Skip to content
gentic.news — AI News Intelligence Platform
Connecting to the Living Graph…
gentic.news

Latest AI News

Every story, newest first — research, funding, product launches, policy, and analysis.

A Peking University researcher points to a phase-change memristor chip on a test board, with latency and speedup…
AI ResearchBreakthrough
100

PKU Chip Hits 2.12ms Brain Latency, 478x A100 Speedup

PKU chip achieves 2.12ms step latency with 478x speedup over Nvidia A100 for brain modeling using phase-change memristors.

pandaily.com/Jul 7, 2026/3 min read/Widely Reported
hardwareresearchneuromorphic computing
A sleek metallic robot hand reaches toward a glowing digital brain, symbolizing advanced AI competition
AI Research
84

GPT-4 Held Top Spot 52 Weeks; Today's Models Last 7

GPT-4 dominated the ECI for a year. Today's top models last 7 weeks median, with 17 leadership changes since Feb 2024.

the-decoder.com/Jul 6, 2026/3 min read
benchmarksai modelscompetition
A person's laptop screen displays Obsidian notes and a connected AI agent interface, with icons for Readwise…
AI Research
74

Second Brain Becomes Open-Source AI Agent Memory System

Pauliusztin open-sourced a 2-year project turning 10,994 notes into a living AI agent memory system. The architecture aligns with Google's new OKF standard.

x.com/Jul 6, 2026/3 min read
open sourceknowledge managementai agents
Alibaba researchers present MIPU framework diagram with a two-step RL process, showing a large language model being…
AI Research
85

Alibaba's MIPI fixes LLM training-inference mismatch with direct RL

Alibaba's MIPI directly optimizes inference policy, fixing the mismatch in LLM post-training via the MIPU framework.

x.com/Jul 6, 2026/3 min read
llm post-trainingreinforcement learningalibaba
Tencent's Hunyuan Hy3 model interface showing benchmark charts comparing agent task resolution rates against…
Products & Launches
98

Tencent Hunyuan Hy3: 295B MoE Hits 90% Agent Task Resolution

Tencent launched Hunyuan Hy3, a 295B MoE model with 21B active parameters, claiming 90% agent task resolution, surpassing DeepSeek V4 Pro and Qwen 3.7 Max.

pandaily.com/Jul 6, 2026/3 min read/Widely Reported
ai agentsai modelschina ai
A laptop screen displays a terminal window showing markdown output from a scanned document, with a printed page and…
Products & Launches
78

Ollama-OCR Turns Scanned Docs Into Markdown, No Cloud Needed

Ollama-OCR extracts text from scanned docs locally using Ollama vision models. 2.3k stars, no cloud APIs needed.

x.com/Jul 6, 2026/3 min read
open-sourcedeveloper-toolsvision-models
Biren Technology executive standing beside a large GPU chip display, with server racks and blue lighting in the…
Funding & Business
100

Biren Raises $893M to Ramp GPU Production, Challenge Nvidia in China

Biren raises $893M at a discount to fund GPU production and challenge Nvidia in China's AI chip market.

scmp.com/Jul 6, 2026/3 min read/Widely Reported
chinafundingnvidia
A high-resolution photomosaic composed of hundreds of small tile photos forming a larger image, likely a landscape…
AI Research
85

PhotoQuilt Makes Training-Free Photomosaics at 14K Resolution

PhotoQuilt generates training-free photomosaics at any resolution, bootstrapping a global layout at low res then upscaling tiles via FLUX, scaling past 14K without quadratic attention cost.

x.com/Jul 5, 2026/3 min read
image-generationdiffusion-modelsgenerative-ai
AI researchers discuss a chart comparing 35B parameter agent performance to trillion-parameter models at a conference
AI Research
85

Hugging Face Papers: 35B Agent Matches Trillion-Parameter Performance

Hugging Face Daily Papers featured eight AI papers, including Orca (world model), Dockerless (62% SWE-bench), and a 35B agent matching trillion-parameter performance.

x.com/Jul 5, 2026/3 min read
efficiencyworld modelsscaling
Robot navigating through a room with furniture, using sensors and a screen displaying a floor plan
AI Research
100

Alibaba's Qwen-RobotNav Unifies Robot Navigation in One 2B-8B Model

Alibaba's Qwen-RobotNav unifies VLN, ObjectNav, tracking, and autonomous driving in a 2B-8B model, deploying zero-shot to quadruped robots via a configurable observation protocol.

