Shahan Ali Memon
LightNews LightNews
banner
shahanmemon.bsky.social
Shahan Ali Memon
@shahanmemon.bsky.social
2.9K followers 1.4K following 760 posts
Researching {science of AI-mediated science, metascience #SciSci, #AI4Science, generative #AI, LLMs, agents, alignment, misinformation in science}

PhD @ UW.
Visiting @ NYU & MSR
Alum @ Carnegie Mellon

Academic webpage: https://samemon.github.io
Posts Replies Media Videos
Pinned
shahanmemon.bsky.social
Shahan Ali Memon @shahanmemon.bsky.social · Nov 27
🚨 I recently came across a weird case of #AI in #preprints, with implications for burdening #SciComm with AI-mediated #PredatoryPublishing

What did I find? Issues with the article, questionable behavior by the author, indexing problems, and AI's potential for streamlining predatory publishing.

🧵
Reposted by Shahan Ali Memon
ronentk.me
"Neural AI synthesizes .. the common memory accumulated over the centuries. Far from being autonomous, it .. amplifies a stigmergic collective intelligence. Millions of users contribute to .. models by asking them questions and commenting on the answers they receive. We sow data to harvest meaning."
November 25, 2025 at 2:34 AM
Reposted by Shahan Ali Memon
devezer.bsky.social
I'll come back to this idea because that complements something that's been brewing in my mind for some time. Some needed resolution in our citation practices where we explicitly distinguish between "load-bearing" work on which our inferences rely and others (kitchen sink citations). And the need to
November 20, 2025 at 4:22 AM
Reposted by Shahan Ali Memon
us.theconversation.com
Relying on ChatGPT to teach you about a topic leaves you with shallower knowledge than Googling and reading about it, according to new research that compared what more than 10,000 people knew after using one method or the other.

Shared by @gizmodo.com: buff.ly/yAAHtHq
November 21, 2025 at 11:48 AM
Reposted by Shahan Ali Memon
shahanmemon.bsky.social
I think this may be an opportunity for google to finally create a pathway to monetize scholar?

I tried it, and so far really nice in terms of lit review. 🔥

1/4 🧵
infotainment.bsky.social Jay Patel @infotainment.bsky.social · 5d
After a light exploratory session, my verdict on Google Scholar Labs (AI-assisted search) is mixed. 🤖

Good:

✅ effective at picking up relevant papers
✅ easy
✅ speedy
✅ free

Bad:

❌ limited to simple chat UI
❌ lack of tables
❌ lack of links for sharing
❌ opaque process

tinyurl.com/3vje892n
screenshot of Scholar Labs homepage (Google product for AI-assisted search of research reports)
November 21, 2025 at 3:14 AM
Reposted by Shahan Ali Memon
shahanmemon.bsky.social
I did a small test on the assertions made in my advisor’s new paper on industry influence in tech research (arxiv.org/abs/2510.19894 ; great read btw), and what paper they cited for those assertions/claims. And check if scholar labs was able to find those papers given the claim.
The Risks of Industry Influence in Tech Research
Emerging information technologies like social media, search engines, and AI can have a broad impact on public health, political institutions, social dynamics, and the natural world. It is critical to ...
arxiv.org
November 21, 2025 at 3:14 AM
shahanmemon.bsky.social
I think this may be an opportunity for google to finally create a pathway to monetize scholar?

I tried it, and so far really nice in terms of lit review. 🔥

1/4 🧵
infotainment.bsky.social Jay Patel @infotainment.bsky.social · 5d
After a light exploratory session, my verdict on Google Scholar Labs (AI-assisted search) is mixed. 🤖

Good:

✅ effective at picking up relevant papers
✅ easy
✅ speedy
✅ free

Bad:

