Tootfinder

Universal Music lässt Musik zum Training von KI verwenden
Universal Music öffnet die Tür für das Training Künstlicher Intelligenz mit Werken seiner Künstler. Eine Klage gegen den Dienst Udio wird beigelegt.
https…

Make passive income today by training a crow to pick pockets for cash! :brdRobe:

Elon Musk says xAI bought a third building called "MACROHARDRR", reportedly adjacent to Colossus 2, that will take the company's training compute to almost 2GW (Kurt Wagner/Bloomberg)
https://www.bloomberg.com/news/articles/20

After physician residency applicants submit and certify their rank order list,
there isn’t much else left to do but wait on "Match Day". 
In a few weeks
—as was the case in 2024—
overwhelming majority of senior applicants from U.S. allopathic and osteopathic medical schools will find a training position
through the National Resident Matching Program (NRMP) Main Residency Match. 
But how does the system decide?
It all comes down to a
N…

"Après l'effort, le réconfort!"
(Following-up on this: #TombRaider2 works on

An interview with the team behind the News Creator Corps, whose fellowship provides 20 news influencers with journalism training and a $5K stipend (Neel Dhanesha/Nieman Lab)
https://www.niemanlab.org/2025/10/a-new-fellowship-is-all-about-putt…

A new fellowship is all about putting the news in news creator
The goal of the News Creator Corps is to "flood social platforms with accurate information."

PCIe / FPGA folks: the ThunderScope people are chasing a strange issue where the TS PCIe card appears to work properly on AMD hosts (even with fairly long lossy channels like OCuLink cables or >1 foot riser ribbons) but is failing to train the link at all (i.e. not even coming up as gen1 x1) on Intel hosts.
Ideas on root cause / fixes? I haven't scoped the training yet, as soon as I get an updated unit in my lab I will.

Microsofts Gaming Copilot macht Screenshots – aber nicht für KI-Training
Auf Vorwürfe, dass PC-Spieler mit der neuen Funktion überwacht würden, hat Microsoft nun reagiert. Die Antworten sind jedoch nicht ganz eindeutig.

Feature Flagging with #OpenFeature
https://training.linuxfoundation.org/training/feature-flagging-with-openfeature-lfs140/

Amazon opens Project Rainier, an $11B AI data center on 1,200 acres in Indiana that trains and runs Anthropic's AI models using 500K Amazon Trainium 2 chips (MacKenzie Sigalos/CNBC)
https://www.cnbc.com/2025/10/29/amazon-opens-11-bi…

Amazon opens $11 billion AI data center in rural Indiana as rivals race to break ground
Amazon has opened a date center on 1,200 acres of land in Indiana dedicated to training and running models from Anthropic.

Interim HC Mike Kafka reviews Week 17 vs. Raiders https://www.giants.com/video/interim-hc-mike-kafka-reviews-week-17-vs-raiders

In what may be close to an impossible job,
the “head of preparedness” at OpenAI
will be directly responsible for defending against risks from ever more powerful AIs to human mental health, cybersecurity and biological weapons.
That is before the successful candidate has to start worrying about the possibility that AIs may soon begin training themselves amid fears from some experts they could “turn against us”.
“This will be a stressful job,
and you’ll jump into …

Pretty Fit
A podcast for those who love training, running and pushing their limits...
Great Australian Pods Podcast Directory: #GreatAusPods

Cowboys Reunite With Training Camp Star Amid Trevon Diggs Update https://heavy.com/sports/nfl/dallas-cowboys/trevon-diggs-training-camp-star-poached-by-texans/?adt_ei=[email]

Was in Japan übrigens so nebenbei zum Thema gemacht wird:
https://www.instagram.com/p/DSqy86IAJPZ/?utm_source=ig_web_copy_link&igsh=NTc4MTIwNjQ2YQ==

Japan Daily on Instagram: "For the first time in Japan, foreign nationals have begun driving local bus routes under the Specified Skilled Worker program to address severe labor shortages. Tokyo Bus launched the initiative on December 23 in Tokyo and Okinawa, hiring five drivers from the Philippines after long-term training. The company says even higher wages have not attracted enough domestic drivers, making outside talent necessary. New drivers initially operate with support from Japanese staff and are expected to work independently after several months. Industry leaders hope the move will stabilize bus services and ensure passenger safety. Public understanding and continued training will be important as Japan adapts to workforce shortages in essential public services. @YAHOO"
85 likes, 11 comments - japandaily_jp on December 24, 2025: "For the first time in Japan, foreign nationals have begun driving local bus routes under the Specified Skilled Worker program to address severe labor shortages. Tokyo Bus launched the initiative on December 23 in Tokyo and Okinawa, hiring five drivers from the Philippines after long-term training. The company says even higher wages have not attracted enough domestic drivers, making outside talent necessary. New drivers initially ope…

