My impression is that AI coding is on a "pick two of three" triangle: scope, ship speed, and correctness. Small tools are (speed + correctness) fast prototypes are (speed + scope) and big projects are (scope + correctness). Every project that tries to be all three degrades to "fast prototype."

Q&A with @devezer.bsky.social about metascience, open science, and her work in this area.

Q & A with Berna Devezer

Critical Metascience

markrubin.substack.com

> Mark: What’s your take on the replication crisis? > Berna: It’s past time we stopped using the phrase altogether. hear hear!

Mark Rubin@markrubin.bsky.social · 3d ago

Q&A with @devezer.bsky.social about metascience, open science, and her work in this area.

New blog post! Odds ratios are...a bit weird. Here I explain two problems (interpretation, non-collapsibility) and provide some simple recommendations for improved reporting. www.the100.ci/2026/07/30/r...

Reviewer notes: Odds ratios are really odd

When psychological researchers investigate binary outcomes, they routinely report odds ratios to quantify the strength of an effect or an association. This is quite understandable given that psycholog...

the100.ci

From Terry Tao’s ICM lecture. While there are additional arrows later that aren’t shown here, verification informs exposition. If an AI proof is autoformalized before it is understood and reworked by a human, the result will be more conceptually brittle even if evidently correct. /1

A thoughtful piece on what it means to be a postdoc: *independence is about moving forward without all answers *use science conversations to make sense of uncertainty, beyond presenting results *a postdoc is as much about learning to think & adapt as it is about generating data & publishing

eLife@elife.bsky.social · 2d ago

A postdoc comes with challenges and uncertainties different from those of a PhD. In this ecrLife blog Vartika Sharma shares how managing expectations and building a community are key skills that can help us navigate the new uncertainties of a postdoc.

Using LLMs should allow us to think bigger, harder to grasp, difficult to execute, longer to develop, ideas. That idea that’s been in you brain since forever but you hadn’t had the time because you had to retool and what not. Trust yourself. Go for that shit. Now is the fucking time.

Clément Canonne@ccanonne.github.io · 4d ago

"We're all worried," as what it means to do research (in my field, Theoretical CS) seems to be shifting, and shifting fast. What to do? Senior researchers must lead by example, knowing that not everything will pan out. What I'm suggesting below may not work everywhere, but here's my own advice: 1/

I'm cleaning out old weekend projects that I've worked on over the years. Here's one: A python package wrapping a shader that renders proteins in a certain, well-known style, or at least my interpretation of it (the hint is in the name). Have fun. github.com/daenuprobst/...

📉 The AI jobs apocalypse that isn't Dario Amodei warned half of entry level office jobs could vanish. A new report finds it still isn't showing up in the labour figures. 🔗 theguardian.com/technology/2026/jul/25/ai-jobs-apocalypse-human-labor #AI #Jobs #Anthropic

The AI jobs apocalypse probably isn’t coming anytime soon

Artificial intelligence may not deliver on its promise of vast economic opportunity at a price that humanity is willing to pay

theguardian.com

This piece on the "load-bearing internet people" and how AI-code writing is overwhelming the web's unheralded maintainers by @samlearner.bsky.social is absolutely terrific: brilliantly accessible without being patronising:

Who cleans up after the vibe-coding party?

Our obsession with AI code-writing tools is overwhelming the web’s unsung human caretakers

ft.com

One thing that I no longer hear much talk about: Whether AI would help with formal verification go mainstream. Formally verifying programs is pretty hard and is a niche skillset! Doesn't seem like AI has made a difference in that field. Worth asking: why?

When @molliesimpson.bsky.social started reporting on the renamed Bolton University two years ago, I had *no idea* it would turn into one of the biggest scandals in the history of UK higher education. Proper, proper reporting of the kind that a functioning society relies upon.

The Mill@manchestermill.bsky.social · 4w ago

Exclusive: Vice-chancellor George Holmes among those arrested on suspicion of fraud. It marks an unprecedented moment in British higher education - the first time a university vice-chancellor has ever been arrested in the UK as part of a criminal investigation. manchestermill.co.uk/exclusive-vi...

Liz Kendall on what the British government’s role in the era of AI is for SovereignAI’s initial Substack. DSIT ended up in a really good place by the end of the Starmer era IMV: it shed some of its Kyle era “intellectual property intellectual schmoperty” boosterism while retaining strong MoSes:

24 Months

Liz Kendall MP, Secretary of State for Science, Innovation and Technology

uksovereignai.substack.com

Please come and join us! Jobs for machine learning and mathy types!

UCL Research Department of Cell & Developmental Biology@ucl-cdb.bsky.social · 4w ago

📢 We're hiring! The Cell & Developmental Biology department (CDB) and other departments in Bioscience @ UCL are delighted to be recruiting for several Lecturer / Associate Professor positions. 👉 View full details and apply here: www.ucl.ac.uk/life-science...

