lithos Twitter
Lithos Header
Last Updated
Age in hrs 
1
2
3
5
8
13
21
34
55

 Text-Vision Co-Instructed Image Editing

 Learning from the Self-future: On-policy Self-distillation for dLLMs

 Rethinking the Role of Efficient Attention in Hybrid Architectures

 Looped World Models

 MotionVLA: Vision-Language-Action Model for Humanoid Motion

 Aligning Quantum Operators with Large Language Models

 LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling

 Unified Multimodal Autoregressive Modeling with Shared Context-Visual Tokenizer is Key to Unification

 ProCUA-SFT Technical Report

 Beyond Monolingual Deep Research: Evaluating Agents and Retrievers with Cross-Lingual BrowseComp-Plus

 Variable-Width Transformers

 ActWorld: From Explorable to Interactive World Model via Action-Aware Memory

 Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients

 A Gradient Perspective on RLVR Stability and Winner Advantage Policy Optimization

 ChLogic: Evaluating Robustness of Logical Reasoning in Chinese Expressions

 LectūraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching

 TRIAGE: Dialectical Reasoning for Explainable Risk Prediction on Irregularly Sampled Medical Time Series with LLMs

 Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion

 ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining

 OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillation

 GameCraft-Bench: Can Agents Build Playable Games End-to-End in a Real Game Engine?

 Topics: AI • US-China • Why Do (Some) Chinese AI Labs Distill?  - And why the Mythos/Fable drama doesn’t matter here

 Insurers pivot AI strategy toward core risk underwriting  - AI investments by insurers are now expected to generate tangible business value beyond mere efficiency. According to findings in the 2026 Evident AI Index, insurers are now embedding AI technologies into workflows that directly influence underwriting discipline and capital allocation. Christian Preece,...

 HPE AI Factory With NVIDIA Expands for the Era of Agents  - Enterprises are moving agentic AI from proof of concept to production — and the next generation of AI factories are built for the era of agents. At HPE Discover Las Vegas, running through Thursday, June 18, NVIDIA and HPE are expanding the HPE AI Factory with NVIDIA, including NVIDIA Vera CPU and NVIDIA...

 AMD Snaps MEXT to Break the Memory Wall  - AMD acquires MEXT to slash AI memory costs and break the memory wall. The post AMD Snaps MEXT to Break the Memory Wall appeared first on EE Times .

 No breakthrough in Anthropic talks with Trump administration  - Officials with artificial intelligence startup Anthropic met with the Trump administration Monday as it seeks to restore models blocked over national security concerns, a spokesperson for the company said.

 Epic Errands

 Building the NPC Generator

 🌿 Elysium: Building a Persistent Agentic Civilization on a Bioluminescent Hypergraph

 Field Notes: building a shop-floor AI on a small local model

 Teaching a 4B Model to Run a City-Wide Scavenger Hunt — Without Naming a Single Street

 TropicalGT and ToricGT: Tropical, Toric, and Graph-Token Geometry for Auditable Reasoning

 Building an AI Interview Coach for the BuildSmall Hackathon 2026

 AI Authorship Signals Are Review Prompts, Not Verdicts

 FitCheck: What AI can your computer actually run?

 Professor Pip - a tiny teacher who lives in your browser

 🧬 Carbon-VEPor: Efficient Variant Effect Prediction with Carbon

 DiffSense: A Local-First Pull Request Reviewer Built During Build Small

 Teaching an LLM to Decompile the Nintendo 64

 From "Play My Workout Playlist" to a Real Android Tap Plan

 Ink-Wash Universe: A Developer’s Chronicle of Building a Local LLM Traditional Liuyao Divination App (FateForge) at a Hackathon

 Small model, big speech: a speech coach under 32B parameters

 I Built an AI That Finished the Song I Quit On Ten Years Ago

 6ixPulse: piloting agentic neighbourhood research, starting with Toronto

 Slipstream: a benchmark for mid-flight project forecasting, and a small agent that blends the best methods