To Perplexity AI: Short-term, the money may be in implantable visual prostheses for the blind. Longer-term, the money may be in training and support for noninvasive visual prostheses for the blind. Microsoft Windows vs Linux, Neuralink Blindsight vs The vOICe. https://www.perplexity.ai/s…

As NFL tight ends enjoy a golden era, passing on lessons has become personal https://www.nytimes.com/athletic/6735452/2025/10/25/nfl-tight-ends-career-training-position/

I so hope this blows up in his face spectacularly.
Note that I do not believe that any LLM can become as skilled as the least skilled #InfoSec professional using conventional tools, so I’m quite confident this will not reach any useful goal.
I’m saying he needs to learn a lesson, and it would be great if it were useful for others as well. @…

Today's alarm in the Radisson Blu Hotel will just be a DRILL!
Der heutige Alarm im Radisson Blu Hotel ist nur eine ÜBUNG!
#radissonbluhotel #39c3

Russia keeps hitting Ukraine's training centers — and Ukraine doesn't have an answer: https://benborges.xyz/2025/10/18/russia-keeps-hitting-ukraines-training.html

Inside an ICE Defense Training on Fortnite https://www.404media.co/ice-defense-training-on-fortnite-new-save-collective/

Anybody besides me thinking that maybe somebody’s got an #orca training program somewhere? We know the US has been training #dolphins / #porpoises for various purposes, but orcas are bigger, and meaner.

Some new ICE recruits have showed up to training without full vetting (NBC News)
https://www.nbcnews.com/politics/immigration/new-ice-recruits-showed-training-full-vetting-rcna238739
http://www.memeorandum.com/251022/p116#a251022p116

Who would have guessed.
Big Tech is paying millions to train teachers on AI, in a push to bring chatbots into classrooms
https://apnews.com/article/artificial-intelligence-teacher-union-microsoft-f7554b6550fb90519dd8129acac8e2…

Training Dynamics Impact Post-Training Quantization Robustness
Albert Catalan-Tatjer, Niccol\`o Ajroldi, Jonas Geiping
https://arxiv.org/abs/2510.06213 https://

Training Dynamics Impact Post-Training Quantization Robustness
While post-training quantization is widely adopted for efficient deployment of large language models, the mechanisms underlying quantization robustness remain unclear. We conduct a comprehensive analysis of quantization degradation across open-source language model training trajectories up to 32B parameters and 15T training tokens to accurately assess the relationship between training dynamics and quantization performance. Our key finding is that quantization errors in large-scale training runs…

Half of ICE's recruits couldn't pass an open book test... including allowing candidates to use their notes from the course they took.
https://www.nbcnews.com/politics/immigration/new-ice-recruits-showed-training-full-vetting-rcna2…

Jessie is training to be a stuntman
#BrookeBurns
Season 9 Episode 5 "The Natural"
#RandomBaywatch #lvdlpx

I escaped the #fog today for a while, took some photos, noticed that I should start training a bit more in the next weeks.
But most importantly, I just enjoyed being out, had some friendly words with some hikers and had some fresh air.
#hiking

Microsoft says its Gaming Copilot feature, which recently hit Windows 11 as a public beta, uses screenshots to understand in-game events, not to train AI models (Jowi Morales/Tom's Hardware)

Microsoft responds to Gaming Copilot controversy, says it uses screenshots to understand in-game events, not for training AI models — optional feature can be turned off, but not easily uninstalled
Microsoft says it isn't using your data to train its AI.

„Only in the world of AI algorithm training can you claim that you were torrenting 2,400 porn videos for personal use and have that seem like the lesser of two evils.“ 😂
https://www.vice.com/en/article/meta-says-the-2400-ad…

Meta Says the 2,400 ‘Adult Movies’ They Torrented Were for Personal Use, Not Training AI
Two adult entertainment companies say Meta’s appetite for information has gotten a little kinky... and now they're suing.

Sparky has been working as part of Antares canine search-and-rescue unit since 2013.
She also helped develop a training program for search dogs.
Over the years of service, Sparky has participated in hundreds of search operations,
saved people's lives
&helped find the bodies of more than 200 people

SentinelLabs' Dakota Cary linked Yu Yang and Qiu Daibing, two alleged members of the Chinese state hacking group, to participants of the 2012 Cisco Networking Academy Cup.
https://www.theregister.com/2025/12/11/salt_typhoon_cisco_training/

Researcher claims Salt Typhoon spies attended Cisco training scheme
: Skills gained later fed Beijing's cyber operations, according to SentinelLabs expert

Luca Antongirolami named to 20-man Under-16 Boys' National Team Roster for International Training Camp in Spain https://www.sportingkc.com/news/luca-antongirolami-named-to-20-man-under-16-boys-natio…

I wonder if programming language innovation will drop off post-AI. Like there's not much point learning a new language if you're not programming it directly, and LLMs know less about them.
You'll just reach for the language with the most (non-LLM generated) training data which will always be the older ones.