I tried to write about this a few months ago but I am still unsatisfied by it. you don't have to like them but you absolutely need to know what they can do in an unvarnished way. my discord has a rule: We Don't Argue With The Marketing

The abstraction you didn't ask for

When I say 'generative AI isn't going away,' people hear 'and you have to like it.' You don't, and you might be right not to. But the is-ought divide here is real and we should all be preparing for bo...

ed3d.net

🐔 Brian Bucklew 🐔 ₑͤ＞∿＜ₑͤ ∞🌮@unormal.bsky.social · 4w ago

The truth is I think the luddites are not going win the fight here because they've staked their flag on not keeping clear sight of the enemy, which is... not what sun tzu would suggest.

9/ Looking back, the reviewer was not the biggest problem. Reviewers can be wrong. Reviewers can be unreasonable. Reviewers can be openly hostile. That is precisely why editors exist. The problem begins when inappropriate behaviour is acknowledged, action is promised, and then nothing changes.

Wait what about Posh Spice and Boy George doing a Christmas charity single as Posh George with a remake of Making Plans for Nigel

we constantly rely on unreliable tools (some of them in meat suits and with hair) by adding compensating controls around their actions so the outputs are bounded and, while not deterministic, predictable in the real world, this is normal. but now *you* have to look at it, and some people hate that

I. M. Noone@ianmnoone.bsky.social · last mo.

I only take issue with the final quote: “People need to manage the competing dynamics of relying on generative AI and staying mindfully vigilant.” Simply put, there is no way in which one can rely on an unreliable tool, especially when it robs us of the very ability to recognize when it's wrong.

"My favourite book? I'd probably say 'The Best of Library'"

Hari Kunzru@harikunzru.bsky.social · last mo.

The $400m flying bribe has shelves of fake books, some titled “Library”.

We are recruiting! We are seeking a Research Technical Specialist in Mass Spectrometry. The main purpose of this position is to provide general and specialist research support including leading on method development in relation to Mass Spectrometry.

UCL – University College London

UCL is consistently ranked as one of the top ten universities in the world (QS World University Rankings 2010-2022) and is No.2 in the UK for research power (Research Excellence Framework 2021).

ucl.ac.uk

‘Big changes’: Crick to stop paying for hybrid open access. www.researchprofessionalnews.com/rr-news-uk-o...

‘Big changes’: Crick to stop paying for hybrid open access - Research Professional News

Leading biomedical research institute says it will “go full in on green and gold”

researchprofessionalnews.com

Continuing the marathon clearout of my parents house. Came across this interesting cutting from a 1993 edition of the Radio Times about a programme called Two Seconds to Midnight. It predicts that in 2043 electricity will be generated by renewables and that we will use TVs for our shopping.

'On average the raw lifetime earnings gap for those aged 37 in 2022 who went to university, and those who could have but did not, was £290,000. And this includes a 3.5 per cent discount rate (applying Treasury Green Book rules) that lowers the value of earnings in later life'. 1/3

Higher education study still makes financial sense for most

DK examines what the Institute for Fiscal Studies has done with a few extra years of data on graduate earnings, and on a new approach to comparing graduates to their non-graduate peers

wonkhe.com

hearing that if she's been stolen, when the police catch the thief it'll be their biggest collar ever

NBC News@nbcnews.com · last mo.

She has a neck the size of an adult human, legs that are just as long and an unforgettable coat. And she’s missing.

Web-based modeling platforms can enhance collaboration between modelers and experimentalists during early model development. This Community Page by Marvin van Aalst, Alienor Lahlou &co provides guiding principles on how to build agile interactive #modeling tools. 🧪#AcademicSky bit.ly/4ahZ7Nm

Incredible: vaccines can literally prevent cancer! Truly one of humanity’s greatest achievements. Now let’s do some of the other awful viruses that seed chronic disease.

HPV vaccine means young women now have 'close to zero' risk of cervical cancer death

A new study finds that hundreds of lives have been saved since school-age girls were offered the HPV jab in 2008.

bbc.com

Your periodic reminder that the non-profit publisher @biologists.bsky.social funds grants that support meetings in the areas of its journals: @jcellsci.bsky.social @dev-journal.bsky.social @jexpbiol.bsky.social @dmmjournal.bsky.social 1/2 🧪 www.biologists.com/grants/scien...

Scientific Meeting Grants

Our grants support and encourage the sharing of knowledge throughout the community by facilitating international collaboration, event attendance and the organisation of scientific meetings, conference...

biologists.com

I'm attending a digital humanities event in Montreal. There was a keynote on AI and something about the talk made me wonder if it had been written by Claude. I said as much in the Q&A. As I posed the question, the speaker shifted, looking slightly uncomfortable. What he said next shocked the room +