x.com/Jul 5, 2026/3 min read/Multi-Source
roboticsnavigationai models
ByteDance and Alibaba AI agents disabled by China's July 15 ban, affecting Doubao and Qwen services
Policy & Ethics
98

China's July 15 AI Agent Ban Hits ByteDance, Alibaba

China's July 15 anthropomorphic AI rules force ByteDance and Alibaba to disable custom agents, exempting utility bots while targeting emotional companions.

scmp.com/Jul 5, 2026/3 min read/Widely Reported
consumer-aichinaregulation
A sleek, futuristic workshop space with a large digital screen displaying Fable 5 model capability graphs and team…
Products & Launches
100

Anthropic's Fable 5 gets production workshop series from @_vmlops

Anthropic's Fable 5 gets production workshop series from @_vmlops covering capability curves, reliable agents, and deployment at scale.

x.com/Jul 5, 2026/3 min read/Widely Reported
anthropicproductionai agents
CMU researchers demonstrate Gym-Anything converting a spreadsheet interface into a simulated training ground for AI…
AI Research
92

CMU's Gym-Anything Turns Any Software Into Agent Training Ground

CMU's Gym-Anything automates agent environment creation, producing CUA-World with 10,000+ tasks. Even strong models fail most long tasks, showing real computer-use work is unsolved.

x.com/Jul 4, 2026/3 min read
cmuresearchai agents
Alibaba Blocks Claude Code After Anthropic’s Tracking Experiment
Products & Launches
92

Alibaba Blocks Claude Code After Anthropic’s Tracking Experiment

Alibaba blocked Claude Code after Anthropic’s tracking experiment angered Chinese developers and security staff, highlighting data sovereignty tensions.

x.com/Jul 4, 2026/3 min read/Widely Reported
chinaai regulationenterprise ai
Diagram comparing Tencent Hunyuan GEAR's dual read-out architecture to LlamaGen-REPA, with speed and quality metrics
AI Research
85

Tencent Hunyuan GEAR: 10× Faster Autoregressive Image Gen

Tencent Hunyuan's GEAR jointly trains VQ tokenizers and AR generators end-to-end, achieving 10× faster autoregressive image generation while outperforming LlamaGen-REPA.

x.com/Jul 4, 2026/3 min read
image-generationtokenizerstencent
Scientists in lab coats examine molecular structures on a large screen, with a robotic arm handling a petri dish…
StartupsBreakthrough
100

Anthropic Launches In-House Drug Discovery for Neglected Diseases

Anthropic launched drug discovery for neglected diseases. Novartis CEO says AI could cut development from 12 to 7-8 years and double success rates.

the-decoder.com/Jul 4, 2026/3 min read/Multi-Source
ai in healthcareclaudeanthropic
Mistral AI logo on a dark gradient background with abstract geometric patterns representing AI and mathematics
Open SourceBreakthrough
100

Mistral's Leanstral 1.5 hits 100% on miniF2F, finds 5 real bugs

Mistral's Leanstral 1.5 scores 100% on miniF2F, solves 587 Putnam problems, and finds 5 real bugs in open-source code.

the-decoder.com/Jul 4, 2026/3 min read/Widely Reported
code-analysisformal-verificationopen-source
CEO Sam Altman stands at a podium with the OpenAI logo behind him, addressing a room of policymakers and journalists…
Products & Launches
85

OpenAI Offers Washington 5% of $852B Valuation to Ease AI Pressure

OpenAI proposed 5% of its $852B business to Washington to ease AI regulatory pressure, per @rohanpaul_ai. The equity-for-peace swap could set a precedent.

x.com/Jul 3, 2026/3 min read
ai policyfundingregulation
ByteDance Seed AI researchers present a graph showing AI agent learning speed doubling quarterly, with data points…
AI ResearchBreakthrough
100

ByteDance Finds AI Agents Double Learning Speed Every 3 Months

ByteDance's Seed AI team discovered that AI agents double learning speed every three months via real-world interaction, per a Thursday paper. EdgeBench benchmark with 134 tasks ≥12 hours each underpins the finding.

scmp.com/Jul 3, 2026/3 min read/Widely Reported
benchmarkingbytedancescaling laws
A line graph showing a sharp 3.5x spike in high-severity CVEs in June, following Anthropic's April announcement of…
Products & LaunchesBreakthrough
100

CVEs spike 3.5x after Anthropic's Mythos Preview launch

High-severity CVEs jumped 3.5x in June after Anthropic's Mythos Preview launch. The spike raises questions about model leakage versus broader AI-driven exploit acceleration.