Kick it out!
https://winfuture.de/news,155123.html

Gmail & Co: Google liest Mails & Nutzerdaten jetzt für KI-Training aus
Google hat laut dem Sicherheitsdienstleister MalwareBytes begonnen, die E-Mails der Nutzer von Gmail sowie andere User-Daten über die meisten seiner Dienste hinweg zum KI-Training zu verwenden. In der EU gilt dabei wohl eine erfreuliche Ausnahme.

Getting Your Indices in a Row: Full-Text Search for LLM Training Data for Real World
Ines Altemir Marinas, Anastasiia Kucherenko, Alexander Sternfeld, Andrei Kucharavy
https://arxiv.org/abs/2510.09471 …

Getting Your Indices in a Row: Full-Text Search for LLM Training Data for Real World
The performance of Large Language Models (LLMs) is determined by their training data. Despite the proliferation of open-weight LLMs, access to LLM training data has remained limited. Even for fully open LLMs, the scale of the data makes it all but inscrutable to the general scientific community, despite potentially containing critical data scraped from the internet. In this paper, we present the full-text indexing pipeline for the Apertus LLM training data. Leveraging Elasticsearch parallel i…

I am wondering how easy it is for a bigtech company to collect all fediverse data for AI training? Are there signs that they are doing that? In my personal opinion it is fine dat AI models train on public posts but i was just wondering if it happens...
#AI #trainingdata

i am assigned to two exam committees for january prelims, one on energy-efficient model training and the other on i/o-compute balance in GPU-based data analytics. yummy.

You can really see the results from all the training I've been doing

ACE-G: Improving Generalization of Scene Coordinate Regression Through Query Pre-Training
Leonard Bruns, Axel Barroso-Laguna, Tommaso Cavallari, \'Aron Monszpart, Sowmya Munukutla, Victor Adrian Prisacariu, Eric Brachmann
https://arxiv.org/abs/2510.11605

ACE-G: Improving Generalization of Scene Coordinate Regression Through Query Pre-Training
Scene coordinate regression (SCR) has established itself as a promising learning-based approach to visual relocalization. After mere minutes of scene-specific training, SCR models estimate camera poses of query images with high accuracy. Still, SCR methods fall short of the generalization capabilities of more classical feature-matching approaches. When imaging conditions of query images, such as lighting or viewpoint, are too different from the training views, SCR models fail. Failing to genera…

Russian missile strikes Ukrainian training ground, casualties reported, military says: https://benborges.xyz/2025/10/16/russian-missile-strikes-ukrainian-training.html

This is a truly fascinating use of LLMs. I hope we get to see this soon.
https://github.com/DGoettlich/history-llms
A family of 4 billion (B) parameter LLMs based on the Qwen3 architecture trained on 80B tokens of historical data up to knowledge-cutoffs : 1913, 1929, 1933, 1939…

GitHub - DGoettlich/history-llms: Information hub for our project training the largest possible historical LLMs.
Information hub for our project training the largest possible historical LLMs. - DGoettlich/history-llms

Kathi asks ChatGPT "how to test a graph algo".
Repeats the query 6 months later…now it mentions PBT.
Huh! We ask it why it mentioned PBT. It cites back…
Our own papers and blog, and specifically mentions our own education research work on why graph algos are a great eg.
Moving the needle!

GOAT: A Training Framework for Goal-Oriented Agent with Tools
Hyunji Min, Sangwon Jung, Junyoung Sung, Dosung Lee, Leekyeung Han, Paul Hongsuck Seo
https://arxiv.org/abs/2510.12218

GOAT: A Training Framework for Goal-Oriented Agent with Tools
Large language models (LLMs) have recently been extended beyond traditional text generation to serve as interactive agents capable of using external tools based on user intent. However, current LLM agents still show limited ability to handle goal-oriented queries, which require decomposing a high-level objective into multiple interdependent API calls with correct planning and execution. Current approaches mainly rely on zero-shot evaluation due to the absence of training data. While proprietary…

Sharing this trick is my contribution to improve the quality of life of every corporate and government employee:
https://mastodon.cloud/@lrz/115350971038126826

