Tootfinder

A look at UAE-based Khazna, a real estate company that builds and rents out data centers and controls 70% of the country's operational data center capacity (Divsha Bhat/Rest of World)
https://restofworld.org/2025/khazna-data-center-uae/

The invisible backbone of the UAE’s AI superpower dreams
UAE's largest data center operator Khazna is challenging Western assumptions while building regional dominance.

More progress on the building of the woodwork here. Going pretty well.
But that will be it for a few days. Mostly coz I've caught a damn virus and wanna try and protect the builders from catching it but also because he's used up all the wood. More to be delivered next week.

It's time for building work. Uh ho.
The bedroom that has remained undecorated and mostly dark since I moved in more than 20 years ago will finally be stripped and rebuilt according to my specifications.
What is my specifications? Humm. Well. I live alone and nobody can tell me what to do so I'm doing it Tardis themed.
Mostly I built the shape of it in Blender last year and the tardis theme only popped out when I readdressed that design this spring with an eye to actual colors. Adding the frosted glass to the windows on the top cupboards is the main thing that sells it.
Here's the video I made for the builder.
#diy #interiorDesign #rebuild #bedroom #doctorWho #tardis

Just finished Transiruby, along with a 9k-line journal file for it. I almost got 100, but not quite; I don't have time to go back to it before the semester starts though.
If you like exploration games, it's an excellent one, with great level design & tons of secrets. It actually makes you do significant secret-finding and map-reading in order to beat the game, not just for extras or a special ending, which is something that a lot of metroidvania games since Super Metroid don't do. My one complaint is that the map system isn't perfect, and finding obscure secrets to progress is fine when the map hints at them but much less fun when it hints incorrectly (I looked one progress item up in a speedrun video because of this).
Decently cool movement mechanics, although the combat does take a back seat and almost all of the bosses are easy (I beat the final two bosses in the third and second tries respectively). I don't think that's any better or worse than a game like Nine Sols where the final boss took me hundreds of tries though; just a different flavor. The world-building isn't as rich as the more epic metroidvanias like Hollow Knight or Lone Fungus (or again, Nine Sols) but again I'm fine with that. It's just a more casual game that has really excellent level design & exploration poetics.
#AmPlaying

How xAI is building Colossus 2, which is on track to be the world's first gigawatt-scale AI datacenter after reaching 200MW in six months (Jeremie Eliahou Ontiveros/SemiAnalysis)
https://semianalysis.com/2025/09/16/xais-colossus-2-first-gigawatt-datacenter/…

xAI’s Colossus 2 – First Gigawatt Datacenter In The World, Unique RL Methodology, Capital Raise
Much has been written about xAI’s Colossus 1. The Memphis build belongs in the history books: the largest AI training cluster, erected from scratch in 122 days. With roughly 200,000 H100/H200s and …

What News Recommendation Research Did (But Mostly Didn't) Teach Us About Building A News Recommender
Karl Higley, Robin Burke, Michael D. Ekstrand, Bart P. Knijnenburg
https://arxiv.org/abs/2509.12361

What News Recommendation Research Did (But Mostly Didn't) Teach Us About Building A News Recommender
One of the goals of recommender systems research is to provide insights and methods that can be used by practitioners to build real-world systems that deliver high-quality recommendations to actual people grounded in their genuine interests and needs. We report on our experience trying to apply the news recommendation literature to build POPROX, a live platform for news recommendation research, and reflect on the extent to which the current state of research supports system-building efforts. Ou…

Building Empathy Through the Sounds of 'a World in Motion'
https://reasonstobecheerful.world/crossing-borders-music/?utm_source=flipboard&utm_content=muz4now/magazine%2…

Building Empathy Through the Sounds of 'a World in Motion'
Audiences at Crossing Borders concerts come to understand refugees and other marginalized communities in a whole new way.

Great panel with @… and Nancy Kress on rules in and for novels. On the box it said "rules and when to break them". Even thou they didn't said a single word about the breaking of rules, I learned much about writing, world-building, and the favorite clothing-colour of vikings (magenta!).

First week with the plastic cast off of my fractured wrist so I guess it's back to normal? No more sitting on the couch doing nothing. Two months off doing tarot shows, back with a marxist rant. Nice when it works that way.
Trouble is the back-to-normal won't last long, building work starts this weekend with taking apart the old bedroom and a new carpet.
https://wordcloudtarot.com/@wordcloudtarot/statuses/01K58ZYQH4SRDTQJTS9J051RSR

From Diagnosis to Improvement: Probing Spatio-Physical Reasoning in Vision Language Models
Tiancheng Han, Yunfei Gao, Yong Li, Wuzhou Yu, Qiaosheng Zhang, Wenqi Shao
https://arxiv.org/abs/2508.10770

From Diagnosis to Improvement: Probing Spatio-Physical Reasoning in Vision Language Models
Spatio-physical reasoning, a foundation capability for understanding the real physics world, is a critical step towards building robust world models. While recent vision language models (VLMs) have shown remarkable progress in specialized domains like multimodal mathematics and pure spatial understanding, their capability for spatio-physical reasoning remains largely unexplored. This paper provides a comprehensive diagnostic analysis of mainstream VLMs, revealing that current models perform ina…

Q&A with Willie Shubert, VP of Programs at environmental news outlet Mongabay, about its global newsroom, publishing news daily in many languages, and more (Rhett Butler/Mongabay)
https://news.mongabay.com/2025/10/buil

Building a global newsroom for a planet in crisis: A conversation with Willie Shubert
For Willie Shubert, journalism is a kind of cartography. “I studied geography,” he says. Mapmaking taught him how to see the world as a system. That systems view now guides his work as Executive Editor and Vice President of Programs at Mongabay, as he helps steer the environmental news organization through one of the most […]

"China starts building world’s biggest hydropower dam"
#China #Energy #Renewables

Building Bigraphs of the real world
Kang Rong Roy Ang
https://arxiv.org/abs/2508.00003 https://arxiv.org/pdf/2508.00003

Building Bigraphs of the real world
This report proposes a formal specification for organising all buildings, streets and administrative areas in the world into a hierarchical space-partitioning tree using data from OpenStreetMap. This hierarchical structure is encoded into a bigraph, serving as a digital twin of the world and capturing complete street connectivity. It presents a tool implemented in OCaml (source code at https://github.com/royangkr/bigraph-of-the-world ) that constructs bigraphs for regions from any part of the w…

Fundamentals of Building Autonomous LLM Agents
Victor de Lamo Castrillo, Habtom Kahsay Gidey, Alexander Lenz, Alois Knoll
https://arxiv.org/abs/2510.09244 https://

Fundamentals of Building Autonomous LLM Agents
This paper reviews the architecture and implementation methods of agents powered by large language models (LLMs). Motivated by the limitations of traditional LLMs in real-world tasks, the research aims to explore patterns to develop "agentic" LLMs that can automate complex tasks and bridge the performance gap with human capabilities. Key components include a perception system that converts environmental percepts into meaningful representations; a reasoning system that formulates plans, adapts t…

Building and Evaluating a Realistic Virtual World for Large Scale Urban Exploration from 360{\deg} Videos
Mizuki Takenawa, Naoki Sugimoto, Leslie W\"ohler, Satoshi Ikehata, Kiyoharu Aizawa
https://arxiv.org/abs/2510.11447

