Tootfinder

Is anyone aware of publications or research on what sort of bugs LLM-generated or LLM-assisted code tends to have?
Like, we have a huge body of knowledge in the security community about how to audit human-generated codebases for the types of bugs that human developers commonly write.
But we don't have that kind of data yet (AFAIK) for the vibe-coded monstrosities all of us are going to be pentesting soon.
Gut feelings:
* There are some common threads and patterns …

US political contradictions; knowledge systems
As Trump at least partially succeeds in constructing an alternate reality for his most ardent followers, it's tempting to think of his dogma as false, in contrast to some imagined "truth" which his non-followers are smart enough to believe in. But a more nuanced view of knowledge would admit that different groups of people have different shared truths, constituting different knowledge systems which each deviate from what's objectively measurable in different ways, and in fact they each accept different standards of what is objective, so there's not really a single "ground truth" we can even compare to to determine which of these knowledge systems is "more correct" (similar problems arise even if we only care about "more useful").
To make this more concrete, we can see that e.g., competing quantum physics theories, or likewise competing religious beliefs, have no reasonable basis on which to judge between them, either in terms of "truth" or "utility." So the Trump-dogma knowledge system, although bad, morally repugnant, etc., can't so easily be dismissed as "false" in my view. "Distorted" or "malignant" or "evil" or "contradictory" are better monikers, in my opinion.
But what I'm even more interested in thinking about is: in what ways does the current American liberal "common sense" knowledge system already bear the scars of past fascist lies & contradictions? I can think of a few:
"Columbus was an explorer."
This is "factually accurate" in the same way some of Trump's propaganda is, but it's also a cruel distortion of "Columbus was a child murderer," and it's a misrepresentation that serves an evil purpose, yet which is widely taught in elementary schools today.
Another: "dropping atomic bombs on civilians in Japan was necessary to end WWII."
Perhaps in the future we'll have "family separation & the 2025 ICE crackdowns were necessary to end the immigration crisis," although I dearly hope not.
"Reparations for slavery aren't reasonable," is yet another...
I'll close this rambling with a question: what other fascist lies have you noticed that are normalized in America right now from past Trump-like leaders (or even from less overtly fascist institutions)?

Your ATs to Ts: MITRE ATT&CK Attack Technique to P-SSCRM Task Mapping
Sivana Hamer, Jacob Bowen, Md Nazmul Haque, Chris Madden, Laurie Williams
https://arxiv.org/abs/2507.18037

Your ATs to Ts: MITRE ATT&CK Attack Technique to P-SSCRM Task Mapping
The MITRE Adversarial Tactics, Techniques and Common Knowledge (MITRE ATT&CK) Attack Technique to Proactive Software Supply Chain Risk Management Framework (P-SSCRM) Task mapping described in this document helps software organizations to determine how different tasks mitigate the attack techniques of software supply chain attacks. The mapping was created through four independent strategies to find agreed-upon mappings. Because each P-SSCRM task is mapped to one or more tasks from the 10 framewo…

OpticalAging: Real-time Presbyopia Simulation for Inclusive Design via Tunable Lenses
Qing Zhang, Zixiong Su, Yoshihito Kondoh, Kazunori Asada, Thad Starner, Kai Kunze, Yuta Itoh, Jun Rekimoto
https://arxiv.org/abs/2506.19307

OpticalAging: Real-time Presbyopia Simulation for Inclusive Design via Tunable Lenses
Presbyopia, a common age-related vision condition affecting most people as they age, often remains inadequately understood by those unaffected. To help bridge the gap between abstract accessibility knowledge and a more grounded appreciation of perceptual challenges, this study presents OpticalAging, an optical see-through simulation approach. Unlike VR-based methods, OpticalAging uses dynamically controlled tunable lenses to simulate the first-person visual perspective of presbyopia's distance-…

Finance Language Model Evaluation (FLaME)
Glenn Matlin, Mika Okamoto, Huzaifa Pardawala, Yang Yang, Sudheer Chava
https://arxiv.org/abs/2506.15846 https://…

Finance Language Model Evaluation (FLaME)
Language Models (LMs) have demonstrated impressive capabilities with core Natural Language Processing (NLP) tasks. The effectiveness of LMs for highly specialized knowledge-intensive tasks in finance remains difficult to assess due to major gaps in the methodologies of existing evaluation frameworks, which have caused an erroneous belief in a far lower bound of LMs' performance on common Finance NLP (FinNLP) tasks. To demonstrate the potential of LMs for these FinNLP tasks, we present the first…

https://boydkane.com/essays/experts

Architecture is All You Need: Improving LLM Recommenders by Dropping the Text
Kevin Foley, Shaghayegh Agah, Kavya Priyanka Kakinada
https://arxiv.org/abs/2506.15833

