All entries

Geometrically Principled Randomized Optimization for Efficient LLM Training

model-releasesarxiv-cs-lg

arXiv:2510.01878v2 Announce Type: replace Abstract: Low-rank gradient optimization for large language models is currently divided into two categories: structured methods that rigorously identify subsp

Industry

Getty Images announces plans to end its Shutterstock merger after the UK CMA conditioned approval of the deal on Shutterstock selling its editorial business (Elias Schisgall/Wall Street Journal)

DGX agent

Elias Schisgall / Wall Street Journal: Getty Images announces plans to end its Shutterstock merger after the UK CMA conditioned approval of the deal on Shutterstock selling its editorial business — A

industrytechmeme

30 Jun 2026

Giving a talk on agent-to-agent and AI network effects at @swyx 's AI Engineer World Fair today at 1:30p in Room 2010. Come say hi! I think …

Giving a talk on agent-to-agent and AI network effects at @swyx 's AI Engineer World Fair today at 1:30p in Room 2010. Come say hi! I think this talk will be a good one if I may say so myself. https:/

agentsswyx--x

Giving agents the ability to write code makes them dramatically more capable. It also makes security a lot harder. At LangChain, we have spe…

Giving agents the ability to write code makes them dramatically more capable. It also makes security a lot harder. At LangChain, we have spent a lot of time this year figuring out how to do both. http

GLACIER: Rethinking Mass Spectrum Prediction as an Object Detection Problem

arXiv:2606.29161v1 Announce Type: new Abstract: Predicting tandem mass spectra (MS/MS) from molecular structures represents a central task in analytical chemistry with direct relevance to clinical met

Local Ai

GLIP: Graph and LLM Joint Pretraining for Graph-Level Tasks

arXiv:2606.29773v1 Announce Type: new Abstract: Graphs are widely used to model relational systems, with applications in domains such as social networks, finance, and biomedicine. Graph neural network

local-aiarxiv-cs-lg

GLM-5.2 is now @Zai_org's most-liked model on Hugging Face of all time. http://huggingface.co/models?p=1&sort=likes

GLM-5.2 has become the most-liked model of all time on Hugging Face's model repository, surpassing previous records. This achievement reflects significant community recognition and adoption of the mod

industryclem-delangue--x

GLM-5.2 just tied for #2 on SportEval’s World Cup Exact Score Hit Rate leaderboard. ⚽ 13 exact-score hits out of 76 matches 📊 17.1% accurac…

model-releaseszhipu-ai--x

GLM-5.2 just tied for #2 on SportEval’s World Cup Exact Score Hit Rate leaderboard. ⚽ 13 exact-score hits out of 76 matches 📊 17.1% accuracy 🥈 Second only to GPT-5.5 That puts @Zai_org ahead of Claude

Goku: A Million-Scale Universal Dataset and Benchmark for Instruction-Based Video Editing

model-releasesarxiv-cs-cv

arXiv:2606.30599v1 Announce Type: new Abstract: Existing instruction-based video editing datasets commonly focus on single-task appearance editing, failing to meet the complex creative demands of real

Safety

Golden Hour Divide: Trauma Care Accessibility and Resource Vulnerability in Sri Lanka

arXiv:2606.29889v1 Announce Type: new Abstract: Timely intensive care dictates survival, yet emergency infrastructure remains unevenly distributed across Sri Lanka. While pre-hospital services have ex

safetyarxiv-cs-lg

good description of how someone has done wiki memory in the wild

good description of how someone has done wiki memory in the wild @hwchase17 @cognition @FactoryAI @karpathy We implemented wiki idea by hosting the 'files' as doc db objects, added an API for db query

Good Enough? An Investigation on the Impact of Label Quality in Large-Scale Medical Datasets

arXiv:2505.20928v2 Announce Type: replace Abstract: Manually refining radiological segmentation masks is highly resource-intensive. To determine when this expert commitment is truly justified for the

Google kills Tenor GIF API, forcing changes at X, Discord, and more

Google shut down the Tenor API on June 30, 2026, breaking GIF pickers in Discord, X, Bluesky, and WhatsApp. Platforms that built GIF search into their products using Tenor are now scrambling for repla

industryars-technica

Google launches Nano Banana 2 Lite, a low cost text-to-image model that delivers outputs in four seconds, and rolls out Gemini Omni Flash to developers (The Keyword)

The Keyword: Google launches Nano Banana 2 Lite, a low cost text-to-image model that delivers outputs in four seconds, and rolls out Gemini Omni Flash to developers — We're making it easier to experim

model-releasestechmeme

Google shuts down the Tenor API, affecting GIF pickers on platforms like Discord, WhatsApp, and Bluesky; Nikita Bier says X has migrated elsewhere (Ben Schoon/9to5Google)

