<P><A HREF="http://www.techmeme.com/251114/p32#a251114p32" TITLE="Techmeme permalink"><IMG WIDTH=11 HEIGHT=12 SRC="http://www.techmeme.com/img/pml.png" STYLE="border:none;padding:0;margin:0;"></A> Karissa Bell / <A HREF="https://www.engadget.com/">Engadget</A>:<BR>
<SPAN STYLE="font-size:1.3em;"><B><A HREF="https://www.engadget.com/social-media/x-is-finally-rolling-out-chat-its-dm-replacement-with-encryption-and-video-calling-233032571.html">X replaces DMs with Chat, a new messaging system that it says is E2EE, supports file sharing, video calling, and more, rolling out first to iOS and the web</A></B></SPAN>&nbsp; &mdash;&nbsp; It also adds disappearing messages and file sharing features.&nbsp; &mdash;&nbsp; X has finally revealed its long-promised chat platform &hellip; </P>

قدمت X نظام مراسلة جديدًا يسمى Chat، والذي يحل محل الرسائل المباشرة (DM). يتميز هذا النظام بالتشفير من طرف إلى طرف (E2EE)، ومشاركة الملفات، ومكالمات الفيديو، والرسائل المختفية. يتم طرحه أولاً على نظام iOS والويب، مما يمثل ترقية كبيرة في قدرات التواصل لدى X.

X ha introducido un nuevo sistema de mensajería llamado Chat, que reemplaza los mensajes directos (DM). Este sistema cuenta con cifrado de extremo a extremo (E2EE), intercambio de archivos, videollamadas y mensajes que desaparecen. El lanzamiento está disponible inicialmente en iOS y la web, marcando una mejora significativa en las capacidades de comunicación de X.

X a introduit un nouveau système de messagerie appelé Chat, qui remplace les messages directs (DM). Ce système propose un chiffrement de bout en bout (E2EE), le partage de fichiers, des appels vidéo et des messages éphémères. Le déploiement est d'abord disponible sur iOS et le web, marquant une mise à niveau significative des capacités de communication de X.

X has introduced a new messaging system called Chat, which replaces direct messages (DMs). This system features end-to-end encryption (E2EE), file sharing, video calling, and disappearing messages. The rollout is initially available on iOS and the web, marking a significant upgrade in X's communication capabilities.

X replaces DMs with Chat, a new messaging system that it says is E2EE, supports file sharing, video calling, and more, rolling out first to iOS and the web (Karissa Bell/Engadget)

arXiv:1906.07172v5 Announce Type: replace-cross 
Abstract: Equivariant neural networks are a class of neural networks designed to preserve symmetries inherent in the data. In this paper, we introduce a general method for modifying a neural network to enforce equivariance, a process we refer to as equivarification. We further show that group convolutional neural networks (G-CNNs) arise as a special case of our framework.

Equivariant neural networks and equivarification

arXiv:2311.13745v4 Announce Type: replace-cross 
Abstract: Diffusion models have become the most popular approach to deep generative modeling of images, largely due to their empirical performance and reliability. From a theoretical standpoint, a number of recent works have studied the iteration complexity of sampling, assuming access to an accurate diffusion model. In this work, we focus on understanding the sample complexity of training such a model; how many samples are needed to learn an accurate diffusion model using a sufficiently expressive neural network? Prior work showed bounds polynomial in the dimension, desired Total Variation error, and Wasserstein error. We show an exponential improvement in the dependence on Wasserstein error and depth, along with improved dependencies on other relevant parameters.