Building and Evaluating a Realistic Virtual World for Large Scale Urban Exploration from 360° Videos
We propose to build realistic virtual worlds, called 360RVW, for large urban environments directly from 360° videos. We provide an interface for interactive exploration, where users can freely navigate via their own avatars. 360° videos record the entire environment of the shooting location simultaneously leading to highly realistic and immersive representations. Our system uses 360° videos recorded along streets and builds a 360RVW through four main operations: video segmentation by interse…

Why Artists Are Striking Spotify Over Daniel Ek’s AI-Offensive Weapons Bet—and Why It Matters for AI Deals http://musictechpolicy.com/2025/09/11/why-artists-are-striking-spotify-over-daniel-eks-ai-offe…

Why Artists Are Striking Spotify Over Daniel Ek’s AI-Offensive Weapons Bet—and Why It Matters for AI Deals
Over the summer, more artists began removing catalogs from Spotify—not just over bleak royalties but over CEO Daniel Ek’s investment in Helsing, a defense-tech firm building AI-enabled offensive we…

Adaptive Motorized LiDAR Scanning Control for Robust Localization with OpenStreetMap
Jianping Li, Kaisong Zhu, Zhongyuan Liu, Rui Jin, Xinhang Xu, Pengfei Wan, Lihua Xie
https://arxiv.org/abs/2509.11742

Adaptive Motorized LiDAR Scanning Control for Robust Localization with OpenStreetMap
LiDAR-to-OpenStreetMap (OSM) localization has gained increasing attention, as OSM provides lightweight global priors such as building footprints. These priors enhance global consistency for robot navigation, but OSM is often incomplete or outdated, limiting its reliability in real-world deployment. Meanwhile, LiDAR itself suffers from a limited field of view (FoV), where motorized rotation is commonly used to achieve panoramic coverage. Existing motorized LiDAR systems, however, typically emplo…

Unbounded: Object-Boundary Interactions in Mixed Reality
Zhuoyue Lyu, Per Ola Kristensson
https://arxiv.org/abs/2509.10750 https://arxiv.org/pdf/2509.10750…

Unbounded: Object-Boundary Interactions in Mixed Reality
Boundaries such as walls, windows, and doors are ubiquitous in the physical world, yet their potential in Mixed Reality (MR) remains underexplored. We present Unbounded, a Research through Design inquiry into Object-Boundary Interactions (OBIs). Building on prior work, we articulate a design space aimed at providing a shared language for OBIs. To demonstrate its potential, we design and implement eight examples across productivity and art exploration scenarios, showcasing how boundaries can enr…

😶‍🌫️ Generating real-time detailed ground visualisations from sparse aerial point clouds
#3d #xr

Generating real-time detailed ground visualisations from sparse aerial point clouds
Building realistic wide scale outdoor 3D content with sufficient visual quality to observe at walking eye level or from driven vehicles is often carried out by large teams of artists skilled in modelling, texturing, material shading and lighting, which typically leads to both prohibitive costs and reduced accuracy honoring the variety of real world ground truth landscapes. In our proposed method, we define a process to automatically amplify real-world scanned data and render real-time in animat…

Architect Terry Farrell, best-known for the postmodern MI6, died recently. My favorite building is his KK100 skyscraper, which is the world's tallest building by a British #architect.
[Photo by Carsten Schael]

#WritersCoffeeClub 10 Aug
How much of your world-building extends beyond what's shown in the text?
I don't know much about the cities on the northern coast of the continent, beyond what they import or export. I don't know how many there are, or what their political systems and religions are. I suspect that there are in fact other continents, but the people on '…

A Type 2 Fuzzy Set Approach for Building Linear Linguistic Regression Analysis under Multi Uncertainty
Junzo Watada, Pei-Chun Lin, Bo Wang, Jeng-Shyang Pan, Jose Guadalupe Flores Muniz
https://arxiv.org/abs/2509.10498

A Type 2 Fuzzy Set Approach for Building Linear Linguistic Regression Analysis under Multi Uncertainty
In this paper, we propose a novel heuristic algorithm for constructing a Type-2 Fuzzy Set of the Linear Linguistic Regression (T2F-LLR) model, designed to address uncertainty and vagueness in real-world decision-making. We consider a practical scenario involving a cosmetic company's promotional planning across four product categories: Basic Face Care, Face Cleaning, Cosmetics, and Body Care, aimed at both male and female consumers. Data are collected using fuzzy linguistic questionnaires from c…

Texas families offered just $3,000 by the government in exchange for building Trump’s wall on their land | The Independent
https://www.independent.co.uk/news/world/americas/us-politics/trump-border-wall-texas-b2802411.html

DyMoDreamer: World Modeling with Dynamic Modulation
Boxuan Zhang, Runqing Wang, Wei Xiao, Weipu Zhang, Jian Sun, Gao Huang, Jie Chen, Gang Wang
https://arxiv.org/abs/2509.24804 …

DyMoDreamer: World Modeling with Dynamic Modulation
A critical bottleneck in deep reinforcement learning (DRL) is sample inefficiency, as training high-performance agents often demands extensive environmental interactions. Model-based reinforcement learning (MBRL) mitigates this by building world models that simulate environmental dynamics and generate synthetic experience, improving sample efficiency. However, conventional world models process observations holistically, failing to decouple dynamic objects and temporal features from static backg…

Summary on The Multilingual Conversational Speech Language Model Challenge: Datasets, Tasks, Baselines, and Methods
Bingshen Mu, Pengcheng Guo, Zhaokai Sun, Shuai Wang, Hexin Liu, Mingchen Shao, Lei Xie, Eng Siong Chng, Longshuai Xiao, Qiangze Feng, Daliang Wang
https://arxiv.org/abs/2509.13785

Summary on The Multilingual Conversational Speech Language Model Challenge: Datasets, Tasks, Baselines, and Methods
This paper summarizes the Interspeech2025 Multilingual Conversational Speech Language Model (MLC-SLM) challenge, which aims to advance the exploration of building effective multilingual conversational speech LLMs (SLLMs). We provide a detailed description of the task settings for the MLC-SLM challenge, the released real-world multilingual conversational speech dataset totaling approximately 1,604 hours, and the baseline systems for participants. The MLC-SLM challenge attracts 78 teams from 13 c…

How to Rise and Keep Rising
- the fun and folk keeps building towards my upcoming album - thanks for listening/watching -
#MusicVideo #PlayList

"Canadian cities can prepare for climate change by building with nature"
#Canada #Climate #ClimateChange #Nature

Canadian cities can prepare for climate change by building with nature
The housing affordability crisis is top of mind for many around the world, including Canadians. Between 2019 and 2024, house prices in Toronto and Montréal had an average annual increase of 6.7% and 10.2%, respectively.

Gestern bin ich auf "The Disintegration Loops" von William Basinski gestoßen.
https://www.youtube.com/watch?v=mjnAE5go9dI
Heute ist der 11. September, und zieht Euch bitte den Wikipedia-Artikel zu den Disintegration Loops rein (https://en.wikipedia.org/wiki/The_Disintegration_Loops). Wahnsinn!
Auszug:
Basinski finished the project the morning of the September 11, 2001 attacks in New York City, and sat on the roof of his apartment building in Brooklyn with friends as the World Trade Center collapsed. He filmed the fallout during the last hour of daylight from a roof, and the following morning he played "Disintegration Loop 1.1" as a soundtrack to the aftermath. Stills from the video were used as the covers for the set of four CDs, and several weeks later Basinski dedicated the work to the victims in a postscript in the liner notes. He said that "the events gave new meaning to the musical pieces created by catastrophic decay in my studio a few weeks before."