 GrandVoice: A Multilingual Voice Companion for Grandparents, Kids, and Families

 Open to Work: The Musical — distilling a 20B lyricist into a 0.5B model that runs on CPU

 NIGHTWAVE — a whole all-night radio station run by a one-billion-parameter brain

 StoryGrove -- Field Notes

 The Mountain Called Unnecessary

 I built a world you can talk into existence

 Sozai — Build Memories Together

 I fine-tuned a model for free from one prompt, with TRL and the Google Colab CLI

 Floodlight Venue Manager Running Notes

 Field Notes: building a machine that interrogates you

 Moku: The Word That Wouldn't Stay Put

 Viva Mais AI: 1B models for one person

 Building "YourGymBuddy" – A Race Against Time, Modal Fine-Tuning, and the Power of Local AI

 Chhaya: teaching a 4B model to be the eyes, not the doctor

 How I Built AI Picture Book

 Anim Vid AI: Turning Any Concept Into a Manim Animation With Fine-Tuned Small Models

 Constellation of Hope - Build Small

 Introducing Clawd: Solana-Native AI Models for On-Chain Builders

 Meet Puck

 I built my mom a budget app on two tiny MiniCPM models — here's what bit me

 🌾 KrishiSathi — From Farmer Problems to AI-Powered Agricultural Intelligence

 Quantization is the safety blind spot of local AI — so I built the screen for it

 The Office Meets Silicon Valley

 The Shape of Words — Field Notes

 Local Privacy Filtering for Shareable Agent Traces

 How We Built OpenMythos: A Cybersecurity LLM Trained from Scratch

 Introducing Clawd: Solana-Native AI Models for On-Chain Builders

 Field Notes: The Apprentice

 Interview Coach: Real-Time, Local Interview Assistance Without Taking Over the Answer

 How Pruna Optimizes Models: Tracing the Smash Function and DAG Execution

 Bridging the gap between neuromorphic ionic computing and more efficient AI  - The human brain is the ultimate supercomputer. It uses a highly branched and interconnected network of neurons and synapses to achieve massive computational power with extreme efficiency. In the age of AI, the brain, a paradigm ...

 Nebius Cloud Logs are now available in Datadog: Trace AI incidents across every layer  - Nebius AI Cloud Logs now stream into Datadog Log Management. If you already run on Datadog, you can investigate your Nebius workloads alongside the rest of your stack and trace an AI incident across every layer without switching tools.

 US clampdown on Anthropic models sends EU sovereignty surge into overdrive  - Brussels says access curbs prove Europe needs greater technological independence

 The health system CEO imperative: Turning AI’s promise into performance  - Health systems aren’t short on AI, they’re short on impact from AI. Capturing real value requires CEO-led transformation and a fundamental rethink of how care and operations run.

 AI in research: Mapping the terrain of adoption, attitudes and impact in the UK  - RAND Europe is administering, analysing and reporting on the first ‘National AI in Research survey’ in the UK.

 How we leveled up a 3D demo into a playable game  - What started as a showcase for Leonardo's 3D generation capabilities powered by Rodin V2 quickly evolved into a playable browser game. By The Leonardo team

 Japan's tech business SoftBank rolls out OpenAI 'patches' against cyberattacks  - Japanese technology giant SoftBank Group Corp. is launching a service using OpenAI technology to protect against the looming threat of cyberattacks, both companies said Tuesday.

 EU publishes its AI content labelling playbook ahead of the AI Act’s August deadline  - The European Union has published its AI content labelling playbook, a voluntary Code of Practice meant to help companies meet transparency rules that become law across the bloc on August 2 onwards. The European Commission released the final Code on 10 June, setting out practical steps for the businesses...

 Glean named 2026 ISV AI Visionary Partner of the Year by Databricks  - Glean was named the 2026 Databricks Partner Award winner: ISV AI Visionary Partner of the Year. See how Glean and Databricks connect governed data with enterprise context to help teams get trusted answers and move faster.