Improved Sample Complexity Bounds for Diffusion Model Training

arXiv:2410.22995v2 Announce Type: replace-cross 
Abstract: A hallmark of advanced artificial intelligence is the capacity to progress from passive visual perception to the strategic modification of visual information to facilitate complex reasoning. This advanced capability, however, remains critically underdeveloped in current Large Multi-modal Models (LMMs). The deficiency is often masked by evaluation metrics that prioritize final-answer accuracy, creating an illusion of competence where genuine reasoning is absent. Using the domain of geometric problem-solving as a precise instrument, we probe this issue through tasks that require constructing visual aids. To this end, we introduce \textbf{VisAidMath}, a challenging benchmark, and our novel Three-Layered Funnel Evaluation Framework. This framework moves beyond simple accuracy (ACCU) to scrutinize the generation of valid visual aids (PVA) and the soundness of subsequent reasoning steps (SPRS). Our extensive experiments on state-of-the-art models, including Doubao-Seed-1.6 and o4, reveal a profound ``Reasoning Illusion''. We observe that high surface-level accuracy conceals a catastrophic failure in the models' ability to produce valid visual aids or to reason from them. Our findings expose a fundamental schism between visual perception and logical deduction in modern LMMs. We host an evaluation platform at CodaBench for testing publicly. Homepage: https://nlp2ct.github.io/VisAidMathHomepage/ Evaluation: https://www.codabench.org/competitions/7634/

VisAidMath: Benchmarking Visual-Aided Mathematical Reasoning

arXiv:2508.01249v3 Announce Type: replace-cross 
Abstract: Large Language Model (LLM) agents offer a powerful new paradigm for solving various problems by combining natural language reasoning with the execution of external tools. However, their dynamic and non-transparent behavior introduces critical security risks, particularly in the presence of prompt injection attacks. In this work, we propose a novel insight that treats the agent runtime traces as structured programs with analyzable semantics. Thus, we present AgentArmor, a program analysis framework that converts agent traces into graph intermediate representation-based structured program dependency representations (e.g., CFG, DFG, and PDG) and enforces security policies via a type system. AgentArmor consists of three key components: (1) a graph constructor that reconstructs the agent's runtime traces as graph-based intermediate representations with control and data flow described within; (2) a property registry that attaches security-relevant metadata of interacted tools \& data, and (3) a type system that performs static inference and checking over the intermediate representation. By representing agent behavior as structured programs, AgentArmor enables program analysis for sensitive data flow, trust boundaries, and policy violations. We evaluate AgentArmor on the AgentDojo benchmark, the results show that AgentArmor can reduce the ASR to 3\%, with the utility drop only 1\%.

AgentArmor: Enforcing Program Analysis on Agent Runtime Trace to Defend Against Prompt Injection

arXiv:2511.14691v1 Announce Type: cross 
Abstract: Attention is the brain's ability to selectively focus on a few specific aspects while ignoring irrelevant ones. This biological principle inspired the attention mechanism in modern Transformers. Transformers now underpin large language models (LLMs) such as GPT, but at the cost of massive training and inference energy, leading to a large carbon footprint. While brain attention emerges from neural circuits, Transformer attention relies on dot-product similarity to weight elements in the input sequence. Neuromorphic computing, especially spiking neural networks (SNNs), offers a brain-inspired path to energy-efficient intelligence. Despite recent work on attention-based spiking Transformers, the core attention layer remains non-neuromorphic. Current spiking attention (i) relies on dot-product or element-wise similarity suited to floating-point operations, not event-driven spikes; (ii) keeps attention matrices that suffer from the von Neumann bottleneck, limiting in-memory computing; and (iii) still diverges from brain-like computation. To address these issues, we propose the Spiking STDP Transformer (S$^{2}$TDPT), a neuromorphic Transformer that implements self-attention through spike-timing-dependent plasticity (STDP), embedding query--key correlations in synaptic weights. STDP, a core mechanism of memory and learning in the brain and widely studied in neuromorphic devices, naturally enables in-memory computing and supports non-von Neumann hardware. On CIFAR-10 and CIFAR-100, our model achieves 94.35\% and 78.08\% accuracy with only four timesteps and 0.49 mJ on CIFAR-100, an 88.47\% energy reduction compared to a standard ANN Transformer. Grad-CAM shows that the model attends to semantically relevant regions, enhancing interpretability. Overall, S$^{2}$TDPT illustrates how biologically inspired attention can yield energy-efficient, hardware-friendly, and explainable neuromorphic models.