Building organisational resilience through ISO 22301
"Today’s world is unpredictable. From cyber threats and pandemics to natural disasters, organizations must be ready to keep running, no matter what happens. A structured Business Continuity Management System – BCMS is no longer a luxury but a necessity. (...)
Understanding ISO 22301: A Framework for Resilience"

Lazarus was a fun show but it's world building was pretty weird. too many people like "world will end in 3 days, any back to my long term planning meeting"

Building causation links in stochastic nonlinear systems from data
Sergio Chibbaro, Cyril Furtlehner, Th\'eo Marchetta, Andrei-Tiberiu Pantea, Davide Rossetti
https://arxiv.org/abs/2509.07701

Building causation links in stochastic nonlinear systems from data
Causal relationships play a fundamental role in understanding the world around us. The ability to identify and understand cause-effect relationships is critical to making informed decisions, predicting outcomes, and developing effective strategies. However, deciphering causal relationships from observational data is a difficult task, as correlations alone may not provide definitive evidence of causality. In recent years, the field of machine learning (ML) has emerged as a powerful tool, offerin…

Introducing Claude Sonnet 4.5
#LLM

AI-Assisted Geometric Analysis of Cultured Neuronal Networks: Parallels with the Cosmic Web
Wolfgang Kurz, Danny Baranes
https://arxiv.org/abs/2510.10286 https://

AI-Assisted Geometric Analysis of Cultured Neuronal Networks: Parallels with the Cosmic Web
Building on evidence of structural parallels between brain networks and the cosmic web [1], we apply AI-based geometric analysis to cultured neuronal networks. Isolated neurons self-organize into dendritic lattices shaped by reproducible wiring rules. These lattices show non-random features-frequent dendritic convergence, hub nodes, small-world connectivity, and large voids. Synaptic contacts cluster and strengthen at hubs. Strikingly, these properties mirror the cosmic web: dendritic branches …

Empirical Study on Robustness and Resilience in Cooperative Multi-Agent Reinforcement Learning
Simin Li, Zihao Mao, Hanxiao Li, Zonglei Jing, Zhuohang bian, Jun Guo, Li Wang, Zhuoran Han, Ruixiao Xu, Xin Yu, Chengdong Ma, Yuqing Ma, Bo An, Yaodong Yang, Weifeng Lv, Xianglong Liu
https://arxiv.org/abs/2510.11824

Empirical Study on Robustness and Resilience in Cooperative Multi-Agent Reinforcement Learning
In cooperative Multi-Agent Reinforcement Learning (MARL), it is a common practice to tune hyperparameters in ideal simulated environments to maximize cooperative performance. However, policies tuned for cooperation often fail to maintain robustness and resilience under real-world uncertainties. Building trustworthy MARL systems requires a deep understanding of robustness, which ensures stability under uncertainties, and resilience, the ability to recover from disruptions--a concept extensively …

Sources: xAI is building world models for use in gaming and robotics, and has hired two AI researchers, Zeeshan Patel and Ethan He, from Nvidia to work on them (Cristina Criddle/Financial Times)
https://www.ft.com/content/ac566346-53dd-4490-8d4c-5269906c64ee

Just finished "Beasts Made of Night" by Tochi Onyebuchi...
Indirect CW for fantasy police state violence.
So I very much enjoyed Onyebuchi's "Riot Baby," and when I grabbed this at the library, I was certain it would be excellent. But having finished it, I'm not sure I like it that much overall?
The first maybe third is excellent, including the world-building, which is fascinating. I feel like Onyebuchi must have played "Shadow of the Colossus" at some point. Onyebuchi certainly does know how to make me care for his characters.
Some spoilers from here on out...
.
.
.
I felt like it stumbles towards the middle, with Bo's reactions neither making sense in the immediate context, nor in retrospect by the end when we've learned more. Things are a bit floaty in the middle with an unclear picture of what exactly is going on politics-wise and what the motivations are. Here I think there were some nuances that didn't make it to the page, or perhaps I'm just a bit thick and not getting stuff I should be? More is of course revealed by the end, but I still wasn't satisfied with the explanations of things. For example, (spoilers) I don't feel I understand clearly what kind of power the army of aki was supposed to represent within the city? Perhaps necessary to wield the threat of offensive inisisia use? In that case, a single scene somewhere of Izu's faction deploying that tactic would have been helpful I think.
Then towards the end, for me things really started to jumble, with unclear motivations, revelations that didn't feel well-paced or -structured, and a finale where both the action & collapsing concerns felt stilted and disjointed. Particularly the mechanics/ethics of the most important death that set the finale in motion bothered me, and the unexplained mechanism by which that led to what came next? I can read a couple of possible interesting morals into the whole denouement, but didn't feel that any of them were sufficiently explored. Especially if we're supposed to see some personal failing in the protagonist's actions, I don't think it's made clear enough what that is, since I feel his reasons to reject each faction are pretty solid, and if we're meant to either pity or abjure his indecision, I don't think the message lands clearly enough.
There *is* a sequel, which honestly I wasn't sure of after the last page, and which I now very interested in. Beasts is Onyebuchi's debut, which maybe makes sense of me feeling that Riot Baby didn't have the same plotting issues. It also maybe means that Onyebuchi couldn't be sure a sequel would make it to publication in terms of setting up the ending.
Overall I really enjoyed at least 80% of this, but was expecting even better (especially politically) given Onyebuchi's other work, and I didn't feel like I found it.
#AmReading

It's interesting to see Saudi Arabia acquiring Western cultural touchstones. I also saw they're building a "Casino City" in Saudi Arabia. I am so confused by the world right now. https://flipboard.com/@theverge/gaming-4oihgk74z/-/a-…

The Solar System School is now permanent! 🥳🥳
The unique graduate program offers students from all over the world the opportunity to obtain a doctorate in solar system research. 🛰️☀️🔭
Joint PR @…

Author Michelle Tea on making art your main focus (and not taking your day job too seriously) https://thecreativ…

Author Michelle Tea on making art your main focus (and not taking your day job too seriously) – The Creative Independent
Author and publisher Michelle Tea discusses surviving as an artist, building community, and staying true to your voice

A methodology for clinically driven interactive segmentation evaluation
Parhom Esmaeili, Virginia Fernandez, Pedro Borges, Eli Gibson, Sebastien Ourselin, M. Jorge Cardoso
https://arxiv.org/abs/2510.09499

A methodology for clinically driven interactive segmentation evaluation
Interactive segmentation is a promising strategy for building robust, generalisable algorithms for volumetric medical image segmentation. However, inconsistent and clinically unrealistic evaluation hinders fair comparison and misrepresents real-world performance. We propose a clinically grounded methodology for defining evaluation tasks and metrics, and built a software framework for constructing standardised evaluation pipelines. We evaluate state-of-the-art algorithms across heterogeneous and…

How Far Do Time Series Foundation Models Paint the Landscape of Real-World Benchmarks ?
Lujun Li, Lama Sleem, Yiqun Wang, Yangjie Xu, Niccol\`o Gentile, Radu State
https://arxiv.org/abs/2509.26347

