AI News · 2026-07-09

💡

Jason Says

Grok 4.5 claims Opus-class, Lovable hits $13.2B, and China flags Claude Code security risks — the AI coding race is intensifying on both competition and geopolitical fronts. Developers, hedge your bets.

🛠️
AI ToolsTechCrunch AI

SpaceXAI Releases Grok 4.5 Opus-Class Model

SpaceXAI launched Grok 4.5, which Musk describes as an 'Opus-class model' offering cheaper and more efficient performance. Latent Space calls it the first flagship model post-Cursor acquisition, signaling SpaceXAI's aggressive pace in the frontier AI race.

🛠️
AI ToolsOpenAI Blog

OpenAI Introduces GPT-Live Voice Models

OpenAI launched GPT-Live, a new generation of voice models for natural human-AI interaction, now powering ChatGPT Voice. This marks a shift to end-to-end native voice understanding, promising lower latency and more natural conversational experiences for developers.

🔥
SkillsGitHub Trending

Addy Osmani Open-Sources Production-Grade Agent Skills

Addy Osmani open-sourced agent-skills, packaging senior engineering workflows and quality gates into reusable skills for AI coding agents across the full dev lifecycle. A significant addition to the growing Skills ecosystem for Claude Code and similar agents.

🛠️
AI ToolsGitHub Trending

Tencent Open-Sources 4-Tier Agent Memory System

Tencent open-sourced TencentDB Agent Memory, a fully local long-term memory system using a 4-tier progressive pipeline with zero external API dependencies. It compresses tool logs into compact Mermaid diagrams for short-term memory and layers long-term recall — a practical memory solution for stateful agents.

🛠️
AI ToolsOpenAI Blog

OpenAI Challenges SWE-Bench Pro Coding Benchmark Reliability

OpenAI published an analysis revealing reliability and accuracy issues in SWE-Bench Pro, a popular coding benchmark. For developers evaluating AI coding tools, this means single-benchmark rankings can be misleading — actual coding ability may be over- or under-estimated.

💰
MonetizationTechCrunch AI

Lovable in Talks to Double Valuation to $13.2B

AI app builder Lovable is reportedly raising $300M led by Menlo Ventures, doubling its valuation to $13.2B. The deal reflects the surging 'AI + No-Code' market, though it raises questions about whether current valuations are sustainable.

💰AI Funding Roundup

EdVisorly

Series A$13.3M

AI原生大学招生自动化平台,用AI替代人工审核流程,教育AI赛道新融资

深度智控(DeepCtrls)

Series B数亿元人民币

Investors: 晶科能源、国投创新、招银国际

物理AI引擎驱动能源基础设施自治控制,已实现L4/L5级AI闭环,正拓展海外市场

昇视唯盛(3Srobotics)

Series B数亿元人民币

Investors: 上海半导体产投、金桥基金

焊接具身智能机器人,自研焊接大脑模型+运动控制系统,切入工业制造AI赛道

Subscribe for daily AI updates + free playbook

📘 Subscribe Free