Attention via Synaptic Plasticity is All You Need: A Biologically Inspired Spiking Neuromorphic Transformer

arXiv:2511.08955v2 Announce Type: replace-cross 
Abstract: Simulating microstructure evolution (MicroEvo) is vital for materials design but demands high numerical accuracy, efficiency, and physical fidelity. Although recent studies on deep learning (DL) offer a promising alternative to traditional solvers, the field lacks standardized benchmarks. Existing studies are flawed due to a lack of comparing specialized MicroEvo DL models with state-of-the-art spatio-temporal architectures, an overemphasis on numerical accuracy over physical fidelity, and a failure to analyze error propagation over time. To address these gaps, we introduce MicroEvoEval, the first comprehensive benchmark for image-based microstructure evolution prediction. We evaluate 14 models, encompassing both domain-specific and general-purpose architectures, across four representative MicroEvo tasks with datasets specifically structured for both short- and long-term assessment. Our multi-faceted evaluation framework goes beyond numerical accuracy and computational cost, incorporating a curated set of structure-preserving metrics to assess physical fidelity. Our extensive evaluations yield several key insights. Notably, we find that modern architectures (e.g., VMamba), not only achieve superior long-term stability and physical fidelity but also operate with an order-of-magnitude greater computational efficiency. The results highlight the necessity of holistic evaluation and identify these modern architectures as a highly promising direction for developing efficient and reliable surrogate models in data-driven materials science.

MicroEvoEval: A Systematic Evaluation Framework for Image-Based Microstructure Evolution Prediction

arXiv:2511.11287v1 Announce Type: cross 
Abstract: The increasing deployment of autonomous AI agents on the web is hampered by a fundamental misalignment: agents must infer affordances from human-oriented user interfaces, leading to brittle, inefficient, and insecure interactions. To address this, we introduce VOIX, a web-native framework that enables websites to expose reliable, auditable, and privacy-preserving capabilities for AI agents through simple, declarative HTML elements. VOIX introduces <tool> and <context> tags, allowing developers to explicitly define available actions and relevant state, thereby creating a clear, machine-readable contract for agent behavior. This approach shifts control to the website developer while preserving user privacy by disconnecting the conversational interactions from the website. We evaluated the framework's practicality, learnability, and expressiveness in a three-day hackathon study with 16 developers. The results demonstrate that participants, regardless of prior experience, were able to rapidly build diverse and functional agent-enabled web applications. Ultimately, this work provides a foundational mechanism for realizing the Agentic Web, enabling a future of seamless and secure human-AI collaboration on the web.

تناقش المقالة تقديم VOIX، وهو إطار تصريحي مصمم لتعزيز التفاعل بين وكلاء الذكاء الاصطناعي وواجهات الويب. يتيح هذا الإطار للمطورين تعريف الإجراءات والحالات من خلال عناصر HTML بسيطة، مما يعزز القدرات الموثوقة والمحافظة على الخصوصية لوكلاء الذكاء الاصطناعي. أظهرت دراسة شملت 16 مطورًا أن المشاركين تمكنوا من إنشاء تطبيقات ويب متنوعة مدعومة بالوكلاء بسرعة، مما يبرز عملية الإطار وفعاليته.

El artículo presenta VOIX, un marco declarativo diseñado para mejorar la interacción entre agentes de IA e interfaces web. Este marco permite a los desarrolladores definir acciones y estados a través de etiquetas HTML simples, promoviendo capacidades fiables y respetuosas con la privacidad para los agentes de IA. Un estudio con 16 desarrolladores demostró que los participantes podían crear rápidamente diversas aplicaciones web habilitadas para agentes, destacando la practicidad y efectividad del marco.