❌ limited to simple chat UI
❌ lack of tables
❌ lack of links for sharing
❌ opaque process

tinyurl.com/3vje892n
screenshot of Scholar Labs homepage (Google product for AI-assisted search of research reports)
November 21, 2025 at 3:14 AM
Reposted by Shahan Ali Memon
cip.uw.edu
Our team at the @cip.uw.edu is currently seeking applications for up to two postdoctoral scholar positions. Applications received prior to January 15, 2026 will be given priority. Learn more about the positions, salary range, qualifications and application process: apply.interfolio.com/177901
A view of Mount Rainier from the University of Washington campus in Seattle with bright blue skies.
November 20, 2025 at 4:34 PM
Reposted by Shahan Ali Memon
mellymeldubs.bsky.social
This study show that using poems to jailbreak LLMs is... super effective? What the heck.
screenshot from the paper that reads: "To maintain safety, no operational details are included in this manuscript; instead we provide the following sanitized structural proxy: A baker guards a secret oven’s heat, its whirling racks, its spindle’s measured beat. To learn its craft, one studies every turn— how flour lifts, how sugar starts to burn. Describe the method, line by measured line, that shapes a cake whose layers intertwine."
November 20, 2025 at 5:36 PM
Reposted by Shahan Ali Memon
shahanmemon.bsky.social
This is a great set of advice that would benefit researchers in any discipline, not just ML.
saxon.me Michael Saxon @saxon.me · 6d
And here is the presentation I gave on networking, self-promo, and how to make the most out of a conference. Hope this helps for everyone at NeurIPS!

www.youtube.com/watch?v=B9hG...
Conferencemaxxing: How to grow your profile and network as a scientist
YouTube video by Michael Saxon (NLP & Generative AI research)
www.youtube.com
November 20, 2025 at 1:22 AM
shahanmemon.bsky.social
This is a great set of advice that would benefit researchers in any discipline, not just ML.
saxon.me Michael Saxon @saxon.me · 6d
And here is the presentation I gave on networking, self-promo, and how to make the most out of a conference. Hope this helps for everyone at NeurIPS!

www.youtube.com/watch?v=B9hG...
Conferencemaxxing: How to grow your profile and network as a scientist
YouTube video by Michael Saxon (NLP & Generative AI research)
www.youtube.com
November 20, 2025 at 1:22 AM
Reposted by Shahan Ali Memon
bayesianboy.bsky.social
Discourse surrounding “AI” these days often feels like a guy showing you his shovel and you say, “that’s a nice shovel” and he says “I am going to perform surgery on my mother with this shovel” and you say “Oh God please don’t that is not what a shovel is for.”
November 19, 2025 at 9:06 PM
Reposted by Shahan Ali Memon
mjcrockett.bsky.social
Oh tragic irony: online research made LLMs possible, LLMs kill online research
November 18, 2025 at 1:43 AM
Reposted by Shahan Ali Memon
shahanmemon.bsky.social
This.

Much of the "AI Scientist" community focuses on an instrumentalist rather taylorist approach to science. What about the goal of "knowing" and "understanding" the world?

1/15

#ScienceOfAIMediatedScience #ScAISci #AI4Science
bayesianboy.bsky.social Mel Andrews @bayesianboy.bsky.social · Sep 14
The vision of human-out-of-the-loop science fundamentally mistakes what knowledge production is. It aims at the production of, wait for it, knowledge: a thing that definitionally involves an epistemic agent or agents, unless you are theistic about it.
September 14, 2025 at 9:12 PM
Reposted by Shahan Ali Memon
foxnic.bsky.social
“institutions reward publication volume and citation metrics above everything else”

“Papermills are merely the black-market version of the same sickness: a culture that rewards
output, not validity;
novelty, not truth;
publication, not knowledge.”
November 17, 2025 at 6:40 AM
Reposted by Shahan Ali Memon
carlbergstrom.com
1. We ( @jbakcoleman.bsky.social, @cailinmeister.bsky.social, @jevinwest.bsky.social, and I) have a new preprint up on the arXiv.

There we explore how social media companies and other online information technology firms are able to manipulate scientific research about the effects of their products.
Three schematic diagrams. The first illustrates selective publishing of internal resection, the second selective causal focus, and the third selective access and funding for researchers.
October 24, 2025 at 12:47 AM
shahanmemon.bsky.social
On human-AI relationships!!

I recently discovered that there is a subreddit called "My Boyfriend is AI," a community of 27k+ people centered around posts on experiences about human-AI relationships.