Ben Schoon / 9to5Google: Google shuts down the Tenor API, affecting GIF pickers on platforms like Discord, WhatsApp, and Bluesky; Nikita Bier says X has migrated elsewhere — GIFs are a huge part of th

industrytechmeme

Google’s Gemini Omni Flash and Nano Banana 2 Lite support slick media content creation at lower costs

model-releasessiliconangle

Google LLC is enhancing its generative artificial intelligence capabilities for creators with the debut of a pair of new media-focused models in the Gemini Enterprise Agent Platform. The new additions

Google's new Nano Banana 2 Lite image model is its fastest and cheapest yet

Google introduced Nano Banana 2 Lite, its fastest and most cost-efficient image model built for high throughput, speed and scale. The model generates text-to-image outputs in about four seconds and co

industryars-technica

Google’s NotebookLM can sum up your research in a TikTok-style clip

Google's NotebookLM is adding a new way to catch up on your notes: TikTok-style AI videos. The new feature is rolling out to Google AI Ultra and Pro subscribers, allowing NotebookLM to generate 60-sec

industrythe-verge-ai

Tools

Got this at ai engineer world fair lol @swyx

This post appears to be a casual social media share from Swyx about an item or experience from an AI Engineer World Fair event, likely featuring a photo or humorous take on something encountered at th

toolsswyx--x

Governance Decay: How Context Compaction Silently Erases Safety Constraints in Long-Horizon LLM Agents

arXiv:2606.22528v2 Announce Type: replace Abstract: Modern LLM agents increasingly rely on context compaction, summarization, or eviction to keep long-running sessions within a token budget. We show t

Safety

GPC: Large-Scale Generative Pretraining for Transferable Motor Control

arXiv:2606.29148v1 Announce Type: cross Abstract: Developing controllers capable of completing a wide range of tasks in a natural and life-like manner is a key challenge in enabling practical applicat

safetyarxiv-cs-ai

GPTNT: Benchmarking Real-Time Collaboration Between Multimodal Agents on Keep Talking And Nobody Explodes

arXiv:2606.28514v1 Announce Type: new Abstract: Multimodal models are increasingly deployed to solve tasks collaboratively with humans or other artificial agents. Existing benchmarks show that these m

Hardware

GPU-Accelerated Inverse Structural Anastylosis from Block Collapse Dynamics

arXiv:2606.28394v1 Announce Type: new Abstract: The physical anastylosis of collapsed architectural monuments -- the meticulous reassembly of fallen stone elements into their original structural confi

hardwarearxiv-cs-cv

Hardware

GPU Parallelization Strategies for Forward and Backward Propagation in Shallow Neural Networks: A CUDA-Based Comparative Study

arXiv:2606.30497v1 Announce Type: cross Abstract: We present a comparative study of CUDA optimization strategies applied to forward and backward propagation in a shallow neural network. Three stacked

hardwarearxiv-cs-lg

Applications

Gradient Boosted Mixed Models: Flexible Estimation of Mean and Variance Components for Clustered Data

arXiv:2511.00217v2 Announce Type: replace-cross Abstract: We introduce Gradient Boosted Mixed Models (GBMixed), a framework which extends boosting to clustered data by jointly modeling the mean and va

applicationsarxiv-cs-lg

Gradient boosting with vector-valued leafs

arXiv:2606.29326v1 Announce Type: cross Abstract: Gradient boosting in the form of decision tree ensembles has successfully been applied to a variety of problems using simple objective functions based

Granular-ball computing: an efficient, robust, and interpretable adaptive multi-granularity representation and computation method

arXiv:2304.11171v5 Announce Type: replace-cross Abstract: To overcome the limitations of point-based inputs, overly fine computation and limited adaptability in existing artificial intelligence method

researcharxiv-cs-ai

Graph-GSReg: Leveraging 3D Scene Graphs for Gaussian Splatting Registration

arXiv:2606.29782v1 Announce Type: new Abstract: Merging multiple 3D Gaussian Splatting (3DGS) scenes into a single unified Gaussian representation is essential for large-scale 3D mapping and long-term

GraphChase: A Platform and Benchmark for Urban Network Security Games

arXiv:2501.17559v2 Announce Type: replace Abstract: After the achievement of solving two-player zero-sum games, more AI researchers focus on solving multiplayer games. Urban Network Security Games (ex

Safety

Grasp-Oriented Non-Prehensile Manipulation via Learning a Graspability Field

arXiv:2606.30474v1 Announce Type: new Abstract: Non-prehensile manipulation is often used as a preparatory step for robotic grasping, yet existing approaches typically require a predefined target obje

safetyarxiv-cs-ro

Tools

grats to @swyx 99% of my timeline is @aiDotEngineer now.

