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.

Alibaba T-Head engineer demonstrates SAIL stack on a server at WAIC Shanghai, with a screen displaying code and a…
Open SourceBreakthrough
82

Alibaba Open-Sources SAIL Stack to Break Nvidia CUDA Lock-In

Alibaba T-Head open-sourced SAIL stack for Zhenwu chips at WAIC, targeting Nvidia CUDA dominance with 7-day migration claim.

scmp.com/4h ago/3 min read
open sourceai infrastructurenvidia competitors
A 3D-rendered cartoon otter swimming in a digital ocean, created by AI coding agent Codex using Blender animation…
Products & Launches
82

Codex Computer Use Generates Blender Animation From Scratch

Codex installed Blender and made a 3D otter animation from a single prompt, needing one human click.

x.com/11h ago/3 min read
agentic aicomputer-usecodex
Diagram showing LongStraw's branch replay method on 8 H20 GPUs reaching 2.1M token positions for RL post-training
AI Research
87

LongStraw Reaches 2.1M Tokens on 8 H20 GPUs via Branch Replay

LongStraw reaches 2.1M token positions for RL post-training on 8 H20 GPUs by replaying short response branches, cutting compute 8-16x vs prior art.

x.com/18h ago/3 min read
efficient traininglong contextreinforcement learning

Microsoft Merges AutoGen and Semantic Kernel into Agent F…

Products & Launches
85

Microsoft Merges AutoGen and Semantic Kernel into Agent Framework

Microsoft merged AutoGen and Semantic Kernel into Agent Framework, a unified production-grade framework for .NET and Python with graph-based workflows and Foundry deployment.

x.com/1d ago/3 min read
ai-infrastructuremicrosoftagent-frameworks
A futuristic tech interface shows a glowing AI chip labeled 'Kimi K3' with data streams and numbers like 2.8T and…
AI Research
100

Moonshot AI's Kimi K3: 2.8T params, 1M token window, $3/M input

Moonshot AI released Kimi K3, a 2.8T-parameter mixture-of-experts model with 1M token context window and $3/M input pricing, claiming autonomous chip design and research capabilities.

x.com/1d ago/3 min read/Widely Reported
long contextmoonshot aiai models
NVIDIA's #1 RTEB Embedding Model Skips Token Generation Entirely
AI Research
81

NVIDIA's #1 RTEB Embedding Model Skips Token Generation Entirely

NVIDIA's RTEB hit #1 on MTEB by skipping token generation. The cheapest reasoning token is the unused one.

x.com/1d ago/3 min read
ai-modelsbenchmarksnvidia
A line graph showing a steep upward curve quickly reaching a flat ceiling, with a person pointing at the saturation…
Opinion & Analysis
75

gdb: Benchmarks Saturate Too Fast for Reliable AI Progress Tracking

@gdb notes benchmarks saturate quickly. This undermines AI progress tracking and may force shift to dynamic evaluations.

x.com/1d ago/3 min read
industry-analysisanthropicbenchmarks
TSMC Chairman CC Wei speaking at a podium, Arizona semiconductor fabrication facility construction visible behind…
Funding & BusinessBreakthrough
100

TSMC Pledges Extra $100B for Arizona, Total US Investment Hits $265B

TSMC pledged $100B more for Arizona fabs, total US investment $265B, driven by AI chip demand. Net income surged 77% to $22B in Q2.

scmp.com/2d ago/3 min read/Widely Reported
manufacturingai chipsinvestment
Mira Murati stands beside a large server rack with glowing blue lights, holding a tablet displaying a chart labeled…
StartupsBreakthrough
100

Murati's Thinking Machines Ships 975B Inkling — Leads US Open Models

Murati's Thinking Machines releases Inkling, a 975B-parameter MoE model that leads US open models but trails Chinese rivals on benchmarks and cost.

the-decoder.com/2d ago/3 min read/Multi-Source
open-sourcestartupsbenchmarks
A close-up of a sleek Nvidia Rubin chip on a circuit board, surrounded by robotic arms and glowing blue lights
Policy & EthicsBreakthrough
100

Japan to Buy 27,500 Nvidia Rubin Chips for Robot AI

Japan to buy 27,500 Nvidia Rubin chips for sovereign robot AI model, signaling strategic push in embodied intelligence.

bloomberg.com/2d ago/3 min read/Widely Reported
japanroboticssovereign ai
A large neural network model diagram with 1T parameters, showing reinforcement learning pipeline steps and emergent…
AI Research
99