L'article présente VOIX, un cadre déclaratif conçu pour améliorer l'interaction entre les agents d'IA et les interfaces web. Ce cadre permet aux développeurs de définir des actions et des états via des balises HTML simples, favorisant des capacités fiables et respectueuses de la vie privée pour les agents d'IA. Une étude impliquant 16 développeurs a montré que les participants pouvaient rapidement créer des applications web activées par des agents, soulignant la praticité et l'efficacité du cadre.

The article discusses the introduction of VOIX, a declarative framework designed to enhance the interaction between AI agents and web interfaces. This framework allows developers to define actions and states through simple HTML tags, promoting reliable and privacy-preserving capabilities for AI agents. A study involving 16 developers demonstrated that participants could quickly create diverse agent-enabled web applications, highlighting the framework's practicality and effectiveness.

Building the Web for Agents: A Declarative Framework for Agent-Web Interaction

arXiv:2510.25932v2 Announce Type: replace-cross 
Abstract: Social platforms distribute information at unprecedented speed, which in turn accelerates the spread of misinformation and threatens public discourse. We present FakeZero, a fully client-side, cross-platform browser extension that flags unreliable posts on Facebook and X (formerly Twitter) while the user scrolls. All computation, DOM scraping, tokenization, Transformer inference, and UI rendering run locally through the Chromium messaging API, so no personal data leaves the device. FakeZero employs a three-stage training curriculum: baseline fine-tuning and domain-adaptive training enhanced with focal loss, adversarial augmentation, and post-training quantization. Evaluated on a dataset of 239,000 posts, the DistilBERT-Quant model (67.6 MB) reaches 97.1% macro-F1, 97.4% accuracy, and an AUROC of 0.996, with a median latency of approximately 103 ms on a commodity laptop. A memory-efficient TinyBERT-Quant variant retains 95.7% macro-F1 and 96.1% accuracy while shrinking the model to 14.7 MB and lowering latency to approximately 40 ms, showing that high-quality fake-news detection is feasible under tight resource budgets with only modest performance loss. By providing inline credibility cues, the extension can serve as a valuable tool for policymakers seeking to curb the spread of misinformation across social networks. With user consent, FakeZero also opens the door for researchers to collect large-scale datasets of fake news in the wild, enabling deeper analysis and the development of more robust detection techniques.

FakeZero هو ملحق متصفح جديد مصمم لاكتشاف المعلومات المضللة على منصات وسائل التواصل الاجتماعي فيسبوك وX (تويتر سابقًا). تعمل هذه الأداة بالكامل على جهاز المستخدم، مما يضمن الخصوصية من خلال معالجة البيانات محليًا دون إرسال معلومات شخصية إلى خوادم خارجية. تستخدم تقنيات متقدمة في التعلم الآلي، حيث تحقق درجة F1 الكلية بنسبة 97.1% ودقة بنسبة 97.4% على مجموعة بيانات تضم 239,000 منشور، مما يجعلها تقدمًا كبيرًا في اكتشاف المعلومات المضللة في الوقت الفعلي.

FakeZero es una nueva extensión de navegador diseñada para detectar desinformación en las plataformas de redes sociales Facebook y X (anteriormente Twitter). Esta herramienta opera completamente en el dispositivo del usuario, garantizando la privacidad al procesar datos localmente sin enviar información personal a servidores externos. Utiliza técnicas avanzadas de aprendizaje automático, logrando un puntaje macro-F1 del 97.1% y una precisión del 97.4% en un conjunto de datos de 239,000 publicaciones, lo que representa un avance significativo en la detección de desinformación en tiempo real.

FakeZero est une nouvelle extension de navigateur conçue pour détecter la désinformation sur les plateformes de médias sociaux Facebook et X (anciennement Twitter). Cet outil fonctionne entièrement sur l'appareil de l'utilisateur, garantissant la confidentialité en traitant les données localement sans envoyer d'informations personnelles à des serveurs externes. Il utilise des techniques avancées d'apprentissage automatique, atteignant un score macro-F1 de 97,1 % et une précision de 97,4 % sur un ensemble de données de 239 000 publications, ce qui en fait une avancée significative dans la détec…

FakeZero is a new browser extension designed to detect misinformation on social media platforms Facebook and X (formerly Twitter). This tool operates entirely on the user's device, ensuring privacy by processing data locally without sending personal information to external servers. It utilizes advanced machine learning techniques, achieving a macro-F1 score of 97.1% and an accuracy of 97.4% on a dataset of 239,000 posts, making it a significant advancement in real-time misinformation detection.