This post celebrates @swyx's influence on the poster's social media feed, noting that approximately 99% of their timeline now consists of content from @aiDotEngineer (likely a reference to Swyx's AI e

toolsswyx--x

Gravitational Duals from Equations of State II: Large Hierarchies and False Vacua

arXiv:2606.30117v1 Announce Type: cross Abstract: We investigate the reconstruction of holographic duals for strongly coupled quantum field theories in regimes characterized by large hierarchies and t

researcharxiv-cs-ai

Tools

great vibes at SFMOMA Speakers Dinner today! if you’re at @aiDotEngineer, come join my talk on Thursday at 11:40. see you there!

Swyx announced a speaking engagement at an AI-focused event, highlighting a positive experience at an SFMOMA Speakers Dinner and inviting attendees to join their talk on Thursday at 11:40 AM. The post

toolsswyx--x

Ground Truths in Suicide Research: The Current State of AI-Based Suicide Detection in Social Media

arXiv:2606.28334v1 Announce Type: cross Abstract: Recent advances in artificial intelligence (AI) and social media data have led to growing optimism about the ability to detect suicide risk at scale.

researcharxiv-cs-ai

Ground4D: Consistency-Aware 4D Reconstruction from Monocular Video

arXiv:2606.28828v1 Announce Type: new Abstract: Learning a 4D scene representation from a single monocular video that supports dynamic novel-view synthesis while maintaining faithful geometry over tim

Tutorials

Grounding LLM Reasoning under Incomplete Graph Evidence

arXiv:2606.30247v1 Announce Type: new Abstract: Knowledge graphs can guide large language models (LLMs) reasoning, but the graph seen by a system is usually a retrieved, linked, temporally scoped, and

tutorialsarxiv-cs-cl

Grounding Sim-to-Real Generalization in Robotic Manipulation: An Empirical Study with Vision-Language-Action Models

arXiv:2603.22876v2 Announce Type: replace-cross Abstract: Learning a generalist control policy for robotic manipulation typically relies on large-scale datasets. Given the high cost of real-world data

GROW^2: Grounding Which and Where for Robot Tool Use

arXiv:2606.30632v1 Announce Type: cross Abstract: Can the robot use a plate to cut a cake if no knife is available? Tool use greatly expands robot capabilities, but to use tools creatively beyond thei

agentsarxiv-cs-ai

GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots

arXiv:2606.29705v1 Announce Type: new Abstract: Data, as the fundamental substrate of modern intelligence, has greatly driven the development of current foundation models. Naturally, researchers aim t

agentsarxiv-cs-ai

Guided Unconditional and Conditional Generative Models for Super-Resolution and Inference of Quasi-Geostrophic Turbulence

arXiv:2507.00719v3 Announce Type: replace-cross Abstract: Typically, numerical simulations of Earth systems are coarse, and Earth observations are sparse and gappy. We apply four generative diffusion

H-GRPO: Permutation-Invariant Reinforcement Learning for Grounded Visual Reasoning

arXiv:2606.29915v1 Announce Type: new Abstract: Vision-Language Models (VLMs) often achieve high performance on benchmarks while remaining 'black boxes', yet they remain prone to hallucination or rely

Had a lot of fun at @aiDotEngineer talking about BM25!

Had a lot of fun at @aiDotEngineer talking about BM25! Cool Search paradigm by @jobergum: “BM25 + Grep is all you need”: 1. Use BM25 to narrow down a large universe of documents to a few candidates 2.

agentsswyx--x

Half price Starlink for people in the Memphis region

Half price Starlink for people in the Memphis region SpaceXAI has announced that they will be applying a 50% discount on the standard @Starlink monthly price for both new and existing customers in the

industryelon-musk--x

harbor is a great framework for running evals for long running, stateful agents its becoming industry standard, powering benchmarks like ter…

harbor is a great framework for running evals for long running, stateful agents its becoming industry standard, powering benchmarks like terminal bench 2 we've integrated deeply with harbor: across La

.@harborframework can now integrate directly with Deep Agents, LangSmith Sandboxes, and LangSmith Observability. You need to run agents in a…

.@harborframework can now integrate directly with Deep Agents, LangSmith Sandboxes, and LangSmith Observability. You need to run agents in a real, reproducible, isolated environment, many times in par

Hard-constraint physics-residual networks for hydrogen crossover prediction and high-pressure extrapolation in PEM water electrolysis

arXiv:2511.05879v5 Announce Type: replace-cross Abstract: Hydrogen crossover is a critical safety and efficiency constraint in high-pressure polymer electrolyte membrane water electrolysis (PEMWE), bu

Hardware

HARD-KV: Head-Adaptive Regularization for Decoding-time KV Compression

arXiv:2606.28831v1 Announce Type: cross Abstract: Long-context LLM inference faces a fundamental conflict: head-adaptive compression algorithms (e.g., Top-p nucleus sampling) offer superior accuracy b

hardwarearxiv-cs-ai

Harvesting AI Computation at the Edge via Generic Approximation

arXiv:2606.29518v1 Announce Type: cross Abstract: With the widespread adoption of AI in various IoT scenarios such as smart sensing and processing, AI chips have become a common component at the edge.