How Far Do Time Series Foundation Models Paint the Landscape of Real-World Benchmarks ?
Recent evaluations of time-series foundation models (TSFMs) have emphasized synthetic benchmarks, leaving real-world generalization less thoroughly examined. This work proposes a novel benchmarking approach that bridges synthetic and realistic data by extracting temporal signals from real-world video using optical flow and curating datasets reflecting everyday temporal dynamics. Building upon this pipeline, we introduce REAL-V-TSFM, a novel dataset designed to capture rich and diverse time seri…

#PreciousPlastic is a global community of people working towards a solution to plastic pollution. Knowledge, tools and techniques are shared online, for free
LET'S BUILD V.5:
==
We spent the last 10 years building a network of plastic recyclers around the world. Yet despite progress, only 10% of all plastic ever made has been recycled, the same as when we started

Weird Greco III 🇬🇷
怪异希腊 III 🇬🇷
📷 Nikon FE
🎞️Ilford FP4 Plus, expired 1994
buy me ☕️ ?/请我喝杯☕️？
#filmphotography

Navigating the Data Space Landscape: Concepts, Applications, and Future Directions
Bojana Marojevikj, Riste Stojanov
https://arxiv.org/abs/2509.06983 https://

Navigating the Data Space Landscape: Concepts, Applications, and Future Directions
This paper explores the evolving landscape of data spaces, focusing on key concepts, practical applications, and emerging future directions. It begins by introducing the foundational principles that underpin data space architectures, emphasizing their importance in facilitating secure and efficient data exchange. The core design principles and essential building blocks that form the backbone of data-space systems are then examined. Several real-world implementations are presented, showcasing ho…

China unveils world's first zero-carbon tower with innovative green technology. #climatechange #climatesolutions #climate

China unveils world's first zero-carbon tower with 24,000 sensors
Discover how the zero-carbon building in Qingdao utilizes photovoltaic glass to significantly reduce carbon emissions and energy loss.

🎯 Real-world validation through extended #CCBench testing with human evaluators completing multi-turn tasks in isolated #Docker containers across frontend development, tool building, data analysis, testing & algorithms
🔧 Near parity with

One of our most important projects was the "food security committee" which made sure that our members always ate. It expanded out and now members of that committee still can food and bring it to houseless camps (among other things).
This is disaster prep work. We are in a disaster, how do you prepare for it to get worse? How to you prepare to come out the other side?
The only way we survive this level of disaster is together. Organize around that idea and you'll already be building the world we want to see on the other side.

Visual artist David Hendren on building community as an introvert – #creativity and #community http…

Visual artist David Hendren on building community as an introvert – The Creative Independent
Visual artist David Hendren discusses dealing with rejection, sustaining an art practice, and working with intention

Tonight, we're throwing a Retro Pixel Party! My Linux PC is geared up to stream The Settlers (1993). Get ready to witness some vintage city-building, pixel by glorious pixel. Will my settlers ever find their way? Tune in to find out! https://www.twitch.tv/tuxramus

At Berlin Buzzwords 2025, Radu POP and Pietro Mele discussed building an extensible hybrid search solution with Elasticsearch. They covered functional modeling, cluster architecture, and practical insights on managing billions of vectors in real-world scenarios. They also addressed hybrid reranking challenges and the limitations of standard fusion techniques, explaining their innovative approach.
Watch the full session: https://youtu.be/C73Pk6PUTl4?si=CTtmjqPdaxp6PeFM

CrediBench: Building Web-Scale Network Datasets for Information Integrity
Emma Kondrup, Sebastian Sabry, Hussein Abdallah, Zachary Yang, James Zhou, Kellin Pelrine, Jean-Fran\c{c}ois Godbout, Michael M. Bronstein, Reihaneh Rabbany, Shenyang Huang
https://arxiv.org/abs/2509.23340

CrediBench: Building Web-Scale Network Datasets for Information Integrity
Online misinformation poses an escalating threat, amplified by the Internet's open nature and increasingly capable LLMs that generate persuasive yet deceptive content. Existing misinformation detection methods typically focus on either textual content or network structure in isolation, failing to leverage the rich, dynamic interplay between website content and hyperlink relationships that characterizes real-world misinformation ecosystems. We introduce CrediBench: a large-scale data processing …

Just finished "Decelerate Blue" written by Adam Rapp and illustrated by Mike Cavallaro. It's a dystopian graphic novel that I found... not that great. Maybe the best thing about it was the world-building around the propaganda language everyone is forced to speak, but overall I found the plot construction and characters to be underwhelming.
Despite being a book about resistance to fascist oppression, it doesn't meaningfully engage with any major axes of oppression like patriarchy, racism, capitalism, or colonialism, and it doesn't offer an interesting lessons on how to conduct resistance or what long-term outcomes one might hope for.
First graphic novel I've checked out in a while that I didn't really like that much.
#AmReading

Anthropic launches Claude Sonnet 4.5, claims it is the world's best coding model, strongest for building complex agents, and its "most aligned" frontier model (Anthropic)
https://www.anthropic.com/news/claude-sonnet-4-5

Introducing Claude Sonnet 4.5
Claude Sonnet 4.5 is the best coding model in the world, strongest model for building complex agents, and best model at using computers.

ChineseEEG-2: An EEG Dataset for Multimodal Semantic Alignment and Neural Decoding during Reading and Listening
Sitong Chen, Beiqianyi Li, Cuilin He, Dongyang Li, Mingyang Wu, Xinke Shen, Song Wang, Xuetao Wei, Xindi Wang, Haiyan Wu, Quanying Liu
https://arxiv.org/abs/2508.04240

ChineseEEG-2: An EEG Dataset for Multimodal Semantic Alignment and Neural Decoding during Reading and Listening
EEG-based neural decoding requires large-scale benchmark datasets. Paired brain-language data across speaking, listening, and reading modalities are essential for aligning neural activity with the semantic representation of large language models (LLMs). However, such datasets are rare, especially for non-English languages. Here, we present ChineseEEG-2, a high-density EEG dataset designed for benchmarking neural decoding models under real-world language tasks. Building on our previous ChineseEE…

🏗️ Tear it down they said... but he kept on building
(looks like China has an answer to the Watts Towers!)
#buildings #housing #china

Tear It Down, They Said. He Just Kept Building.
Defying demolition orders, a Chinese man turned his home into a rickety 11-story tower. Now tourists are coming.

Computational Concept of the Psyche
Anton Kolonin, Vladimir Kryukov
https://arxiv.org/abs/2509.07009 https://arxiv.org/pdf/2509.07009

Computational Concept of the Psyche
The article provides an overview of approaches to modeling the human psyche in the perspective of building an artificial one. Based on the review, a concept of cognitive architecture is proposed, where the psyche is considered as an operating system of a living or artificial subject, including a space of needs that determines its life meanings in connection with stimuli from the external world, and intelligence as a decision-making system for actions in relation to this world in order to satisf…

Singer-songwriter Nemahsis on the importance of being uncompromising – The Creative Independent https://thecreativeindependent.com/peo

Singer-songwriter Nemahsis on the importance of being uncompromising – The Creative Independent
Singer-songwriter Nemahsis discusses breaking through despite hardships, freedom of expression, and the power of building a genuine community online

No paywall. One of the folks I read when I can to stay sane. Always on the cutting edge of The Moment.
Welcome to The Continental. We Do Hope You Enjoy Your Stay. - Epsilon Theory https://new.epsilontheory.com/a/welcome-to-the-continental-we-do-hope…

Welcome to The Continental. We Do Hope You Enjoy Your Stay.
Our autonomy of mind is our birthright. It is not ladled out to us from some central pot, and it cannot be taken away from us. But we can give it away, sometimes in pieces and sometimes all at once, and that’s what the semantic system of world-building and script-injection encourages us to do. Enough. Today we begin to reclaim our minds.