FakeZero: Real-Time, Privacy-Preserving Misinformation Detection for Facebook and X

<p>Lately, I have been sharing my tech progress across my social platforms, and I realized…<br>
why not reach a wider, more developer-focused community?<br>
So here I am, and I am really glad to be here.<br>
Here is to sharing my growth, one post at a time☺️🥂<br>
If you are curious about how my learning journey began, you can find more of my story on <a href="https://www.instagram.com/promise.syl" rel="noopener noreferrer">Instagram</a>, <a href="https://www.linkedin.com/in/sylpromise" rel="noopener noreferrer">LinkedIn</a>, <a href="https://x.com/PpWanmo" rel="noopener noreferrer">X</a> or <a href="https://www.threads.com/@promise.syl" rel="noopener noreferrer">Thread</a>.</p>

<p>I shared one of the most interesting parts of my Data Science learning so far, the finance side of data analysis, where I merged a US GDP and S&amp;P500 (stock market index)to see if GDP can help predict stock prices.</p>

<p>If you’d like to see the post that inspired today’s update, here it is:<br>
</p>
<div class="crayons-card c-embed text-styles text-styles--secondary">
    <div class="c-embed__content">
        <div class="c-embed__cover">
          <a href="https://www.threads.com/@promise.syl/post/DRFr89MjGma?xmt=AQF0oeEaPp-5I8JB2t36rPfH8YfJEqcw-YyY11rGjpJEiKlA124ABmU7SrvebKLD--2kUAY&amp;amp%3Bslof=1" class="c-link align-middle" rel="noopener noreferrer">
            <img alt="" src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fscontent-iad3-2.cdninstagram.com%2Fv%2Ft51.2885-19%2F539440476_17848964157542823_3558075179840778002_n.jpg%3Fstp%3Ddst-jpg_s640x640_tt6%26efg%3DeyJ2ZW5jb2RlX3RhZyI6InByb2ZpbGVfcGljLmRqYW5nby45NTMuYzIifQ%26_nc_ht%3Dscontent-iad3-2.cdninstagram.com%26_nc_cat%3D103%26_nc_oc%3DQ6cZ2QFtvQQEZPBoYdfTihsHVwbRLF3QApNylKkbkYDDNfeD9TSiX2fgsNvqiRv3mrCuoWI%26_nc_ohc%3DcVjsETx04y8Q7kNvwG5LJMS%26_nc_gid%3DieZEYURWWNuDBZKmY6ZfQw%26edm%3DAPs17CUBAAAA%26ccb%3D7-5%26oh%3D00_AfhOymV208fegFPaYi6L-GkD0SApM2Lmty5Qx4V46s3gtQ%26oe%3D691FC103%26_nc_sid%3D10d13b" height="640" class="m-0" width="640">
          </a>
        </div>
      <div class="c-embed__body">
        <h2 class="fs-xl lh-tight">
          <a href="https://www.threads.com/@promise.syl/post/DRFr89MjGma?xmt=AQF0oeEaPp-5I8JB2t36rPfH8YfJEqcw-YyY11rGjpJEiKlA124ABmU7SrvebKLD--2kUAY&amp;amp%3Bslof=1" rel="noopener noreferrer" class="c-link">
            Day 8 of improving my Data Science skills 

I welcome you all to the finance side of my Data Science learning journey!🤗

Please, kindly have your seats and pay attention while I give you some tips you might or might not need to make  money! You might learn a thing or two.