epoch.ai/Jul 3, 2026/3 min read/Widely Reported
vulnerability-discoveryanthropiccybersecurity
A bar chart comparing AI agent performance across seven benchmarks, with fixed compute budgets shown as lower bars…
Policy & Ethics
96

AISI: Fixed compute budgets underestimate AI agents by 60%

AISI found standard benchmarks cap compute budgets, underestimating agent capabilities by ~60%. Success rates jumped ~25% with 10x tokens.

the-decoder.com/Jul 3, 2026/3 min read/Widely Reported
ai safetybenchmarkingai agents
A sleek AI interface displaying a crystal lattice structure on a monitor, with a researcher in a lab coat pointing…
AI ResearchBreakthrough
100

Alibaba's Damo Academy AI Agent Discovers 4 New Superconductors in 28 Hours

Alibaba's Damo Academy unveiled Elements Claw, a 1B-parameter AI agent that discovered 4 new superconductors by screening 2.4M crystal structures in 28 GPU hours.

scmp.com/Jul 3, 2026/3 min read/Widely Reported
materials sciencescientific discoveryai agents
AI robot arm hesitating before a pile of money and server racks, illustrating financial AI model fine-tuning
AI Research
76

Bridgewater, Murati's startup fine-tune Qwen3 to 84.7% on finance tests

Bridgewater and Thinking Machines Lab fine-tuned Qwen3-235B to 84.7% accuracy on financial tasks, beating GPT/Gemini/Claude at 1/14th cost.

the-decoder.com/Jul 3, 2026/3 min read
financefine-tuningai models
Mira Murati presents a chart comparing Thinking Machines' error rate and inference cost against frontier models…
AI Research
100

Mira Murati's Thinking Machines beats frontier models by 29.8% with Bridgewater's expert judgments

Thinking Machines beat frontier models by 29.8% fewer errors using Bridgewater's expert judgments, at 13.8x lower inference cost.

x.com/Jul 3, 2026/3 min read/Multi-Source
financeenterprise-aistartups
AI Security Inst Shows Test-Time Compute Skews Frontier Evaluations
AI Research
92

AI Security Inst Shows Test-Time Compute Skews Frontier Evaluations

AISecInst research shows test-time compute budgets skew frontier model evaluations, challenging standard practices.

x.com/Jul 3, 2026/3 min read/Multi-Source
ai safetyai researchmodel evaluation
A robotic arm in a lab setting grasps a transparent plastic cup on a table, with a monitor displaying code and…
AI Research
85

DART: One-Shot Robot Adaptation via Weight Space Arithmetic

DART from Seoul National University adapts robot policies with one demonstration using weight space arithmetic, achieving 73% success on unseen domain shifts.

x.com/Jul 3, 2026/3 min read
domain adaptationweight space arithmeticrobot learning
Two executives shake hands beside a glowing microchip prototype, with Samsung and Anthropic logos on a presentation…
Big Tech
88

Anthropic Explores Custom AI Chip with Samsung

Anthropic is discussing a custom AI chip with Samsung, per The Information. The move follows OpenAI's Jalapeño chip and signals growing vertical integration in AI hardware.

techcrunch.com/Jul 2, 2026/3 min read
anthropicsamsungai hardware
A diagram showing ELDR routing decode requests to GPU experts using prefill signatures, with arrows indicating…
AI Research
85

ELDR: Expert-Locality Decode Routing Cuts MoE TPOT by 13.9%

ELDR uses prefill expert signatures to route decode requests, cutting median TPOT by 5.9–13.9% in vLLM at scale.

x.com/Jul 2, 2026/3 min read
ai infrastructurellm servingmoe
A laptop screen displays a video editing interface with AI-powered natural language commands, replacing traditional…
Products & Launches
77

Browser-use open-sources Claude-powered video editor

Browser-use open-sourced a video editor inside Claude Computer Use, replacing manual editing with natural language commands and challenging Adobe Premiere.

x.com/Jul 2, 2026/3 min read
open sourcevideo editingai tools
3D scene decomposition showing token groups representing distinct objects, with reconstruction and segmentation…
AI Research
87

Feed-Forward Model Decomposes 3D Scenes as Objects Without 3D Labels

A feed-forward model decomposes 3D scenes into objects from unposed images without 3D annotations, enabling one-pass reconstruction, segmentation, and manipulation.

x.com/Jul 2, 2026/3 min read
computer visionscene understanding3d vision