iiinova — where your digital self keeps evolving

Dr Nadiah Kristensen @nadiah.bsky.social · 13/06/2026

The whole article is very interesting

Dr Nadiah Kristensen @nadiah.bsky.social · 13/06/2026

My goal here though is not to speed up but to make richer. If anything it slows me down (more to check and learn), so I don’t think that counts as what people usually mean by a “productivity gain”.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 13/06/2026

This matches my experience that Claude is only useful for modelling work when its outputs pass first through my own understanding, which is also one way to catch its mishallucinations From: www.bostonreview.net/articles/kno...

No one seriously doubted that the arguments underlying the computer calculations were correct; the objection instead hinged on the separation of proof from our old friend understanding, as in this reaction by British mathematician Frank Bonsall: “We cannot possibly achieve what I regard as the essential element of a proof—our own personal understanding—if part of the argument is hidden away in a box.”

110

Dr Nadiah Kristensen @nadiah.bsky.social · 04/03/2026

Knuth: "Shock! Shock! I learned yesterday that an open problem I'd been working on for several weeks had just been solved by Claude Opus 4.6 ... What a joy it is to learn not only that my conjecture has a nice solution but also to celebrate this dramatic advance in automatic deduction."

www-cs-faculty.stanford.edu

001

Dr Nadiah Kristensen @nadiah.bsky.social · 02/03/2026

On a headcount basis, approximately 40% of university staff are casual. Source: andrewnorton.id.au/2021/04/16/h...

020

Dr Nadiah Kristensen @nadiah.bsky.social · 02/03/2026

Proportions of respondents by employment status: Ongoing/Continuing (78%), Fixed Term (15%), Casual (6%), and Honorary appointment (1%).

110

Dr Nadiah Kristensen @nadiah.bsky.social · 02/03/2026

Sorry, this is a little vague. I can’t refer back to the email for details because I had some big feelings about it and immediately deleted it.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 02/03/2026

I got an email celebrating a Faculty wellbeing initiative. Some lovely colleagues have been volunteering to do things like deliver flowers and citrus fruits to the tearooms to lift our spirits.

110

Dr Nadiah Kristensen @nadiah.bsky.social · 23/02/2026

(My contribution was the analytic and semi-analytic portions of the paper)

000

Dr Nadiah Kristensen @nadiah.bsky.social · 23/02/2026

You'd think requiring everyone to agree to the fishing level would protect fish better, but this new paper shows consensus-based voting can lead to worse conservation outcomes when the fishery was already highly exploited. Congratulations to Irina Lerner on her first paper 🎉

conbio.onlinelibrary.wiley.com

Voting Systems Influence Conservation Outcomes for International Fisheries

Fishery stocks that range partly or entirely beyond national jurisdictions pose substantial coordination challenges. Regional Fishery Management Organizations (RFMOs) were created to coordinate the s...

150

Dr Nadiah Kristensen @nadiah.bsky.social · 30/01/2026

We need $484 billion/year for nature-based climate solutions by 2030. But how do you convince investors when a single cyclone can erase years of restoration? Jordan Holdorf's new paper shows adaptive timing of investments can make risky projects financially viable.

onlinelibrary.wiley.com

Adaptive temporal optimization of restoration investments with blue carbon returns and climate risks

This framework provides a tool for conservation scientists to develop adaptive investment plans for restoration projects under climate risk. Using project-specific data, it shows how carbon markets c...

001

Dr Nadiah Kristensen @nadiah.bsky.social · 28/12/2025

To be clear, I have no idea what the answer is to all of this

000

Dr Nadiah Kristensen @nadiah.bsky.social · 28/12/2025

It doesn't have to be this way (sincerely)

110

Dr Nadiah Kristensen @nadiah.bsky.social · 26/12/2025

Apparently Claude has been using Grokipedia for at least a month bsky.app/profile/dana...

1197

Dr Nadiah Kristensen @nadiah.bsky.social · 26/12/2025

Thread summarising an arxiv preprint investigating what Grokipedia changed from Wikipedia by @harold.bsky.social and @mantzarlis.com bsky.app/profile/haro...

131

Dr Nadiah Kristensen @nadiah.bsky.social · 26/12/2025

Oh! I only noticed this morning. So it has been at least a month

000

Dr Nadiah Kristensen @nadiah.bsky.social · 26/12/2025

Noticed this morning Claude also cites Grokipedia now

010

Dr Nadiah Kristensen @nadiah.bsky.social · 26/12/2025

Claude (Opus 4.5) has started citing Grokipedia in its results.

A screenshot of a Claude chat from the phone app with most of the text greyed out. The text at the top indicates Opus 4.5 is being used. The first block of text has a link indicating the content was sourced from Wikipedia. The second block has a link indicating the content was sourced from Grokipedia.

1266

Reposted by Dr Nadiah Kristensen

Shannon McSheffrey @medievalmcsheff.bsky.social · 22/12/2025

I posted this👇 to a claim that genAI is better than Ctl+F to search historical docs. @tristanpalmgren.bsky.social replied with a genAI summary of my document and @bookllyfr.bsky.social asked if it was accurate. It turns out that it did exactly what genAI does: made stuff up & got stuff wrong.🧵

612657

Dr Nadiah Kristensen @nadiah.bsky.social · 07/12/2025

That makes me really sad. They were both extremely successful in their fields. They didn't need to do that.

010

Dr Nadiah Kristensen @nadiah.bsky.social · 07/12/2025

Haha, thank you, this was helpful.

