024

"Despite their public commitment to learning, in private many executives I meet express skepticism about the quality and effectiveness of it in their workplace. “How do I know this has any impact on my company’s transformation?” I often hear." hbr.org/2025/01/thre...

hbr.org

Three Ways to Lead Learning

Today the development of new skills, knowledge, and mindsets is critical to most businesses’ key strategic efforts. This has given rise to a new breed of senior leaders—leaders of learning—who advocat...

010

Thomas Brox Røst @brox1974.bsky.social · 23/12/2024

"Chegg is one of an emerging group of firms that have already been pummelled by Chatgpt-like generative ai. [...] These emerging ai victims offer clues as to when a business may be strengthened by the technology and when it will be upended." www.economist.com/business/202...

economist.com

What ChatGPT’s corporate victims have in common

The first casualties of generative AI offer lessons for other businesses

010

Thomas Brox Røst @brox1974.bsky.social · 23/12/2024

"LLM-generated EM-to-IP handoff notes were determined superior compared with physician-written summaries via conventional automated evaluation methods, but marginally inferior in usefulness and safety via a novel evaluation framework." jamanetwork.com/journals/jam...

jamanetwork.com

Developing and Evaluating LLM-Generated Emergency Medicine Handoff Notes

This cohort study evaluates large language model (LLM)–generated emergency medicine–to-inpatient handoff notes and evaluates their accuracy and safety compared with physician-written notes.

010

Thomas Brox Røst @brox1974.bsky.social · 20/12/2024

"For simple research goals and datasets, a fully autonomous cycle can create manuscripts that independently recapitulate the findings of peer-reviewed biomedical publications without major errors in about 80 to 90% of cases." (1/2) ai.nejm.org/doi/full/10....

ai.nejm.org

Autonomous LLM-Driven Research — from Data to Human-Verifiable Research Papers

Artificial intelligence (AI) promises to accelerate scientific discovery, but it remains unclear whether AI systems can perform fully autonomous research, and whether they can do so while adhering ...

100

Reposted by Thomas Brox Røst

Adam Johnson @adamj.eu · 18/12/2024

You can join the party on GitHub, from $5/month. But get your orgranization to sign up to give more, to pay back for the framework that underpins your project! github.com/sponsors/dja...

github.com

Sponsor @django on GitHub Sponsors

021

Reposted by Thomas Brox Røst

Charlie Warzel @cwarzel.bsky.social · 11/12/2024

I wrote about crypto. The knock has been that the technology has few use cases. But, post-election I think crypto's legacy is clearer. Its killer app is creating a money/influence flywheel for a specific type of reactionary anti-institutionalist to influence politics. It's gonna get weird (gift)

theatlantic.com

Crypto’s Legacy Is Finally Clear

An anti-establishment technology for an anti-establishment age

28552151

Thomas Brox Røst @brox1974.bsky.social · 03/12/2024

On LLMs and VLMS for games: "NetHack is so hard it seems (today, autumn of 2024) to be a giant brick wall with the best systems getting scores of between 1% and 2% on it." importai.substack.com/p/import-ai-...

importai.substack.com

Import AI 393: 10B distributed training run; China VS the chip embargo; and moral hazards of AI development

Are you sure you're doing more than curve fitting?

011

Thomas Brox Røst @brox1974.bsky.social · 03/12/2024

"In this cross-sectional study of radiologist burnout, frequent AI use was associated with an increased risk of radiologist burnout, particularly among those with high workload or lower AI acceptance." jamanetwork.com/journals/jam...

jamanetwork.com

Artificial Intelligence and Radiologist Burnout

This cross-sectional study investigates whether there is an association between artificial intelligence (AI) use in radiology and radiologist burnout in China.

000

Reposted by Thomas Brox Røst

Ana Marie Cox @anamariecox.com · 01/12/2024

I don’t know who needs to hear this but you are on this earth for a reason. It may be you are someone’s best friend, it may be you are the caretaker to a hurt animal, it may be you smiled to someone on the bus & so they made it through the day. But you are needed. Thank you for existing.

1333645647