Day 29: Leanne Betasamosake Simpson
I've been sitting on Simpson for a while because there's some overlap in her writing with Robin Wall Kimmerer, and I've had a lot of different genres/styles/subjects/media I've wanted to post at least one author from. But I've now hit repeats on at least YA romance and manga, and Simpson's writing is actually quite different from Kimmerer's in a lot of ways. While Kimmerer is a biologist by training and literally braids that knowledge together with her knowledge of Potawatomi cosmology and ethics, Simpson is an Anishinaabe philosopher and anarchist, and her position as a scholar of Indigenous philosophy adds a different depth to her work: she talks in more depth about knowledge relationships and her connections with specific elders, and she has more citations to other Indigenous theorists, which is the one criticism I've ever seen of Kimmerer's work. Rather than being Indigenous and a scientist, she's Indigenous and a scholar of indigenous studies.
I've only read "Theory of Water" by Simpson, but it was excellent, and especially inspiring to read as an anarchist. Simpson's explicit politics are another difference from Kimmerer's work, which is more implicitly than explicitly political. This allows Simpson to draw extremely interesting connections to other anarchist theorists and movements. "Theory of Water" is probably a bit less accessible than "Braiding Sweetgrass," but it's richer from a theory perspective as a result.
In any case, Simpson is a magnificent writer, sharing personal insights and stories along with (and inseparable from) her theoretical ideas.
#30AuthorsNoMen

A headline emphasising the government's support for Britain's High Streets.
> Thousands to get free digital training so everyone has the chance to shop around for cheaper deals online
https://www.gov.uk/government…

Thousands to get free digital training so everyone has the chance to shop around for cheaper deals online
Free support to get people online is rolling out in communities in every region of the country.

Sources: Colts' Jones expected back for camp https://www.espn.com/nfl/story/_/id/47295871/sources-daniel-jones-expected-cleared-training-camp

Trotz Datenschutzbedenken: Jeder Zweite vertraut Chatbots bei Gesundheitsfragen
Viele begrüßen laut Bitkom-Umfrage KI im Gesundheitswesen und würden ihre Daten sogar zum Training zur Verfügung stellen. Die E-Patientenakte bleibt Sorgenkind.

Took a few minutes over lunch to iron out the last of the bugs in the happy path of PCIe link training.
There's no timeouts or fallback if there's problems but if the other side is happy, it will train up to L0 and then sit there ignoring all incoming traffic.
After a while, the link partner gets mad that it hasn't seen a single DLLP from me and drops the link. I don't implement recovery yet so things go downhill from there.

Interactive Training: Feedback-Driven Neural Network Optimization
Wentao Zhang, Yang Young Lu, Yuntian Deng
https://arxiv.org/abs/2510.02297 https://arxiv.…

Interactive Training: Feedback-Driven Neural Network Optimization
Traditional neural network training typically follows fixed, predefined optimization recipes, lacking the flexibility to dynamically respond to instabilities or emerging training issues. In this paper, we introduce Interactive Training, an open-source framework that enables real-time, feedback-driven intervention during neural network training by human experts or automated AI agents. At its core, Interactive Training uses a control server to mediate communication between users or agents and the…

It's so simple -- ZZZ
https://flipboard.com/@menshealth/fitness-jothu8c2z/-/a-KRphhj0uR2uaYX9aABbrSA:a:3199554-/0

How You Can Avoid Overtraining
Yes, you can work out too hard, too often. Here's what you need to know to keep your training on point. This is Your Quick Training Tip, a chance to learn how to work smarter in just a few moments so you can get right to your workout. Certain things fall squarely in the “can never have enough of” …

Gemini co-lead Oriol Vinyals says Gemini 3's gains come from better pre-training and post-training, contradicting the idea that pre-training gains are falling (Stephanie Palazzolo/The Information)
https://www.theinformation.com/articles/google-unseats-anthropic-…

Google Unseats Anthropic With Gemini 3
Before we get to Gemini 3, we want to highlight this morning’s news that former Harvard University President Larry Summers has resigned from OpenAI’s board of directors following a congressional committee’s release of emails showing a personal correspondence between him and convicted sex ...

So I was going to write a short piece about this, but I'm so plumb out of time, and it would be so weird to put this in Metacurity as a special report that I'm writing a social media thread instead.
Last summer, I wrote a piece about the Coinbase breach based on an extensive conversation I had with the exchange's CSO, Philip Martin, who has serious chops as a cybersecurity guy.

Behind the Coinbase breach: Bribery emerges as enterprise threat
Coinbase’s breach shows how bribery schemes — long used for SIM swaps — can be a potent enterprise attack vector. Experts urge security leaders to add bribery training and red-teaming to their cyber defense toolkits.