Ring-Zero Trains 1T-Parameter Model via Reinforcement Learning

Ring-Zero scales RL with verifiable rewards to 1T parameters, revealing emergent reasoning like self-verification and context anxiety.

x.com/2d ago/3 min read
reinforcement learninglarge language modelsai research
Cursor Doubles Model Usage on All Plans, Adds Grok 4.5
Products & Launches
75

Cursor Doubles Model Usage on All Plans, Adds Grok 4.5

Cursor doubled included model usage on all plans, adding Grok 4.5 and Composer 2.5 without price changes, pressuring competitors like GitHub Copilot.

x.com/2d ago/3 min read
developer experiencegenerative aiai tools
A bar chart comparing SWE-Bench scores for 7B to 14B models, showing a +5.4 improvement with function-aware FIM…
AI Research
85

Function-Aware Fill-in-the-Middle Boosts SWE-Bench by +5.4 on 14B Models

Function-aware FIM mid-training boosts SWE-Bench by +2.8 to +5.4 on 7B-14B models, preserving general abilities. Six checkpoints and 400K dataset open-sourced.

x.com/2d ago/3 min read
coding agentsbenchmarksmodel training
A developer types code in an IDE while a Claude Code artifact dashboard displays live data from MCP connectors, with…
Products & Launches
93

Claude Code Artifacts Now Call MCP Connectors for Live Data

Claude Code artifacts now call MCP connectors, enabling live data fetching and actions in generated dashboards and apps, announced via tweet.

x.com/2d ago/3 min read/Multi-Source
claude codeanthropicmcp
A glowing red robotic hand reaches toward a blue digital brain, symbolizing AI security testing, with data streams…
AI ResearchBreakthrough
91

GPT-Red: OpenAI's LLM Super-Hacker Finds 84% of Attacks, Humans 13%

OpenAI's GPT-Red LLM finds 84% of attacks vs 13% for humans, hardening GPT-5.6 Sol. Automated red-teaming shifts safety paradigm.

technologyreview.com/2d ago/3 min read
llm securityai safetyred-teaming
Researchers examine a synthetic document benchmark on a screen, analyzing long-context visual understanding and…
AI Research
85

ServiceNow's SynthDocBench Teases Apart VLM Long-Context Failure Modes

ServiceNow releases SynthDocBench, a controlled synthetic benchmark for long-context visual document understanding that varies length, layout, modality, and reasoning to diagnose VLM failures.

x.com/2d ago/3 min read
ai benchmarksservicenowvision-language models
Two businesspeople shaking hands in a modern office, symbolizing a partnership for deploying AI systems in enterprises
Opinion & Analysis
100

Anthropic, Blackstone Launch $1.5B AI Implementation Venture Ode

Anthropic and Blackstone launched Ode, a $1.5B AI implementation venture, embedding engineers in enterprises. It mirrors OpenAI's The Deployment Company, signaling a shift from model sales to services.

techcrunch.com/3d ago/3 min read/Widely Reported
servicesenterprise-aianthropic
OpenAI Staff Donate $215K to Super PAC Opposing Brockman's $50M Fund
Policy & EthicsBreakthrough
82

OpenAI Staff Donate $215K to Super PAC Opposing Brockman's $50M Fund

OpenAI employees donated over $215,000 to Guardrails Alliance, a super PAC opposing Greg Brockman's $50M pro-industry fund, highlighting internal tensions over AI regulation.

wired.com/3d ago/3 min read
ai policyopenaipolitics
A Xiaomi humanoid robot with articulated arms stands in a bright lab, surrounded by glowing holographic screens…
Open SourceBreakthrough
100

Xiaomi Open-Sources 38B Robotics-U0 Unifying Four Embodied Tasks

Xiaomi open-sourced 38B-parameter Robotics-U0, unifying four embodied tasks in a single model. No benchmarks or training data disclosed yet.

pandaily.com/3d ago/3 min read/Widely Reported
open-sourceroboticsxiaomi
A person holds a smartphone displaying a vivid AI-generated landscape, with text overlay reading 'SpectraReward' and…
AI Research
85

ByteDance SpectraReward: Training-Free Reward Reads Prompt Back From Image

ByteDance Seed releases SpectraReward, a training-free reward that reads a prompt back from a generated image using prompt log-likelihood. No training or preference labels needed.

x.com/3d ago/3 min read
ai alignmentbytedancetext-to-image
A large red whale swimming through digital data streams, symbolizing DeepSeek's massive scale and growth
Funding & Business
100