Architecture is All You Need: Improving LLM Recommenders by Dropping the Text
In recent years, there has been an explosion of interest in the applications of large pre-trained language models (PLMs) to recommender systems, with many studies showing strong performance of PLMs on common benchmark datasets. PLM-based recommender models benefit from flexible and customizable prompting, an unlimited vocabulary of recommendable items, and general ``world knowledge'' acquired through pre-training on massive text corpora. While PLM-based recommenders show promise in settings whe…

Private Continual Counting of Unbounded Streams
Ben Jacobsen, Kassem Fawaz
https://arxiv.org/abs/2506.15018 https://arxiv.org/pdf/250…

Private Continual Counting of Unbounded Streams
We study the problem of differentially private continual counting in the unbounded setting where the input size $n$ is not known in advance. Current state-of-the-art algorithms based on optimal instantiations of the matrix mechanism cannot be directly applied here because their privacy guarantees only hold when key parameters are tuned to $n$. Using the common `doubling trick' avoids knowledge of $n$ but leads to suboptimal and non-smooth error. We solve this problem by introducing novel matrix…

DatasetAgent: A Novel Multi-Agent System for Auto-Constructing Datasets from Real-World Images
Haoran Sun, Haoyu Bian, Shaoning Zeng, Yunbo Rao, Xu Xu, Lin Mei, Jianping Gou
https://arxiv.org/abs/2507.08648

DatasetAgent: A Novel Multi-Agent System for Auto-Constructing Datasets from Real-World Images
Common knowledge indicates that the process of constructing image datasets usually depends on the time-intensive and inefficient method of manual collection and annotation. Large models offer a solution via data generation. Nonetheless, real-world data are obviously more valuable comparing to artificially intelligence generated data, particularly in constructing image datasets. For this reason, we propose a novel method for auto-constructing datasets from real-world images by a multiagent colla…

General search techniques without common knowledge for imperfect-information games, and application to superhuman Fog of War chess
Brian Hu Zhang, Tuomas Sandholm
https://arxiv.org/abs/2506.01242

General search techniques without common knowledge for imperfect-information games, and application to superhuman Fog of War chess
Since the advent of AI, games have served as progress benchmarks. Meanwhile, imperfect-information variants of chess have existed for over a century, present extreme challenges, and have been the focus of significant AI research. Beyond calculation needed in regular chess, they require reasoning about information gathering, the opponent's knowledge, signaling, etc. The most popular variant, Fog of War (FoW) chess (aka. dark chess) is a recognized challenge problem in AI after superhuman perform…

Identifiability in Unlinked Linear Regression: Some Results and Open Problems
Fadoua Balabdaoui, Martin Slawski, Jonathan Steffani
https://arxiv.org/abs/2507.14986

Identifiability in Unlinked Linear Regression: Some Results and Open Problems
A tacit assumption in classical linear regression problems is the full knowledge of the existing link between the covariates and responses. In Unlinked Linear Regression (ULR) this link is either partially or completely missing. While the reasons causing such missingness can be different, a common challenge in statistical inference is the potential non-identifiability of the regression parameter. In this note, we review the existing literature on identifiability when the $d \ge 2$ components of…

Dual Approach to Inverse Covariance Intersection Fusion
Ji\v{r}\'i Ajgl, Ond\v{r}ej Straka
https://arxiv.org/abs/2506.05955 https://

Dual Approach to Inverse Covariance Intersection Fusion
Linear fusion of estimates under the condition of no knowledge of correlation of estimation errors has reached maturity. On the other hand, various cases of partial knowledge are still active research areas. A frequent motivation is to deal with "common information" or "common noise", whatever it means. A fusion rule for a strict meaning of the former expression has already been elaborated. Despite the dual relationship, a strict meaning of the latter one has not been considered so far. The pap…

Zambia's pre-colonial writing system invented by women but almost killed by colonialism - BBC News
https://www.bbc.co.uk/news/articles/c5ye50xgw8vo

Zambia's lost language invented by women but almost killed by colonialism
How social media is pushing back against the still common idea that African societies never had a knowledge system.