Breaking: Raiders DE Maxx Crosby placed on IR, will have knee surgery https://raiderswire.usatoday.com/story/sports/nfl/raiders/2025/12/27/raiders-de-maxx-crosby-injured-reserve-surgery-knee-injury/8792903…

Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers
Wenhan Ma, Hailin Zhang, Liang Zhao, Yifan Song, Yudong Wang, Zhifang Sui, Fuli Luo
https://arxiv.org/abs/2510.11370

Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers
Reinforcement learning (RL) has emerged as a crucial approach for enhancing the capabilities of large language models. However, in Mixture-of-Experts (MoE) models, the routing mechanism often introduces instability, even leading to catastrophic RL training collapse. We analyze the training-inference consistency of MoE models and identify a notable discrepancy in routing behaviors between the two phases. Moreover, even under identical conditions, the routing framework can yield divergent expert …

Can an octopus learn to play a piano? This guy, Mattias Krantz, tried to find out. He rescues a random octopus from a fish market and then tries to work out how to teach it to play. You can skip the training to 15:20 to hear their first duet, but the training is interesting too. https://youtu.be/PcWnQ7fYzwI?si=BMFFE7

MuckRock and Sunlight Research Center merge; Sunlight provides training, tools, and support for investigative reporting, with a focus on local news (Sarah Scire/Nieman Lab)
https://www.niemanlab.org/2025/10/rebuilding-the-research-desk-in-local-newsrooms/…

Rebuilding the research desk in local newsrooms
A merger between MuckRock and Sunlight Research Center will help slimmed-down newsrooms investigate local governments, public officials, and more.

ViCO: A Training Strategy towards Semantic Aware Dynamic High-Resolution
Long Cui, Weiyun Wang, Jie Shao, Zichen Wen, Gen Luo, Linfeng Zhang, Yanting Zhang, Yu Qiao, Wenhai Wang
https://arxiv.org/abs/2510.12793

ViCO: A Training Strategy towards Semantic Aware Dynamic High-Resolution
Existing Multimodal Large Language Models (MLLMs) suffer from increased inference costs due to the additional vision tokens introduced by image inputs. In this work, we propose Visual Consistency Learning (ViCO), a novel training algorithm that enables the model to represent images of varying semantic complexities using different numbers of vision tokens. The key idea behind our method is to employ multiple MLP connectors, each with a different image compression ratio, to downsample the vision …

KI-Training: Normalsprachlicher Nutzungsvorbehalt reicht OLG Hamburg nicht
Das OLG in Hamburg hat im Fall der Klage eines Fotografen gegen den KI-Modell-Trainingsverein Laion die Berufung zurückgewiesen, eine Revision aber zugelassen.

Microsoft, OpenAI, and Anthropic are providing millions to the American Federation of Teachers to build AI training hubs aimed at educating 400,000 teachers (Jocelyn Gecker/Associated Press)
https://apnews.com/article/artificial-int…

Microsoft and OpenAI invest millions in AI training for teachers
The country’s largest teachers unions are teaming up with tech companies like Microsoft and OpenAI to help educators build literacy in artificial intelligence. The tech giants are paying for training for teachers to apply AI skills in the classroom. For the companies, the partnerships are also opportunities to make inroads into schools and win over students in the race for AI dominance. American Federation of Teachers president Randi Weingarten says skepticism guided her negotiations, but the…

To ChatGPT: Can training (sighted) humans for visual-to-auditory sensory substitution help build cognitive reserve to effectively delay or help protect against future cognitive decline? https://chatgpt.com/share/68fa9147-eca0-8004-b4d9-333510229d1f Spoiler: "Put toge…

ChatGPT - VASS and cognitive reserve
Shared via ChatGPT

A bit annoying that free/open source communities are almost all hosted by a closed source platform owned by microsoft AI who use it to IP-wash all those projects by training LLMs.. While also supporting genocide etc.
Migrating the big @… and @…

New military training center for Ukrainian, NATO troops opens in Poland: https://benborges.xyz/2025/10/03/new-military-training-center-for.html

Colts QB Anthony Richardson has surgery on orbital fracture, hopes to return this season https://www.nytimes.com/athletic/6751396/2025/10/26/anthony-richardson-surgery-orbital-fracture-colts/

Sources: Alibaba and ByteDance are among the top Chinese companies training their latest LLMs in Southeast Asian data centers to access Nvidia's AI chips (Financial Times)
https://www.ft.com/content/96fe9898-a3a4-4a33-be1d-da06bdb6cb2b

Montag: US-Cyberangriffe auf Chinas Zeitbehörde, Wikipedia findet getarnte Bots
Pekings Vorwürfe an die NSA KI-Training mit Wikipedia Deutsche Bahn mit neuem ICE L Kommentar zum Platzen der KI-Blase Googles Foldable mit Entzündung

Car bomb in Moscow kills senior Russian general
-- the third such attack in a year

Lt. Gen. Fanil Sarvarov, who headed the army’s
Directorate for Operational Training,
died as a result of his injuries Monday,
when a bomb planted under his car exploded,
making him the third general killed by a car bomb in the past year.
His department plays a critical role in planning military operations and ensuring the combat readiness of the Russian army.