Just finished "Donuts and Doom" by Balazs Lorinczi. A lovely queer romance graphic novel featuring a witch and a rock star.
I like the way the book builds their romance through different stages, even if the plot is fairly basic. The world building is also fun.
#AmReading

LLMs in the SOC: An Empirical Study of Human-AI Collaboration in Security Operations Centres
Ronal Singh, Shahroz Tariq, Fatemeh Jalalvand, Mohan Baruwal Chhetri, Surya Nepal, Cecile Paris, Martin Lochner
https://arxiv.org/abs/2508.18947

LLMs in the SOC: An Empirical Study of Human-AI Collaboration in Security Operations Centres
The integration of Large Language Models (LLMs) into Security Operations Centres (SOCs) presents a transformative, yet still evolving, opportunity to reduce analyst workload through human-AI collaboration. However, their real-world application in SOCs remains underexplored. To address this gap, we present a longitudinal study of 3,090 analyst queries from 45 SOC analysts over 10 months. Our analysis reveals that analysts use LLMs as on-demand aids for sensemaking and context-building, rather th…

Generating real-time detailed ground visualisations from sparse aerial point clouds
Aidan Murray, Eddie Waite, Caleb Ross, Scarlet Mitchell, Alexander Bradley, Joanna Jamrozy, Kenny Mitchell
https://arxiv.org/abs/2507.18664

Generating real-time detailed ground visualisations from sparse aerial point clouds
Building realistic wide scale outdoor 3D content with sufficient visual quality to observe at walking eye level or from driven vehicles is often carried out by large teams of artists skilled in modelling, texturing, material shading and lighting, which typically leads to both prohibitive costs and reduced accuracy honoring the variety of real world ground truth landscapes. In our proposed method, we define a process to automatically amplify real-world scanned data and render real-time in animat…

Human-like Navigation in a World Built for Humans
Bhargav Chandaka, Gloria X. Wang, Haozhe Chen, Henry Che, Albert J. Zhai, Shenlong Wang
https://arxiv.org/abs/2509.21189 https:…

Human-like Navigation in a World Built for Humans
When navigating in a man-made environment they haven't visited before--like an office building--humans employ behaviors such as reading signs and asking others for directions. These behaviors help humans reach their destinations efficiently by reducing the need to search through large areas. Existing robot navigation systems lack the ability to execute such behaviors and are thus highly inefficient at navigating within large environments. We present ReasonNav, a modular navigation system which …

Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmark
Yuxuan Cai, Yipeng Hao, Jie Zhou, Hang Yan, Zhikai Lei, Rui Zhen, Zhenhua Han, Yutao Yang, Junsong Li, Qianjun Pan, Tianyu Huai, Qin Chen, Xin Li, Kai Chen, Bo Zhang, Xipeng Qiu, Liang He
https://arxiv.org/abs/2508.19005

Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmark
As AI advances toward general intelligence, the focus is shifting from systems optimized for static tasks to creating open-ended agents that learn continuously. In this paper, we introduce Experience-driven Lifelong Learning (ELL), a framework for building self-evolving agents capable of continuous growth through real-world interaction. The framework is built on four core principles: (1) Experience Exploration: Agents learn through continuous, self-motivated interaction with dynamic environment…

MLLM-based Speech Recognition: When and How is Multimodality Beneficial?
Yiwen Guan, Viet Anh Trinh, Vivek Voleti, Jacob Whitehill
https://arxiv.org/abs/2507.19037 https://

MLLM-based Speech Recognition: When and How is Multimodality Beneficial?
Recent advances in multi-modal large language models (MLLMs) have opened new possibilities for unified modeling of speech, text, images, and other modalities. Building on our prior work, this paper examines the conditions and model architectures under which multiple input modalities can improve automatic speech recognition (ASR) accuracy in noisy environments. Through experiments on synthetic and real-world data, we find that (1) harnessing more modalities usually improves ASR accuracy, as each…

graph framework: A Domain Specific Compiler for Building Physics Applications
M. Cianciosa, D. Batchelor, W. Elwasif
https://arxiv.org/abs/2508.15967 https://

graph framework: A Domain Specific Compiler for Building Physics Applications
Modern supercomputers are increasingly relying on Graphic Processing Units (GPUs) and other accelerators to achieve exa-scale performance at reasonable energy usage. The challenge of exploiting these accelerators is the incompatibility between different vendors. A scientific code written using CUDA will not operate on a AMD gpu. Frameworks that can abstract the physics from the accelerator kernel code are needed to exploit the current and future hardware. In the world of machine learning, sever…

Learning Linear Regression with Low-Rank Tasks in-Context
Kaito Takanami, Takashi Takahashi, Yoshiyuki Kabashima
https://arxiv.org/abs/2510.04548 https://a…

Learning Linear Regression with Low-Rank Tasks in-Context
In-context learning (ICL) is a key building block of modern large language models, yet its theoretical mechanisms remain poorly understood. It is particularly mysterious how ICL operates in real-world applications where tasks have a common structure. In this work, we address this problem by analyzing a linear attention model trained on low-rank regression tasks. Within this setting, we precisely characterize the distribution of predictions and the generalization error in the high-dimensional li…

TL;DR: what if nationalism, not anarchy, is futile?
Since I had the pleasure of seeing the "what would anarchists do against a warlord?" argument again in my timeline, I'll present again my extremely simple proposed solution:
Convince the followers of the warlord that they're better off joining you in freedom, then kill or exile the warlord once they're alone or vastly outnumbered.
Remember that even in our own historical moment where nothing close to large-scale free society has existed in living memory, the warlord's promise of "help me oppress others and you'll be richly rewarded" is a lie that many understand is historically a bad bet. Many, many people currently take that bet, for a variety of reasons, and they're enough to coerce through fear an even larger number of others. But although we imagine, just as the medieval peasants might have imagined of monarchy, that such a structure is both the natural order of things and much too strong to possibly fail, in reality it takes an enormous amount of energy, coordination, and luck for these structures to persist! Nations crumble every day, and none has survived more than a couple *hundred* years, compared to pre-nation societies which persisted for *tends of thousands of years* if not more. I'm this bubbling froth of hierarchies, the notion that hierarchy is inevitable is certainly popular, but since there's clearly a bit of an ulterior motive to make (and teach) that claim, I'm not sure we should trust it.
So what I believe could form the preconditions for future anarchist societies to avoid the "warlord problem" is merely: a widespread common sense belief that letting anyone else have authority over you is morally suspect. Given such a belief, a warlord will have a hard time building any following at all, and their opponents will have an easy time getting their supporters to defect. In fact, we're already partway there, relative to the situation a couple hundred years ago. At that time, someone could claim "you need to obey my orders and fight and die for me because the Queen was my mother" and that was actually a quite successful strategy. Nowadays, this strategy is only still working in a few isolated places, and the idea that one could *start a new monarchy* or even resurrect a defunct one seems absurd. So why can't that same transformation from "this is just how the world works" to "haha, how did anyone ever believe *that*? also happen to nationalism in general? I don't see an obvious reason why not.
Now I think one popular counterargument to this is: if you think non-state societies can win out with these tactics, why didn't they work for American tribes in the face of the European colonizers? (Or insert your favorite example of colonialism here.) I think I can imagine a variety of reasons, from the fact that many of those societies didn't try this tactic (and/or were hierarchical themselves), to the impacts of disease weakening those societies pre-contact, to the fact that with much-greater communication and education possibilities it might work better now, to the fact that most of those tribes are *still* around, and a future in which they persist longer than the colonist ideologies actually seems likely to me, despite the fact that so much cultural destruction has taken place. In fact, if the modern day descendants of the colonized tribes sow the seeds of a future society free of colonialism, that's the ultimate demonstration of the futility of hierarchical domination (I just read "Theory of Water" by Leanne Betasamosake Simpson).
I guess the TL;DR on this is: what if nationalism is actually as futile as monarchy, and we're just unfortunately living in the brief period during which it is ascendant?