 We’re strengthening our presence in Alabama through new investments and community support.  - Google has announced a $1.5 billion investment for 2026 and 2027 to expand its data center campus in Jackson County, Alabama. Operating since 2019 on a repurposed former…

 Accenture: Consumers show growing trust in AI shopping agents  - Consumers are showing a willingness to let AI agents take on more shopping-related tasks, according to new research from Accenture. The company’s 2026 Consumer Pulse Research, based on a survey of 25,590 consumers across 16 countries, found that 74% of respondents would trust a personal AI agent more...

 LegalHalluLens: Typed Hallucination Auditing and Calibrated Multi-Agent Debate for Trustworthy Legal AI  - Subjects: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG); Multiagent Systems (cs.MA)

 Learning Cardiac Electrophysiology Digital Twins Through Agentic Discovery of Hybrid Structure

 EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation

 Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers

 The Stanford EDGAR Filings Dataset: Reconstructing U.S. Corporate and Financial Disclosures into Layout-Faithful and Token-Efficient Pretraining Data

 DRFLOW: A Deep Research Benchmark for Personalized Workflow Prediction

 WEQA: Wearable hEalth Question Answering with Query-Adaptive Agentic Reasoning

 Memory as a Wasting Asset: Pricing Flash Endurance for Embodied Agents, and the Limits of Doing So

 Your AI Travel Agent Would Book You a Bullfight: An Agentic Benchmark for Implicit Animal Welfare in Frontier AI Models

 Knowledge Reutilization in Meta-Reinforcement Learning

 First Proof Second Batch

 Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding

 IsabeLLM: Automated Theorem Proving Applied to Formally Verifying Consensus

 A Unified Framework for Context-Aware and Relation-Aware Graph Retrieval-Augmented Generation

 Agentic AI-based Framework for Mitigating Premature Diagnostic Handoff and Silent Hallucination in Healthcare Applications

 PseudoBench: Measuring How Agentic Auto-Research Fuels Pseudoscience

 ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

 LLM Consumer Behavior Theory: Foundations of a Novel Research Field

 STAR: SpatioTemporal Adaptive Reward Allocation for Text-to-Image RL Post-Training

 MoCo-AIS: A Contrastive Learning Framework for Similarity Computation of Vessel Trajectories

 MLPerf® Training 6.0: Leading NVIDIA HGX B300 and competitive NVIDIA GB300 NVL72 results on NVIDIA Blackwell Ultra systems  - MLPerf® Training 6.0 results are in. Across six configurations on NVIDIA Blackwell Ultra systems, Nebius posted the #1 single-node HGX B300 times for Llama-3.1-8B and GPT-OSS 20B pre-training and came within 3.1% of the fastest GB300 NVL72 result on all three benchmarks at 72 GPUs. This post covers the...

 Prompt Hungary: The impact of AI on the competitiveness of the economy  - AI could become a powerful growth engine for the country by helping overcome the economy’s challenges and limitations, and generating improved productivity and lasting competitive advantage.

 Non-x86 servers now nearly half the market, IDC says  - Demand for AI systems plus the shortage of DRAM and NAND are shaping the global market

 How customer teams became agent builders  - The people closest to your customers have the best context for improving the experience. Now they're the ones building the agents.

 LiteLLM Vulnerability Chain Lets Low-Privilege Users Take Over AI Gateway Servers  - A default low-privilege account on a LiteLLM proxy can climb to full admin and run code on the server by chaining three vulnerabilities, researchers at Obsidian Security disclosed LiteLLM is a widely deployed open-source AI gateway that brokers calls to more than 100 model providers behind one OpenAI-compatible...

 Verifiable Audit Trails for AI-Enabled Biological Design Tools: A Proposal for Verifiable Biodesign Logging  - The authors propose a tamper-evident audit-trail framework using cryptographic signing to establish a trustworthy chain of custody that can be used with any artificial intelligence–enabled biological design tool.


 [EOT] @lithos_graphein2026-Jun-17 05:20:54