Car bomb in Moscow kills senior Russian general
Lt. Gen. Fanil Sarvarov, 56, had headed the Russian army’s Directorate for Operational Training and participated in Moscow’s war in Ukraine, investigators said.

Automated Evolutionary Optimization for Resource-Efficient Neural Network Training
Ilia Revin, Leon Strelkov, Vadim A. Potemkin, Ivan Kireev, Andrey Savchenko
https://arxiv.org/abs/2510.09566

Automated Evolutionary Optimization for Resource-Efficient Neural Network Training
There are many critical challenges in optimizing neural network models, including distributed computing, compression techniques, and efficient training, regardless of their application to specific tasks. Solving such problems is crucial because the need for scalable and resource-efficient models is increasing. To address these challenges, we have developed a new automated machine learning (AutoML) framework, Parameter Efficient Training with Robust Automation (PETRA). It applies evolutionary op…

Mid-Training of Large Language Models: A Survey
Kaixiang Mo, Yuxin Shi, Weiwei Weng, Zhiqiang Zhou, Shuman Liu, Haibo Zhang, Anxiang Zeng
https://arxiv.org/abs/2510.06826 https:…

Mid-Training of Large Language Models: A Survey
Large language models (LLMs) are typically developed through large-scale pre-training followed by task-specific fine-tuning. Recent advances highlight the importance of an intermediate mid-training stage, where models undergo multiple annealing-style phases that refine data quality, adapt optimization schedules, and extend context length. This stage mitigates diminishing returns from noisy tokens, stabilizes convergence, and expands model capability in late training. Its effectiveness can be ex…

Dallas Cowboys Week 8 inactives include promising wide receiver https://www.si.com/nfl/cowboys/news/dallas-cowboys-week-8-inactives-include-promising-wide-receiver

Advancing End-to-End Pixel Space Generative Modeling via Self-supervised Pre-training
Jiachen Lei, Keli Liu, Julius Berner, Haiming Yu, Hongkai Zheng, Jiahong Wu, Xiangxiang Chu
https://arxiv.org/abs/2510.12586

Advancing End-to-End Pixel Space Generative Modeling via Self-supervised Pre-training
Pixel-space generative models are often more difficult to train and generally underperform compared to their latent-space counterparts, leaving a persistent performance and efficiency gap. In this paper, we introduce a novel two-stage training framework that closes this gap for pixel-space diffusion and consistency models. In the first stage, we pre-train encoders to capture meaningful semantics from clean images while aligning them with points along the same deterministic sampling trajectory, …

NYT reporter John Carreyrou and five other writers sue xAI, Anthropic, Google, OpenAI, Meta, and Perplexity, accusing them of pirating their books to train AI (Blake Brittain/Reuters)
https://www.reuters.com/legal/government/n

Investigation: How Russia militarizes Ukrainian children, training them for a lifetime of war: https://benborges.xyz/2025/10/23/investigation-how-russia-militarizes-ukrainian.html

Regularized Random Fourier Features and Finite Element Reconstruction for Operator Learning in Sobolev Space
Xinyue Yu, Hayden Schaeffer
https://arxiv.org/abs/2512.17884 https://arxiv.org/pdf/2512.17884 https://arxiv.org/html/2512.17884
arXiv:2512.17884v1 Announce Type: new
Abstract: Operator learning is a data-driven approximation of mappings between infinite-dimensional function spaces, such as the solution operators of partial differential equations. Kernel-based operator learning can offer accurate, theoretically justified approximations that require less training than standard methods. However, they can become computationally prohibitive for large training sets and can be sensitive to noise. We propose a regularized random Fourier feature (RRFF) approach, coupled with a finite element reconstruction map (RRFF-FEM), for learning operators from noisy data. The method uses random features drawn from multivariate Student's $t$ distributions, together with frequency-weighted Tikhonov regularization that suppresses high-frequency noise. We establish high-probability bounds on the extreme singular values of the associated random feature matrix and show that when the number of features $N$ scales like $m \log m$ with the number of training samples $m$, the system is well-conditioned, which yields estimation and generalization guarantees. Detailed numerical experiments on benchmark PDE problems, including advection, Burgers', Darcy flow, Helmholtz, Navier-Stokes, and structural mechanics, demonstrate that RRFF and RRFF-FEM are robust to noise and achieve improved performance with reduced training time compared to the unregularized random feature model, while maintaining competitive accuracy relative to kernel and neural operator tests.
toXiv_bot_toot

MaP: A Unified Framework for Reliable Evaluation of Pre-training Dynamics
Jiapeng Wang, Changxin Tian, Kunlong Chen, Ziqi Liu, Jiaxin Mao, Wayne Xin Zhao, Zhiqiang Zhang, Jun Zhou
https://arxiv.org/abs/2510.09295