Toward Engineering AGI: Benchmarking the Engineering Design Capabilities of LLMs
Xingang Guo, Yaxin Li, Xiangyi Kong, Yilan Jiang, Xiayu Zhao, Zhihua Gong, Yufan Zhang, Daixuan Li, Tianle Sang, Beixiao Zhu, Gregory Jun, Yingbing Huang, Yiqi Liu, Yuqi Xue, Rahul Dev Kundu, Qi Jian Lim, Yizhou Zhao, Luke Alexander Granger, Mohamed Badr Younis, Darioush Keivan, Nippun Sabharwal, Shreyanka Sinha, Prakhar Agarwal, Kojo Vandyck, Hanlin Mai, Zichen Wang, Aditya Venkatesh, Ayush Barik, Jiankun…

Toward Engineering AGI: Benchmarking the Engineering Design Capabilities of LLMs
Today, industry pioneers dream of developing general-purpose AI engineers capable of designing and building humanity's most ambitious projects--from starships that will carry us to distant worlds to Dyson spheres that harness stellar energy. Yet engineering design represents a fundamentally different challenge for large language models (LLMs) compared to traditional textbook-style problem solving or factual question answering. Real-world engineering design demands the synthesis of domain knowle…

OpenZL: A Graph-Based Model for Compression
Yann Collet, Nick Terrell, W. Felix Handte, Danielle Rozenblit, Victor Zhang, Kevin Zhang, Yaelle Goldschlag, Jennifer Lee, Daniel Riegel, Stan Angelov, Nadav Rotem
https://arxiv.org/abs/2510.03203

OpenZL: A Graph-Based Model for Compression
Research in general-purpose lossless compression over the last decade has largely found improvements in compression ratio that come at great cost to resource utilization and processing throughput. However, most production workloads require high throughput and low resource utilization, so most research systems have seen little adoption. Instead, real world improvements in compression are increasingly often realized by building application-specific compressors which can exploit knowledge about th…

ReLAM: Learning Anticipation Model for Rewarding Visual Robotic Manipulation
Nan Tang, Jing-Cheng Pang, Guanlin Li, Chao Qian, Yang Yu
https://arxiv.org/abs/2509.22402 https://

ReLAM: Learning Anticipation Model for Rewarding Visual Robotic Manipulation
Reward design remains a critical bottleneck in visual reinforcement learning (RL) for robotic manipulation. In simulated environments, rewards are conventionally designed based on the distance to a target position. However, such precise positional information is often unavailable in real-world visual settings due to sensory and perceptual limitations. In this study, we propose a method that implicitly infers spatial distances through keypoints extracted from images. Building on this, we introdu…

Subzero Labs, which is building Rialo, a blockchain designed for non-crypto developers to build "real-world" apps, raised a $20M seed led by Pantera Capital (Ben Weiss/Fortune)
https://fortune.com/crypto/2025/08/01/subzero-labs-seed-r…

Subzero Labs raises $20 million to build a blockchain for the ‘real world’
Pantera Capital led the fundraise, with participation from Variant, Coinbase Ventures, and Susquehanna Crypto.

China tests world-class robot arms for 'artificial sun' project, bringing them closer to their goal of building a fusion reactor maintenance system with potential applications in nuclear plant inspections and rescue missions. #climatesolutions

China tests world-class robot arms for ‘artificial sun’ project
China unveils robotic arm trio for fusion reactor maintenance, moving closer to stable clean energy and ITER support.

Today on 3D Printing Adventures Penguin Style!: We're building an AT-AT - Star Wars Plant Pot! Because nothing says "gardening" like a giant, four-legged war machine. May the force (and the filament) be with us! https://www.twitch.tv/tuxramus

AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications
Dawei Gao, Zitao Li, Yuexiang Xie, Weirui Kuang, Liuyi Yao, Bingchen Qian, Zhijian Ma, Yue Cui, Haohao Luo, Shen Li, Lu Yi, Yi Yu, Shiqi He, Zhiling Luo, Wenmeng Zhou, Zhicheng Zhang, Xuguang He, Ziqian Chen, Weikai Liao, Farruh Isakulovich Kushnazarov, Yaliang Li, Bolin Ding, Jingren Zhou

AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications
Driven by rapid advancements of Large Language Models (LLMs), agents are empowered to combine intrinsic knowledge with dynamic tool use, greatly enhancing their capacity to address real-world tasks. In line with such an evolution, AgentScope introduces major improvements in a new version (1.0), towards comprehensively supporting flexible and efficient tool-based agent-environment interactions for building agentic applications. Specifically, we abstract foundational components essential for agen…

Budget Home Recording: Essential Tools to get you Started https://www.gearnews.com/budget-home-recording-studio/?utm_source=dlvr.it&utm_content=djrandomjavabas/magazine/Digital …

Budget Home Recording: Essential Tools to get you Started
With so much gear available, building your own budget home recording setup can be tricky for beginners. We discuss the essentials!

AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications
Dawei Gao, Zitao Li, Yuexiang Xie, Weirui Kuang, Liuyi Yao, Bingchen Qian, Zhijian Ma, Yue Cui, Haohao Luo, Shen Li, Lu Yi, Yi Yu, Shiqi He, Zhiling Luo, Wenmeng Zhou, Zhicheng Zhang, Xuguang He, Ziqian Chen, Weikai Liao, Farruh Isakulovich Kushnazarov, Yaliang Li, Bolin Ding, Jingren Zhou

AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications
Driven by rapid advancements of Large Language Models (LLMs), agents are empowered to combine intrinsic knowledge with dynamic tool use, greatly enhancing their capacity to address real-world tasks. In line with such an evolution, AgentScope introduces major improvements in a new version (1.0), towards comprehensively supporting flexible and efficient tool-based agent-environment interactions for building agentic applications. Specifically, we abstract foundational components essential for agen…

⚡ Living cement: Scientists turn bacteria-infused cement into energy-storing supercapacitors
https://techxplore.com/news/2025-09-cement-scientists-bacteria-infused-energy.html

Living cement: Scientists turn bacteria-infused cement into energy-storing supercapacitors
Imagine a house that doesn't just shelter you but also stores electricity. It may sound like science fiction, but it's now closer to reality than ever before. A research team at Aarhus University has demonstrated how the world's most widely used building material can be transformed into a living energy device. By embedding energy-producing bacteria in cement, they have created a biohybrid supercapacitor with surprisingly high performance and a remarkable ability to regenerate itself over time.