DeepSeek seeks fresh $71B round weeks after $7B close

DeepSeek seeks $71B valuation round weeks after $7B close. Capital for data centers and custom chips to sustain 11x cheaper pricing than GPT-5.5.

the-decoder.com/3d ago/3 min read/Widely Reported
ai infrastructurechinese aifunding
Smartphone screen showing ChatGPT chat interface inside WhatsApp with a welcome message for European users
Policy & EthicsBreakthrough
89

ChatGPT returns to WhatsApp in EU after Meta forced to open platform

OpenAI re-enabled ChatGPT on WhatsApp in the EEA after EU forced Meta to open its platform. Users reach GPT-5.5 via 1-800-CHATGPT.

the-decoder.com/4d ago/3 min read
messagingregulationai assistants
Close-up of a glossy black 14nm AI chip with intricate circuit patterns, labeled 'China AI Processor 520 TFLOPS'…
Products & LaunchesBreakthrough
100

China's 14nm AI Chip Hits 520 TFLOPS Via Architecture, Not Shrink

China's 14nm AI chip claims 520 TFLOPS and 6.4TB/s bandwidth via software-defined and 3D near-memory architecture, bypassing advanced node restrictions.

pandaily.com/4d ago/3 min read/Widely Reported
chinacomputer architectureai hardware
Three OpenAI GPT-5.6 model tiers—Sol, Terra, Luna—launch on Amazon Bedrock, with pricing matching first-party API access
Products & LaunchesBreakthrough
100

OpenAI GPT-5.6 Sol, Terra, Luna Launch on Bedrock at Same Price

OpenAI's GPT-5.6 Sol, Terra, and Luna launch on Amazon Bedrock at matching first-party pricing. Sol scores 80 on Coding Agent Index.

aws.amazon.com/4d ago/3 min read/Widely Reported
amazongpt-5.6openai
A white Google-branded delivery robot rolls along a city sidewalk past a brick building, its cylindrical body topped…
Opinion & Analysis
75

Google alone ships full any-to-any multimodal models

Mollick notes Google alone ships full any-to-any multimodal models; OpenAI and Anthropic lag. This gives Google a structural advantage in agentic workflows.

x.com/4d ago/3 min read
anthropicmultimodalgoogle
Two men in a modern office, one gesturing toward a whiteboard with AI agent diagrams, the other smiling while…
AI ResearchBreakthrough
100

Rich Sutton Launches Oak Lab to Build Self-Learning AI Agents

Rich Sutton founded Oak Lab to build self-learning AI agents. He rejects static datasets for real-time reinforcement learning with a trillion-parameter goal at 20W.

the-decoder.com/4d ago/3 min read/Multi-Source
startupsreinforcement learningai agents
Bar chart showing open-weight model token share rising from 11% to 29%, with Vercel CEO Guillermo Rauch presenting…
Products & Launches
80

Open-weight models now run 29% of gateway tokens, up from 11% in April

Open-weight models now handle 29% of gateway tokens, up from 11% in April. The 18-point jump signals accelerating enterprise adoption of open architectures like Llama 3 and Mistral.

x.com/4d ago/3 min read
open-sourceenterpriseinference
A laptop with 25GB RAM running a 744-billion-parameter AI model, no GPU, displaying inference output on a dark screen
AI Research
85

Colibri Runs 744B-Parameter Model on 25GB RAM, No GPU

Colibri claims to run a 744B-parameter model on 25GB RAM without GPU, but lacks evidence. If true, it could democratize large-model inference.

x.com/4d ago/3 min read
quantizationmodel compressioninference
Three professionals in business attire sit at a conference table, one pointing to a large screen displaying…
Policy & EthicsBreakthrough
84

200+ economists warn AI could surpass Industrial Revolution, offer no plan

200+ economists including 16 Nobel laureates signed a statement warning AI could transform economy faster than Industrial Revolution, but proposed no specific policies.

the-decoder.com/4d ago/3 min read
labor marketeconomicsai policy
Professor Yuji Tachikawa at a whiteboard, with Claude Fable 5 displayed on a laptop, highlighting a calculation…
AI Research
100

Claude Fable 5 Solves String Theory Problem Stalled for Six Months

Claude Fable 5 solved a string theory problem stalled for six months. Professor Yuji Tachikawa says the model made a non-trivial observation and used SymPy for verification.

x.com/5d ago/3 min read/Multi-Source
claudeanthropicphysics