MaP: A Unified Framework for Reliable Evaluation of Pre-training Dynamics
Reliable evaluation is fundamental to the progress of Large Language Models (LLMs), yet the evaluation process during pre-training is plagued by significant instability that obscures true learning dynamics. In this work, we systematically diagnose this instability, attributing it to two distinct sources: \textit{Parameter Instability} from training stochasticity and \textit{Evaluation Instability} from noisy measurement protocols. To counteract both sources of noise, we introduce \textbf{MaP}, …

QB Kyler Murray on Cardinals' plans for new 2028 training facility: 'Hopefully I'm here to see it' https://www.nfl.com/news/qb-kyler-murray-on-cardinals-plans-for-new-2028-training-facility-hopefully-i-m-here-t…

A look at Detroit's Apple Developer Academy, launched after the BLM protests, which spends $20K per student, nearly 2x the budget of local community colleges (Paresh Dave/Wired)
https://www.wired.com/story/apple-app-making-course-michigan-state-university/<…

Apple’s App Course Runs $20,000 a Student. Is It Really Worth It?
Apple, Michigan taxpayers, and one of Detroit’s wealthiest families spent roughly $30 million training hundreds of people to build iPhone apps. Not everyone lands coding jobs right away.

A Framework for Low-Effort Training Data Generation for Urban Semantic Segmentation
Denis Zavadski, Damjan Kal\v{s}an, Tim K\"uchler, Haebom Lee, Stefan Roth, Carsten Rother
https://arxiv.org/abs/2510.11567

A Framework for Low-Effort Training Data Generation for Urban Semantic Segmentation
Synthetic datasets are widely used for training urban scene recognition models, but even highly realistic renderings show a noticeable gap to real imagery. This gap is particularly pronounced when adapting to a specific target domain, such as Cityscapes, where differences in architecture, vegetation, object appearance, and camera characteristics limit downstream performance. Closing this gap with more detailed 3D modelling would require expensive asset and scene design, defeating the purpose of…

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training
Yunxiang Zhang, Muhammad Khalifa, Lechen Zhang, Xin Liu, Ayoung Lee, Xinliang Frederick Zhang, Farima Fatahi Bayat, Lu Wang
https://arxiv.org/abs/2510.09354

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training
Large reasoning models exhibit long chain-of-thought reasoning with strategies such as backtracking and self-correction, though recent studies suggest that these abilities typically require additional training. We first investigate whether such behaviors can be elicited without any training. To this end, we propose a decoding-time approach, ThinkLogit, which utilizes logit arithmetic to tune a target large non-reasoning model for long reasoning using a substantially smaller reasoning model as t…

Laminar: A Scalable Asynchronous RL Post-Training Framework
Guangming Sheng, Yuxuan Tong, Borui Wan, Wang Zhang, Chaobo Jia, Xibin Wu, Yuqi Wu, Xiang Li, Chi Zhang, Yanghua Peng, Haibin Lin, Xin Liu, Chuan Wu
https://arxiv.org/abs/2510.12633

Laminar: A Scalable Asynchronous RL Post-Training Framework
Reinforcement learning (RL) post-training for Large Language Models (LLMs) is now scaling to large clusters and running for extended durations to enhance model reasoning performance. However, the scalability of existing RL frameworks is limited, as extreme long-tail skewness in RL trajectory generation causes severe GPU underutilization. Current asynchronous RL systems attempt to mitigate this, but they rely on global weight synchronization between the actor and all rollouts, which creates a ri…

DiT360: High-Fidelity Panoramic Image Generation via Hybrid Training
Haoran Feng, Dizhe Zhang, Xiangtai Li, Bo Du, Lu Qi
https://arxiv.org/abs/2510.11712 https://

DiT360: High-Fidelity Panoramic Image Generation via Hybrid Training
In this work, we propose DiT360, a DiT-based framework that performs hybrid training on perspective and panoramic data for panoramic image generation. For the issues of maintaining geometric fidelity and photorealism in generation quality, we attribute the main reason to the lack of large-scale, high-quality, real-world panoramic data, where such a data-centric view differs from prior methods that focus on model design. Basically, DiT360 has several key modules for inter-domain transformation a…

Scores of volunteers are training to protect their noncitizen neighbors from the potential deployment of military troops in New York City.
Organizers said they are building on what they’ve learned from the experience of resistance in Los Angeles, Portland, and Chicago,
where masked and unidentified agents have often used excessive force and broken federal laws to carry out the Trump administration’s devastating campaign against immigrants.
“We need to exercise repeatedly w…