ExCyTIn-Bench: Evaluating LLM agents on Cyber Threat Investigation
Yiran Wu, Mauricio Velazco, Andrew Zhao, Manuel Ra\'ul Mel\'endez Luj\'an, Srisuma Movva, Yogesh K Roy, Quang Nguyen, Roberto Rodriguez, Qingyun Wu, Michael Albada, Julia Kiseleva, Anand Mudgerikar
https://arxiv.org/abs/2507.14201

ExCyTIn-Bench: Evaluating LLM agents on Cyber Threat Investigation
We present ExCyTIn-Bench, the first benchmark to Evaluate an LLM agent x on the task of Cyber Threat Investigation through security questions derived from investigation graphs. Real-world security analysts must sift through a large number of heterogeneous alert signals and security logs, follow multi-hop chains of evidence, and compile an incident report. With the developments of LLMs, building LLM-based agents for automatic thread investigation is a promising direction. To assist the developme…

Anthropic prices Claude Sonnet 4.5 at $3/1M input and $15/1M output tokens, same as Sonnet 4, cheaper than Opus at $15/$75 but higher than GPT-5 at $1.25/$10 (Simon Willison/Simon Willison's Weblog)
https://simonwillison.net/2025/Sep/29/claude-sonnet-4-5/

Claude Sonnet 4.5 is probably the “best coding model in the world” (at least for now)
Anthropic released Claude Sonnet 4.5 today, with a very bold set of claims: Claude Sonnet 4.5 is the best coding model in the world. It’s the strongest model for building …

Physics-Informed Regression Modelling for Vertical Facade Surface Temperature: A Tropical Case Study on Solar-reflective Material
Shisheng Chen, Shanshan Tong, Nyuk Hien Wong, May Lwin Oo, Joie Lim, Erna Tan, Ruohan Xu, Marcel Ignatius, Yang He, Zhenjiang Shen
https://arxiv.org/abs/2507.16174

Physics-Informed Regression Modelling for Vertical Facade Surface Temperature: A Tropical Case Study on Solar-reflective Material
Urban heat islands (UHIs) pose a critical challenge in densely populated cities and tropical climates where large amounts of energy are used to meet the cooling demand. To address this, Building and Construction Authority (BCA) of Singapore provides incentives for passive cooling such as using of solar-reflective material in its Green Mark guidelines. Thus, understanding about its real-world effectiveness in tropical urban environments is required. This study evaluated the effectiveness of sola…

The Eloquence team submission for task 1 of MLC-SLM challenge
Lorenzo Concina, Jordi Luque, Alessio Brutti, Marco Matassoni, Yuchen Zhang
https://arxiv.org/abs/2507.19308 https:…

The Eloquence team submission for task 1 of MLC-SLM challenge
In this paper, we present our studies and experiments carried out for the task 1 of the Challenge and Workshop on Multilingual Conversational Speech Language Model (MLC-SLM), which focuses on advancing multilingual conversational speech recognition through the development of speech language models architectures. Given the increasing relevance of real-world conversational data for building robust Spoken Dialogue Systems, we explore three approaches to multilingual ASR. First, we conduct an evalu…

PoseDiff: A Unified Diffusion Model Bridging Robot Pose Estimation and Video-to-Action Control
Haozhuo Zhang, Michele Caprio, Jing Shao, Qiang Zhang, Jian Tang, Shanghang Zhang, Wei Pan
https://arxiv.org/abs/2509.24591

PoseDiff: A Unified Diffusion Model Bridging Robot Pose Estimation and Video-to-Action Control
We present PoseDiff, a conditional diffusion model that unifies robot state estimation and control within a single framework. At its core, PoseDiff maps raw visual observations into structured robot states-such as 3D keypoints or joint angles-from a single RGB image, eliminating the need for multi-stage pipelines or auxiliary modalities. Building upon this foundation, PoseDiff extends naturally to video-to-action inverse dynamics: by conditioning on sparse video keyframes generated by world mod…

Decentralized Federated Learning of Probabilistic Generative Classifiers
Aritz P\'erez, Carlos Echegoyen, Guzm\'an Santaf\'e
https://arxiv.org/abs/2507.17285

Decentralized Federated Learning of Probabilistic Generative Classifiers
Federated learning is a paradigm of increasing relevance in real world applications, aimed at building a global model across a network of heterogeneous users without requiring the sharing of private data. We focus on model learning over decentralized architectures, where users collaborate directly to update the global model without relying on a central server. In this context, the current paper proposes a novel approach to collaboratively learn probabilistic generative classifiers with a parame…

A profile of Egune AI, a startup building LLMs for the Mongolian language, as it navigates geopolitics, a lack of resources, and the nascent local tech scene (Viola Zhou/Rest of World)
https://restofworld.org/2025/mongolia-egune-ai-llm/

The Mongolian startup defying Big Tech with its own LLM
Egune is one of several linguistically and culturally aware AI models built by smaller nations to reduce reliance on American and Chinese tech giants.

Via @…
Challenge: "Name 20 female authors you admire, 1 per day"
N. K. Jemisin.
Incredibly powerful writing and writes science fiction and fantasy that I actually enjoy reading after getting disillusioned with the Tolkein lineage for its deep racism & colonialism. Her Stone Sky series is wildly creative in terms of world building, had an excellent epic plot to rival any of the other greats you could name, has complex and compelling characters, and a satisfying conclusion. She's won the right awards, and I hope that that translates into a lineage of people building in her ideas as rich as Tolkien's.

