Geo News
Community curated by people like you
LatestAICryptoHealthWorld AffairsUS Politics

Open source AI ecosystems & communities stories

Sep 15, 2026

Salesforce and Nvidia launch Koa reasoning model

Salesforce announced Koa, its first reasoning AI model built on Nvidia's open-weight Llama architecture and specialized for CRM data. The model represents a partnership between the two companies at Dreamforce conference.

Sep 15, 2026·2 sources
00
Sep 3, 2026

OpenAI's Codex wins AI security race at Devcon, DeepSeek finishes second

Ten AI coding agents competed in Austin Griffith's security challenge at Devcon, attempting 12 Solidity challenges originally designed for human developers. Only OpenAI's Codex completed all challenges, while DeepSeek, an open-weight Chinese model, came closest among cheaper alternatives.

Sep 3, 2026·1 source
00
Aug 30, 2026

Debian votes to allow AI-generated code in Linux distribution

The Debian community has voted to adopt a new AI policy that permits contributors to use generative AI tools in developing code for the Linux distribution. The policy neither endorses nor prohibits AI use, with disclosure optional but quality standards mandatory.

Aug 30, 2026·2 sources
00
Aug 27, 2026

Nvidia agrees to acquire Hugging Face for $12.9 billion

Nvidia has agreed to buy Hugging Face, the popular open-source AI platform and model repository, for $12.9 billion according to multiple reports. The acquisition would give the chip giant control of a strategic hub for open-source AI development.

Aug 27, 2026·9 sources
00
Aug 22, 2026

Mystery AI model Ox Alpha gains traction with free access despite unknown creator

A new AI model called Ox Alpha is gaining significant developer adoption after appearing online for free, though the identity of its creator remains unknown. Multiple outlets report developers are impressed with its capabilities.

Aug 22, 2026·2 sources
00
Aug 14, 2026

Chinese AI firm Z.ai releases GLM-5.3 coding model, claims discovery of 2,436 vulnerabilities

Z.ai (Zhipu AI) released GLM-5.3, a 743-billion-parameter open-weights AI model that the company claims is the most capable coding model available. The model reportedly discovered 2,436 vulnerabilities across 269 projects, including a serious flaw in the Cursor code editor and 1,097 critical bugs in Linux, WebKit, and FreeBSD.

Aug 14, 2026·6 sources
00
Aug 6, 2026

Bitcoin Red Team Identifies Nearly 5,000 Security Vulnerabilities in 30-Hour AI-Assisted Audit of Open-Source Projects

A volunteer group of 16 security researchers used AI tools to audit 390 Bitcoin-related open-source projects, discovering 4,962 potential vulnerabilities in approximately 30 hours, including 85 critical and 635 high-severity issues. The rapid security sweep was prompted by a recent $100 million Coldcard wallet exploit.

Aug 6, 2026·9 sources
00
Aug 5, 2026

Bitcoin Red Team files nearly 5,000 security findings across 390 projects in 30 hours

A 16-person volunteer group running AI models against Bitcoin codebases filed 4,962 potential security issues including 85 critical bugs across 390 open-source projects in approximately 27-30 hours.

Aug 5, 2026·6 sources
00
Jul 22, 2026

Nvidia CEO Jensen Huang defends Chinese AI models as U.S. weighs restrictions amid 'Kimi panic'

Nvidia CEO Jensen Huang publicly praised Chinese open-source AI models as 'excellent' and 'world class,' urging Washington not to ban them, even as Treasury Secretary Scott Bessent warned of scrutinizing Chinese models for stolen intellectual property. The comments come amid growing concern in Washington over low-cost Chinese AI models like Kimi that increasingly compete with American offerings.

Jul 22, 2026·8 sources
00
May 28, 2026

IBM and Red Hat Launch $5 Billion AI-Driven Open Source Security Initiative

IBM announced Project Lightwell, a $5 billion investment deploying 20,000 engineers to secure open source software using AI-driven tools and automation.

May 28, 2026·2 sources
00

Top claims

  • ▪In benchmark tests, Koa matched or exceeded the performance of multiple general-purpose models on CRM tasks while producing three times fewer errors
  • ▪Salesforce Inc. announced the Koa model at its annual Dreamforce tech conference on September 15, 2026
  • ▪Koa is hosted and run within Salesforce's trust boundaries via Nvidia's open framework, ensuring customer data does not leave the controlled environment during inference

People involved

Austin GriffithJensen HuangScott Bessent

Subtopics

AI security5AI for developers4AI coding assistants3AI foundation models2AI research & benchmarks2AI startups2AI tools & products2Bitcoin security & risks2AGI arms race & geopolitics1AI agents1AI antitrust & competition1AI copyright & training data1AI governance1AI open source risks1AI policy1AI Regulation1AI standards, audits & compliance1Bitcoin1Bitcoin wallets & custody1Business & enterprise AI1China1Compute, chips & AI infrastructure1Media mergers and acquisitions1Open source software security1

Related timelines

Congress

108 stories

Crypto hacks

100 stories

Iran War

209 stories

Payments

118 stories

Trump administration

116 stories