Scores of NYC volunteers prepare for potential military deployment to their communities
Organizers report signs of an impending immigration enforcement after Trump threatened action against New York City if Zohran Mamdani won the mayoral race

EU will DSGVO schleifen – für KI und Cookie-Banner
Der von der EU-Kommission geplante "digitale Omnibus" würde bestehende Datenschutzrechte aufweichen. Es geht etwa um Cookies und das Training von KI-Systemen.
h…

Science-centric streamer CuriosityStream's Q3 revenue rose 46% YoY to $18.4M, driven by IP licensing deals for AI training that generated $23.4M through Sept. (Scharon Harding/Ars Technica)
https://arstechnica.com/gadgets/2025/1

Science-centric streaming service Curiosity Stream is an AI-licensing firm now
Curiosity Stream’s owner has more content for AI companies than it does for subscribers.

From Condensation to Rank Collapse: A Two-Stage Analysis of Transformer Training Dynamics
Zheng-An Chen, Tao Luo
https://arxiv.org/abs/2510.06954 https://a…

From Condensation to Rank Collapse: A Two-Stage Analysis of Transformer Training Dynamics
Although transformer-based models have shown exceptional empirical performance, the fundamental principles governing their training dynamics are inadequately characterized beyond configuration-specific studies. Inspired by empirical evidence showing improved reasoning capabilities under small initialization scales in language models, we employ the gradient flow analytical framework established in [Zhou et al. NeurIPS 2022] to systematically investigate linearized Transformer training dynamics. …

Q&A with Z.ai Director of Product Zixuan Li on Chinese AI models embracing open source, attracting global users for its GLM model, training on memes, and more (ChinaTalk)
https://www.chinatalk.media/p/the-zai-playbook

The Z.ai Playbook
Zixuan Li on the Chinese AI ecosystem

Varjo erneuert Premium-XR-Headsets für hochrealistische Flugsimulatoren
Varjo aktualisiert seine Mixed-Reality-Headsets für Simulatoren. Verbesserte Kameras und Eye-Tracking sollen das Training genauer und effizienter machen.

Peloton unveils its Cross Training Series, including a $1,695 Bike and $6,695 Tread Plus, and AI-based Peloton IQ to track workouts on new and old machines (Victoria Song/The Verge)
https://www.theverge.com/tech/789282/peloton-cross-training-…

Peloton ushers in a new era with revamped hardware and AI
Peloton just announced a complete product lineup refresh, adding swivel displays, upgraded specs, and the Peloton IQ AI into every model.

Bedrock Data, which has debuted ArgusAI to show enterprises what data their AI models access during training and to prevent data leakage, raised a $25M Series A (Chris Metinko/Axios)
https://www.axios.com/pro/enterprise-software-deals/2025/11/1…

Exclusive: Bedrock Data raises $25M to secure enterprise AI
Greylock Partners led the Series A.

Hugging Face details how it used its new tool, Skills, to fine tune LLMs using Claude, including for writing scripts, submitting jobs to cloud GPUs, and more (Hugging Face)
https://huggingface.co/blog/hf-skills-training

We Got Claude to Fine-Tune an Open Source LLM
We’re on a journey to advance and democratize artificial intelligence through open source and open science.

Andrej Karpathy unveils nanochat, a full-stack training and inference implementation of an LLM in a single, dependency-minimal codebase (Andrej Karpathy/@karpathy)
https://x.com/karpathy/status/1977755427569111362

Andrej Karpathy (@karpathy) on X
Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,

Chip giants' efforts to turn Phoenix into a US hub may hinge on training local workers; an estimated 115K local chip jobs are set to be created in four years (Peter S. Goodman/New York Times)
https://www.nytimes.com/2025/12/04/business/tsmc-arizona-workers-…

A Growing U.S. Tech Hub Needs Workers. Colleges Try to Keep Up.
The success of efforts to turn Phoenix into a dominant center of semiconductor manufacturing may hinge on efforts to train local workers.

Micro1, which helps AI labs find experts for data annotation, says it has crossed $100M in annualized revenue and fielded investment offers at a $2.5B valuation (Anna Tong/Forbes)
http://www.forbes.com/sites/annatong/2025

This 24 Year Old Built A Multibillion-Dollar AI Training Empire In Eight Months
Ali Ansari’s decision to turn micro1’s AI recruitment assistant into a data labeling business spiked the company’s valuation from $80 million to $2.5 billion.

Anker paid users of its Eufy security cameras $2 per video of staged or real package and car thefts to train its AI systems from December 2024 to February 2025 (Lorenzo Franceschi-Bicchierai/TechCrunch)
https://techcrunch.com/2025/10/04/anke

Anker offered Eufy camera owners $2 per video for AI training | TechCrunch
Hundreds of Eufy customers have donated hundreds of thousands of videos to train the company’s AI systems.