010

Dr Nadiah Kristensen @nadiah.bsky.social · 06/12/2025

Disrespecting reality seems like about 90% of what’s going wrong in the world right now

010

Dr Nadiah Kristensen @nadiah.bsky.social · 05/12/2025

I really appreciated how this article led with the context I needed to put the newsworthy part into perspective

000

Reposted by Dr Nadiah Kristensen

Charlotte Patterson @charlotterubyp.bsky.social · 21/10/2025

🚨 Excited to share our new research! 🚨 We tackle a big question in conservation: how can we make smarter use of existing surveys to learn about places and species we haven’t yet been able to survey? 🏔️ 🔗 Read the paper here: doi.org/10.1111/2041...

2228

Reposted by Dr Nadiah Kristensen

Kalyani Twyman @kztwyman.bsky.social · 18/11/2025

Happy to share that our paper ‘a formal theory of group-level adaptation for obligate eusociality’ (with @andygardner.bsky.social) is now out in @jevbio.bsky.social advances. doi.org/10.1093/jeb/... #OpenAccess #Eusociality #GroupAdaptation #FormalDarwinism

2299

Dr Nadiah Kristensen @nadiah.bsky.social · 27/11/2025

Though funnily enough, we got a really thorough review at Sci Rep. The reviewer was so careful they picked up a right author but wrong paper mis-citation. I think they even found typos in the supplement, from memory (20 pages of mostly maths). But we could have got somewhere else.

010

Dr Nadiah Kristensen @nadiah.bsky.social · 27/11/2025

A blessing for the system, not me. I really regret not going with my gut and saying “no” to submitting that paper to Scientific Reports. The Ecol. Model. ones can’t be helped because that was a respectable journal at the time, but Sci Rep has always felt dodgy. That’s a lesson learned for me.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 27/11/2025

… this could just be me being optimistic, but I think this could be a blessing in disguise.

110

Dr Nadiah Kristensen @nadiah.bsky.social · 27/11/2025

The first one I posted about 3 months ago here (still up, btw) bsky.app/profile/nadi...

100

Dr Nadiah Kristensen @nadiah.bsky.social · 27/11/2025

Awesome, this is the second journal I’ve published in that has since let through egregious AI slop.

130

Dr Nadiah Kristensen @nadiah.bsky.social · 27/11/2025

Oh awesome, that’s two journals now that I’ve published in that have since let through egregious AI slop.

130

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

Almost certainly not a kin selection model. That was a very my-own-field-biased thing to say.

000

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

Or maybe there's some other way they think about it - I'm not an economist. I guess I'll have to go find out.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

I understand they need to suppress the individual-level model for clarity and tractability, so they need to assume some mechanism like niche-construction/intergenerational kin selection that makes individuals care for group interests in the future.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

The thing that gave me pause was the assumption that groups can care about their future. Their case studies were large-scale and long-term, like the persistence of the US Constitution.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

The institution-voting is modelled as a dynamic game (Markov Voting Equilibria approach). There is a discounting factor which allows groups to take into account the effect of future institutions on their payoffs. It's exciting it captures so much about their historical case studies.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

This was the overview, so I might have the details of the model wrong, but it seems to have two layers: (1) given the institution, the policy is determined by the median voter, and (2) voting on institutions themselves, which determine shares and which groups are enfranchised.

100

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

Recently read a superstar outside my field, summarising their previous work modelling institutional change. A quite simple model, but it can capture, e.g., "strategic stability", where groups vote against a change that will benefit them bc it will open the door to future change that harms them.

sciencedirect.com

Institutional change and institutional persistence

In this essay, we provide a simple conceptual framework to elucidate the forces that lead to institutional persistence and change. Our framework is ba…

120

Dr Nadiah Kristensen @nadiah.bsky.social · 26/11/2025

Clicking <3 so others notice that I noticed you noticing the noticing of your noticing.

110

Dr Nadiah Kristensen @nadiah.bsky.social · 25/11/2025

Why is altruism more common in nature than spite? This new paper gives a general mathematical explanation: the degree of negative assortment that can be achieved is constrained in a way that positive assortment is not, particularly in unbalanced populations. www.sciencedirect.com/science/arti...

Multi panel figure illustrating the following intuition for why group structures favoring altruism might be more common than those favoring spite. It considers the example scenarios with pairs of individuals. When the population is evenly split between 100 actors (altruistic or spiteful) and 100 non-actors, maximum positive assortment produces all same-type pairs (50 pairs actor+actor, 50 pairs actor+nonactor) and maximum negative assortment produces all opposite-type pairs (100 pairs actor+nonactor). Therefore, maximum assortativity produces equally group structures most favorable to altruism and group structures most favorable to spite. However, when actors outnumber non-actors (140 actors, 60 nonactors), positive and negative assortment have asymmetric effects on group structure. Maximum positive assortment still produces all same-type pairs (70 pairs actor+actor, 30 pairs actor+nonactor); however, maximum negative assortment cannot achieve all opposite-type pairs due to the surplus of actors – some same-type pairs (actor–actor) must remain (60 pairs actor+nonactor, 40 pairs actor+actor, ). Therefore, the maximum achievable positive assortment promotes altruism more effectively than the maximum negative assortment promotes spite. A similar asymmetry arises when non-actors outnumber actors (60 actors, 140 nonactors). Maximum positive assortment again produces all same-type pairs (30 pairs actor+actor, 70 pairs actor+nonactor), but maximum negative assortment cannot achieve all opposite-type pairs due to the surplus of non-actors – some same-type pairs (non-actor– non-actor) necessarily remain (60 pairs actor+nonactor, but 40 pairs nonactor+nonactor). Thus, once again, the maximum achievable positive assortment promotes altruism more effectively than the maximum negative assortment promotes spite.

0125