Towards a Knowledge Base of Common Sustainability Weaknesses in Green Software Development
Priyavanshi Pathania, Rohit Mehra, Vibhu Saujanya Sharma, Vikrant Kaulgud, Sanjay Podder, Adam P. Burden
https://arxiv.org/abs/2506.08812

Towards a Knowledge Base of Common Sustainability Weaknesses in Green Software Development
With the climate crisis looming, engineering sustainable software systems become crucial to optimize resource utilization, minimize environmental impact, and foster a greener, more resilient digital ecosystem. For developers, getting access to automated tools that analyze code and suggest sustainabilityrelated optimizations becomes extremely important from a learning and implementation perspective. However, there is currently a dearth of such tools due to the lack of standardized knowledge, whi…

I had this idea for a Cairn lifepath generator where there are three stages of life and you roll 1d6 for your stats at each stage, and also get appropriate items.
It has not been playtested, it's barely been proofread, but I've been having a lot of fun generating guys
perchance.org/lt8m69fg35
#ttrpg #CairnRpg

Research Knowledge Graphs: the Shifting Paradigm of Scholarly Information Representation
Matth\"aus Zloch, Danilo Dess\`i, Jennifer D'Souza, Leyla Jael Castro, Benjamin Zapilko, Saurav Karmakar, Brigitte Mathiak, Markus Stocker, Wolfgang Otto, S\"oren Auer, Stefan Dietze
https://arxiv.org/abs/2506.07285

Research Knowledge Graphs: the Shifting Paradigm of Scholarly Information Representation
Sharing and reusing research artifacts, such as datasets, publications, or methods is a fundamental part of scientific activity, where heterogeneity of resources and metadata and the common practice of capturing information in unstructured publications pose crucial challenges. Reproducibility of research and finding state-of-the-art methods or data have become increasingly challenging. In this context, the concept of Research Knowledge Graphs (RKGs) has emerged, aiming at providing an easy to u…

From Developer Pairs to AI Copilots: A Comparative Study on Knowledge Transfer
Alisa Welter, Niklas Schneider, Tobias Dick, Kallistos Weis, Christof Tinnes, Marvin Wyrich, Sven Apel
https://arxiv.org/abs/2506.04785

From Developer Pairs to AI Copilots: A Comparative Study on Knowledge Transfer
Knowledge transfer is fundamental to human collaboration and is therefore common in software engineering. Pair programming is a prominent instance. With the rise of AI coding assistants, developers now not only work with human partners but also, as some claim, with AI pair programmers. Although studies confirm knowledge transfer during human pair programming, its effectiveness with AI coding assistants remains uncertain. To analyze knowledge transfer in both human-human and human-AI settings, w…

No Trade Under Verifiable Information
Spyros Galanis
https://arxiv.org/abs/2506.04944 https://arxiv.org/pdf/2506.04944

No Trade Under Verifiable Information
No trade theorems examine conditions under which agents cannot agree to disagree on the value of a security which pays according to some state of nature, thus preventing any mutual agreement to trade. A large literature has examined conditions which imply no trade, such as relaxing the common prior and common knowledge assumptions, as well as allowing for agents who are boundedly rational or ambiguity averse. We contribute to this literature by examining conditions on the private information of…

Biden's record on covid
Saw a comment along the lines of: "seems unfair to criticise Biden's record on covid, while the situation now under Trump is so much worse".
But I don't think of them as separate. Biden's government _contributed_ to how covid stands under Trump.
If Biden's lot had taken the opportunity to educate people (at least the ones open to considering science findings) that
- it's airborne like smoke
- an empty room can hold infectious virus
- air filters, UV and fresh air reduce the levels of it
- masks work better the better they fit
- you can be infectious without/before symptoms
- you're fairly likely still to be infectious for 10 days, a few people longer
- current vaccines don't stop you catching it or transmitting it
- it can mess with your immune system so you're more likely to catch other things
- vaccinated people can still get Long Covid
- it's not "mild", it's just that the damage is quiet
then even if they hadn't done anything more to address the problem, people would be in a far better position to deploy their own common sense.
And unlike funding or laws, that investment in _knowledge_ is something it would be difficult for Trump's lot to roll back.
But Biden & co chose instead to play down the risks, and explicitly or implicitly mislead people (e.g. the 5-day quarantine, which contradicts the real infectious period).
So, yes it's worse now, but they _contributed_ to how it is now. They chose to encourage misapprehensions and confusion, and the effects of that choice are still playing out now.
Not letting them off the hook on the grounds of being comparatively "less bad", when they themselves laid some of the foundations of the current state of play.
#covid #Biden #USPol #CovidIsntOver