Should we teach vibe coding? Here's why not.
Should AI coding be taught in undergrad CS education?
1/2
I teach undergraduate computer science labs, including for intro and more-advanced core courses. I don't publish (non-negligible) scholarly work in the area, but I've got years of craft expertise in course design, and I do follow the academic literature to some degree. In other words, In not the world's leading expert, but I have spent a lot of time thinking about course design, and consider myself competent at it, with plenty of direct experience in what knowledge & skills I can expect from students as they move through the curriculum.
I'm also strongly against most uses of what's called "AI" these days (specifically, generative deep neutral networks as supplied by our current cadre of techbro). There are a surprising number of completely orthogonal reasons to oppose the use of these systems, and a very limited number of reasonable exceptions (overcoming accessibility barriers is an example). On the grounds of environmental and digital-commons-pollution costs alone, using specifically the largest/newest models is unethical in most cases.
But as any good teacher should, I constantly question these evaluations, because I worry about the impact on my students should I eschew teaching relevant tech for bad reasons (and even for his reasons). I also want to make my reasoning clear to students, who should absolutely question me on this. That inspired me to ask a simple question: ignoring for one moment the ethical objections (which we shouldn't, of course; they're very stark), at what level in the CS major could I expect to teach a course about programming with AI assistance, and expect students to succeed at a more technically demanding final project than a course at the same level where students were banned from using AI? In other words, at what level would I expect students to actually benefit from AI coding "assistance?"
To be clear, I'm assuming that students aren't using AI in other aspects of coursework: the topic of using AI to "help you study" is a separate one (TL;DR it's gross value is not negative, but it's mostly not worth the harm to your metacognitive abilities, which AI-induced changes to the digital commons are making more important than ever).
So what's my answer to this question?
If I'm being incredibly optimistic, senior year. Slightly less optimistic, second year of a masters program. Realistic? Maybe never.
The interesting bit for you-the-reader is: why is this my answer? (Especially given that students would probably self-report significant gains at lower levels.) To start with, [this paper where experienced developers thought that AI assistance sped up their work on real tasks when in fact it slowed it down] (https://arxiv.org/abs/2507.09089) is informative. There are a lot of differences in task between experienced devs solving real bugs and students working on a class project, but it's important to understand that we shouldn't have a baseline expectation that AI coding "assistants" will speed things up in the best of circumstances, and we shouldn't trust self-reports of productivity (or the AI hype machine in general).
Now we might imagine that coding assistants will be better at helping with a student project than at helping with fixing bugs in open-source software, since it's a much easier task. For many programming assignments that have a fixed answer, we know that many AI assistants can just spit out a solution based on prompting them with the problem description (there's another elephant in the room here to do with learning outcomes regardless of project success, but we'll ignore this over too, my focus here is on project complexity reach, not learning outcomes). My question is about more open-ended projects, not assignments with an expected answer. Here's a second study (by one of my colleagues) about novices using AI assistance for programming tasks. It showcases how difficult it is to use AI tools well, and some of these stumbling blocks that novices in particular face.
But what about intermediate students? Might there be some level where the AI is helpful because the task is still relatively simple and the students are good enough to handle it? The problem with this is that as task complexity increases, so does the likelihood of the AI generating (or copying) code that uses more complex constructs which a student doesn't understand. Let's say I have second year students writing interactive websites with JavaScript. Without a lot of care that those students don't know how to deploy, the AI is likely to suggest code that depends on several different frameworks, from React to JQuery, without actually setting up or including those frameworks, and of course three students would be way out of their depth trying to do that. This is a general problem: each programming class carefully limits the specific code frameworks and constructs it expects students to know based on the material it covers. There is no feasible way to limit an AI assistant to a fixed set of constructs or frameworks, using current designs. There are alternate designs where this would be possible (like AI search through adaptation from a controlled library of snippets) but those would be entirely different tools.
So what happens on a sizeable class project where the AI has dropped in buggy code, especially if it uses code constructs the students don't understand? Best case, they understand that they don't understand and re-prompt, or ask for help from an instructor or TA quickly who helps them get rid of the stuff they don't understand and re-prompt or manually add stuff they do. Average case: they waste several hours and/or sweep the bugs partly under the rug, resulting in a project with significant defects. Students in their second and even third years of a CS major still have a lot to learn about debugging, and usually have significant gaps in their knowledge of even their most comfortable programming language. I do think regardless of AI we as teachers need to get better at teaching debugging skills, but the knowledge gaps are inevitable because there's just too much to know. In Python, for example, the LLM is going to spit out yields, async functions, try/finally, maybe even something like a while/else, or with recent training data, the walrus operator. I can't expect even a fraction of 3rd year students who have worked with Python since their first year to know about all these things, and based on how students approach projects where they have studied all the relevant constructs but have forgotten some, I'm not optimistic seeing these things will magically become learning opportunities. Student projects are better off working with a limited subset of full programming languages that the students have actually learned, and using AI coding assistants as currently designed makes this impossible. Beyond that, even when the "assistant" just introduces bugs using syntax the students understand, even through their 4th year many students struggle to understand the operation of moderately complex code they've written themselves, let alone written by someone else. Having access to an AI that will confidently offer incorrect explanations for bugs will make this worse.
To be sure a small minority of students will be able to overcome these problems, but that minority is the group that has a good grasp of the fundamentals and has broadened their knowledge through self-study, which earlier AI-reliant classes would make less likely to happen. In any case, I care about the average student, since we already have plenty of stuff about our institutions that makes life easier for a favored few while being worse for the average student (note that our construction of that favored few as the "good" students is a large part of this problem).
To summarize: because AI assistants introduce excess code complexity and difficult-to-debug bugs, they'll slow down rather than speed up project progress for the average student on moderately complex projects. On a fixed deadline, they'll result in worse projects, or necessitate less ambitious project scoping to ensure adequate completion, and I expect this remains broadly true through 4-6 years of study in most programs (don't take this as an endorsement of AI "assistants" for masters students; we've ignored a lot of other problems along the way).
There's a related problem: solving open-ended project assignments well ultimately depends on deeply understanding the problem, and AI "assistants" allow students to put a lot of code in their file without spending much time thinking about the problem or building an understanding of it. This is awful for learning outcomes, but also bad for project success. Getting students to see the value of thinking deeply about a problem is a thorny pedagogical puzzle at the best of times, and allowing the use of AI "assistants" makes the problem much much worse. This is another area I hope to see (or even drive) pedagogical improvement in, for what it's worth.
1/2

Day 6: Kamome Shirahama
Before I wander much father afield, I'd be remiss not to include at least one Mangaka (I've got 8 on my planning list; if you think Manga is pushing it just wait until you see what the next few days have in store).
I'm currently following "Witch Hat Atelier," and it's absolutely amazing in several dimensions: first class world-building, deep philosophical themes, nuanced diverse cast, tightly-constructed interwoven plots, deep mysteries that keep everything churning and show up in unexpected places, absolutely stellar art both in terms of in-panel depictions and page layouts (some are Watchmen-quality), especially if you are sartorially inclined, and general kindness of its core messages. This is a series I wish every programmer would read, because it includes excellent advice about software design in multiple ways (did I mention there's an intricate and logical magic system within which the main character innovates in legible-to-the-reader-as-innovation ways?). Also, I bet I would have enjoyed this just a much as a 10-year-old as I'm enjoying it in my 30's, which is something that takes well-honed skill to pull off.
Shirahama is a master of her craft, and I'm honestly kinda surprised to see Witch Hat is only her second series. Definitely thinking how I can get my hands on her earlier work in English.
#20AuthorsNoMen

TCIA: A Task-Centric Instruction Augmentation Method for Instruction Finetuning
Simin Ma, Shujian Liu, Jun Tan, Yebowen Hu, Song Wang, Sathish Reddy Indurthi, Sanqiang Zhao, Liwei Wu, Jianbing Han, Kaiqiang Song
https://arxiv.org/abs/2508.20374

TCIA: A Task-Centric Instruction Augmentation Method for Instruction Finetuning
Diverse instruction data is vital for effective instruction tuning of large language models, as it enables the model to generalize across different types of inputs . Building such diversified instruction dataset is an essential step in this process. Existing approaches often leverage large language models to automatically explore and generate diverse instructions, ensuring both data diversity and quality. However, they tend to overlook an important factor in real-world applications: on-task rel…

Touch in the Wild: Learning Fine-Grained Manipulation with a Portable Visuo-Tactile Gripper
Xinyue Zhu, Binghao Huang, Yunzhu Li
https://arxiv.org/abs/2507.15062

Touch in the Wild: Learning Fine-Grained Manipulation with a Portable Visuo-Tactile Gripper
Handheld grippers are increasingly used to collect human demonstrations due to their ease of deployment and versatility. However, most existing designs lack tactile sensing, despite the critical role of tactile feedback in precise manipulation. We present a portable, lightweight gripper with integrated tactile sensors that enables synchronized collection of visual and tactile data in diverse, real-world, and in-the-wild settings. Building on this hardware, we propose a cross-modal representatio…