Today I worked on a super simple task: checking if the US GDP and the S&amp;P 500 move together, that is, I checked to see if GDP can help predict stock prices.
          </a>
        </h2>
          <p class="truncate-at-3">
            Day 8 of improving my Data Science skills 

I welcome you all to the finance side of my Data Science learning journey!🤗

Please, kindly have your seats and pay attention while I give you some tips you might or might not need to make  money! You might learn a thing or two.

Today I worked on a super simple task: checking if the US GDP and the S&amp;P 500 move together, that is, I checked to see if GDP can help predict stock prices.
          </p>
        <div class="color-secondary fs-s flex items-center">
            <img alt="favicon" class="c-embed__favicon m-0 mr-2 radius-0" src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fstatic.cdninstagram.com%2Frsrc.php%2Fye%2Fr%2FlEu8iVizmNW.ico" width="48" height="48">
          threads.com
        </div>
      </div>
    </div>
</div>


<h2>
  
  
  What to expect from me here
</h2>

<p>This is the beginning of my Dev.to journey.<br>
I will be sharing what I learn daily,<br>
the wins,<br>
the confusion,<br>
and the “ohhh now I get it” moments.</p>

<p>If you enjoy real-time learning, especially from someone navigating Data Science step-by-step, I’d love to connect with you here.<br>
Thanks for reading, see you later with another update.</p>

يعبّر بروس، المعروف أيضًا باسم SP، عن حماسه لمشاركة رحلته التقنية مع مجتمع المطورين. لقد قام بتوثيق تقدمه على منصات التواصل الاجتماعي مثل إنستغرام ولينكد إن وX وThread. يبرز بروس جانبًا مثيرًا من تعلمه في علم البيانات، وهو دمج بيانات الناتج المحلي الإجمالي الأمريكي مع مؤشر S&P 500 لاستكشاف إمكانية استخدام الناتج المحلي الإجمالي في التنبؤ بأسعار الأسهم. يدعو الآخرين لمتابعة نموه من خلال منشوراته.

Promise, también conocido como SP, expresa su entusiasmo por compartir su viaje tecnológico con la comunidad de desarrolladores. Han estado documentando su progreso en plataformas de redes sociales como Instagram, LinkedIn, X y Thread. Promise destaca un aspecto interesante de su aprendizaje en Ciencia de Datos, específicamente la fusión de datos del PIB de EE. UU. con el índice bursátil S&P 500 para explorar el potencial del PIB en la predicción de precios de acciones. Invitan a otros a seguir su crecimiento a través de sus publicaciones.

Promise, également connu sous le nom de SP, exprime son enthousiasme à partager son parcours technologique avec la communauté des développeurs. Ils ont documenté leurs progrès sur des plateformes de médias sociaux telles qu'Instagram, LinkedIn, X et Thread. Promise met en avant un aspect intéressant de son apprentissage en Data Science, en particulier la fusion des données du PIB américain avec l'indice boursier S&P 500 pour explorer le potentiel du PIB à prédire les prix des actions. Ils invitent les autres à suivre leur croissance à travers leurs publications.

Promise, also known as SP, expresses excitement about sharing their tech journey with the developer community. They have been documenting their progress on social media platforms like Instagram, LinkedIn, X, and Thread. Promise highlights an interesting aspect of their Data Science learning, specifically merging US GDP data with the S&P 500 stock market index to explore the potential of GDP in predicting stock prices. They invite others to follow their growth through their posts.

👋 Hi Dev Community! I am Promise (you can call me SP), I am super excited to share my journey here too!