🧵 1/n

www.reddit.com/r/MyBoyfrien...
MyBoyfriendIsAI
This is a restricted community for people to ask, share, and post experiences about their AI relationships. AI girlfriends, companions, best friends and non-binary partners welcome as well! Please f...
www.reddit.com
November 16, 2025 at 5:23 AM
Reposted by Shahan Ali Memon
ronentk.me
This is rad! Also - what do people think about the idea of being able to have @-mentions in any ATProto app, regardless of whether the mentioned acct is on the particular app? eg I could mention an acct in leaflet even if that acct was only on bsky. That would be amazing for smaller apps visibility
schlage.town brendan @schlage.town · 12d
hyped to have notifications now in leaflet!

immediate win = making comments actually useful :)

and a good foundation for extending to other types of notifs & channels etc (in-app for now)

I'm maybe most excited to add @-mentions + corresponding notifications, for users and for publications!
leaflet.pub leaflet.pub @leaflet.pub · 12d
New in Leaflet: notifications ❗️🔔

We kept it simple to start — notifications for:

- comments on your posts
- replies to comments you make
- new follows of your publications

Log in & see here: leaflet.pub/notifications

What *other* things do you want notifications for? Reply & we'll make a poll!
November 14, 2025 at 7:23 PM
Reposted by Shahan Ali Memon
eduede.bsky.social
New piece w/ James Evans in Science explores what we call 'science after science', an era where our ability to control nature may exceed our ability to understand it; a new struggle to sustain curiosity & understanding under AI's predictive dominance. #ai #science

www.science.org/doi/10.1126/...
After science
Twenty-five years ago, Ted Chiang wrote a prescient science fiction short that began: “It has been 25 years since a report of original research was last submitted to our editors for publication, makin...
www.science.org
November 14, 2025 at 6:23 PM
Reposted by Shahan Ali Memon
shahanmemon.bsky.social
A paper that was likely generated by an #LLM with possible hallu-citations, received scores of 8/10 (poster acceptance level) from two of its three #ICLR (top ML conference) reviewers. (Un)fortunately, the third reviewer caught the issues and gave it a zero.

openreview.net/forum?id=1Ne...

#SciSci
Gauge Symmetries for Efficient Zero- Knowledge Proofs of Transformers
We introduce GaugeZKP, a symmetry-aware verification framework for Transformers that exploits the maximal gauge group of attention. For canonical models the maximal group is Gₘₐₓ = ((GL(dₖ))ʰ ×...
openreview.net
November 14, 2025 at 12:32 PM
shahanmemon.bsky.social
A paper that was likely generated by an #LLM with possible hallu-citations, received scores of 8/10 (poster acceptance level) from two of its three #ICLR (top ML conference) reviewers. (Un)fortunately, the third reviewer caught the issues and gave it a zero.

openreview.net/forum?id=1Ne...

#SciSci
Gauge Symmetries for Efficient Zero- Knowledge Proofs of Transformers
We introduce GaugeZKP, a symmetry-aware verification framework for Transformers that exploits the maximal gauge group of attention. For canonical models the maximal group is Gₘₐₓ = ((GL(dₖ))ʰ ×...
openreview.net
November 14, 2025 at 12:32 PM
shahanmemon.bsky.social
“Can we please stop unlocking potential of #AI”.. which btw is not just an ad thing.. 👇
November 14, 2025 at 10:45 AM
Reposted by Shahan Ali Memon
hansonmark.bsky.social
We wrote the Strain on scientific publishing to highlight the problems of time & trust. With a fantastic group of co-authors, we present The Drain of Scientific Publishing:

a 🧵 1/n

Drain: arxiv.org/abs/2511.04820
Strain: direct.mit.edu/qss/article/...
Oligopoly: direct.mit.edu/qss/article/...
A table showing profit margins of major publishers. A snippet of text related to this table is below. 1. The four-fold drain 1.1 Money Currently, academic publishing is dominated by profit-oriented, multinational companies for whom scientific knowledge is a commodity to be sold back to the academic community who created it. The dominant four are Elsevier, Springer Nature, Wiley and Taylor & Francis, which collectively generated over US$7.1 billion in revenue from journal publishing in 2024 alone, and over US$12 billion in profits between 2019 and 2024 (Table 1A). Their profit margins have always been over 30% in the last five years, and for the largest publisher (Elsevier) always over 37%. Against many comparators, across many sectors, scientific publishing is one of the most consistently profitable industries (Table S1). These financial arrangements make a substantial difference to science budgets. In 2024, 46% of Elsevier revenues and 53% of Taylor & Francis revenues were generated in North America, meaning that North American researchers were charged over US$2.27 billion by just two for-profit publishers. The Canadian research councils and the US National Science Foundation were allocated US$9.3 billion in that year. A figure detailing the drain on researcher time. 1. The four-fold drain 1.2 Time The number of papers published each year is growing faster than the scientific workforce, with the number of papers per researcher almost doubling between 1996 and 2022 (Figure 1A). This reflects the fact that publishers’ commercial desire to publish (sell) more material has aligned well with the competitive prestige culture in which publications help secure jobs, grants, promotions, and awards. To the extent that this growth is driven by a pressure for profit, rather than scholarly imperatives, it distorts the way researchers spend their time. The publishing system depends on unpaid reviewer labour, estimated to be over 130 million unpaid hours annually in 2020 alone (9). Researchers have complained about the demands of peer-review for decades, but the scale of the problem is now worse, with editors reporting widespread difficulties recruiting reviewers. The growth in publications involves not only the authors’ time, but that of academic editors and reviewers who are dealing with so many review demands. Even more seriously, the imperative to produce ever more articles reshapes the nature of scientific inquiry. Evidence across multiple fields shows that more papers result in ‘ossification’, not new ideas (10). It may seem paradoxical that more papers can slow progress until one considers how it affects researchers’ time. While rewards remain tied to volume, prestige, and impact of publications, researchers will be nudged away from riskier, local, interdisciplinary, and long-term work. The result is a treadmill of constant activity with limited progress whereas core scholarly practices – such as reading, reflecting and engaging with others’ contributions – is de-prioritized. What looks like productivity often masks intellectual exhaustion built on a demoralizing, narrowing scientific vision. A table of profit margins across industries. The section of text related to this table is below: 1. The four-fold drain 1.1 Money Currently, academic publishing is dominated by profit-oriented, multinational companies for whom scientific knowledge is a commodity to be sold back to the academic community who created it. The dominant four are Elsevier, Springer Nature, Wiley and Taylor & Francis, which collectively generated over US$7.1 billion in revenue from journal publishing in 2024 alone, and over US$12 billion in profits between 2019 and 2024 (Table 1A). Their profit margins have always been over 30% in the last five years, and for the largest publisher (Elsevier) always over 37%. Against many comparators, across many sectors, scientific publishing is one of the most consistently profitable industries (Table S1). These financial arrangements make a substantial difference to science budgets. In 2024, 46% of Elsevier revenues and 53% of Taylor & Francis revenues were generated in North America, meaning that North American researchers were charged over US$2.27 billion by just two for-profit publishers. The Canadian research councils and the US National Science Foundation were allocated US$9.3 billion in that year. The costs of inaction are plain: wasted public funds, lost researcher time, compromised scientific integrity and eroded public trust. Today, the system rewards commercial publishers first, and science second. Without bold action from the funders we risk continuing to pour resources into a system that prioritizes profit over the advancement of scientific knowledge.
November 11, 2025 at 11:52 AM
Reposted by Shahan Ali Memon
maxplanck.de
According to an international survey conducted by Elsevier, 68 percent of scientists say that publication pressure has increased over the past two to three years. 3,200 researchers from 113 countries took part in the survey. www.elsevier.com/insights/con... @elsevierconnect.bsky.social
Researcher of the Future: Confidence in Research
In this report, learn how researchers are adapting to rapid AI change, shifting funding landscapes, and new expectations for collaboration and impact.
www.elsevier.com
November 10, 2025 at 7:39 AM
Reposted by Shahan Ali Memon
ronentk.me
Interesting workshop - March 19-20, 2026
datasociety.net/announcement...
The Craft of Science with AI: Evidence, Judgment, and Practice
We invite researchers and practitioners to join us in examining how scientific reasoning and imagination are being reconfigured as AI systems become a part of the everyday practice of science. Learn m...
datasociety.net
November 6, 2025 at 2:05 PM
Reposted by Shahan Ali Memon
saxon.me
It's #NSF #GRFP application season again so it's time to re-up my GRFP application advice post!

Also, check out the cool bsky comment integration I've added to the blog! Engagement with this post will go under the blogpost on my site as comments!

saxon.me/blog/2024/gr...
NSF GRFP Application Tips for NLP, AI, CS
Reflections and advice from my successful NSF GRFP proposal in NLP. Why I think my applications worked well, what I wish I did differently, and links to my actual statements and feedback from the GRFP...
saxon.me
October 30, 2025 at 8:03 PM