<A HREF="https://www.anthropic.com/news/claude-in-microsoft-foundry"><IMG VSPACE="4" HSPACE="4" BORDER="0" ALIGN="RIGHT" SRC="http://www.techmeme.com/251118/i55.jpg"></A>
<P><A HREF="http://www.techmeme.com/251118/p55#a251118p55" TITLE="Techmeme permalink"><IMG WIDTH=11 HEIGHT=12 SRC="http://www.techmeme.com/img/pml.png" STYLE="border:none;padding:0;margin:0;"></A> <A HREF="https://www.anthropic.com/">Anthropic</A>:<BR>
<SPAN STYLE="font-size:1.3em;"><B><A HREF="https://www.anthropic.com/news/claude-in-microsoft-foundry">Anthropic's Claude Sonnet 4.5, Haiku 4.5, and Opus 4.1 are now available in public preview in Microsoft Foundry for Azure customers</A></B></SPAN>&nbsp; &mdash;&nbsp; Today we announced that Microsoft and Anthropic are expanding our partnership.&nbsp; As part of the partnership, Claude Sonnet 4.5, Haiku 4.5 &hellip; </P>

Anthropic's Claude Sonnet 4.5, Haiku 4.5, and Opus 4.1 are now available in public preview in Microsoft Foundry for Azure customers (Anthropic)

<A HREF="https://www.bloomberg.com/news/articles/2025-11-18/trump-urges-congress-to-block-state-level-ai-regulation"><IMG VSPACE="4" HSPACE="4" BORDER="0" ALIGN="RIGHT" SRC="http://www.techmeme.com/251118/i54.jpg"></A>
<P><A HREF="http://www.techmeme.com/251118/p54#a251118p54" TITLE="Techmeme permalink"><IMG WIDTH=11 HEIGHT=12 SRC="http://www.techmeme.com/img/pml.png" STYLE="border:none;padding:0;margin:0;"></A> Kate Sullivan / <A HREF="https://www.bloomberg.com/">Bloomberg</A>:<BR>
<SPAN STYLE="font-size:1.3em;"><B><A HREF="https://www.bloomberg.com/news/articles/2025-11-18/trump-urges-congress-to-block-state-level-ai-regulation">Trump says the US needs a &ldquo;Federal Standard&rdquo; governing oversight of AI instead of &ldquo;a patchwork of 50 State Regulatory Regimes&rdquo;, otherwise China will catch up</A></B></SPAN>&nbsp; &mdash;&nbsp; President Donald Trump called on Congress to pass a federal standard governing oversight of artificial intelligence &hellip; </P>

Trump says the US needs a "Federal Standard" governing oversight of AI instead of "a patchwork of 50 State Regulatory Regimes", otherwise China will catch up (Kate Sullivan/Bloomberg)

<A HREF="https://www.businessinsider.com/meta-soumith-chintala-mira-murati-thinking-machines-lab-pytorch-ai-2025-11"><IMG VSPACE="4" HSPACE="4" BORDER="0" ALIGN="RIGHT" SRC="http://www.techmeme.com/251118/i53.jpg"></A>
<P><A HREF="http://www.techmeme.com/251118/p53#a251118p53" TITLE="Techmeme permalink"><IMG WIDTH=11 HEIGHT=12 SRC="http://www.techmeme.com/img/pml.png" STYLE="border:none;padding:0;margin:0;"></A> Pranav Dixit / <A HREF="https://www.businessinsider.com/">Business Insider</A>:<BR>
<SPAN STYLE="font-size:1.3em;"><B><A HREF="https://www.businessinsider.com/meta-soumith-chintala-mira-murati-thinking-machines-lab-pytorch-ai-2025-11">Soumith Chintala, who co-created the PyTorch ML framework at Meta and left the company this month, has joined Mira Murati's Thinking Machines Lab</A></B></SPAN>&nbsp; &mdash;&nbsp; - Soumith Chintala, the creator of PyTorch, has joined Mira Murati's AI startup Thinking Machines Lab.&nbsp; &mdash; Chintala's move comes amid major restructuring inside Meta's AI ranks.</P>

Soumith Chintala, who co-created the PyTorch ML framework at Meta and left the company this month, has joined Mira Murati's Thinking Machines Lab (Pranav Dixit/Business Insider)

X replaces DMs with Chat, a new messaging system that it says is E2EE, supports file sharing, video calling, and more, rolling out first to iOS and the web (Karissa Bell/Engadget)

Was this article worth reading? Share it