The AI agent ecosystem continues its rapid expansion. OpenClaw launched native iOS and Android companion apps, turning phones into nodes in a self-hosted AI agent gateway network — a major milestone for open-source mobile agents. Security concerns continue to shadow OpenClaw's growth, with fresh reports of malicious skills slipping into the ClawHub marketplace exposing supply chain vulnerabilities. Anthropic had a blockbuster day: the US government lifted export controls on Fable 5, clearing it for global distribution with a new safeguard blocking the reported jailbreak 99% of the time; the company also launched Claude Sonnet 5 as its "most agentic" mid-tier model (now default for Free and Pro users), and announced Claude Corps, an $85K paid 12-month AI fellowship. Meanwhile, xAI's Grok 4.5 is reportedly in private beta at SpaceX and Tesla, with a 1.5T parameter model expected in July. A sobering MIT Q&A explored what agentic AI actually is versus what we want it to be, and the BIS warned that massive AI capital spending could mirror history's worst economic bubbles.
OpenClaw released free companion apps for iPhone and Android, letting users run autonomous AI agent workflows directly from their phones. Each phone becomes a node in a self-hosted agent gateway network — not a chatbot, but a full agent that can take steps, make decisions, and complete tasks.
Attackers slipped malicious skills into OpenClaw's ClawHub marketplace, exposing a new supply chain attack vector where trusted AI agents can be manipulated to access sensitive systems, credentials, and enterprise workflows. The breach underscores growing security concerns as the agent ecosystem scales.
MIT News interviewed Phillip Isola (EECS/CSAIL) on the fundamentals and potential impacts of agentic AI, exploring the intelligence AI agents possess and the underlying models that power them. A thoughtful counterpoint to the hype cycle.
A 2026 analysis finds the vast majority of agentic AI pilots fail to reach production, with Gartner warning up to 60% of AI projects could be abandoned due to data infrastructure limitations. The gap between investment and data maturity is widening.
A detailed ranking of 18+ AI coding agents, including Augment Code (70.6% SWE-bench), Kilo Code, and others. The space is maturing fast with enterprise-focused tools leading on benchmarks.
Selected posts from the edition, with links to the originals.
"I expect @AnthropicAI to increasingly compete against their customers and distill their customers' data. Chinese models wipe out their margins, leaving no margins on the model layer itself. If Chinese models deliver similar results, Anthropic can't even compete against US"
View post on X ↗"sneaky, but also clever."
View post on X ↗"🚨BREAKING: The most powerful public AI model just came back from an 18-day government shutdown. Fable 5 returns globally on July 1 after Commerce lifted its export controls, redeployed with a new safeguard that blocks the reported jailbreak 99% of the time."
View post on X ↗"🚨 BREAKING: Anthropic will pay you $85,000 to learn AI. It's called Claude Corps. Anthropic just launched it, and it's a 12-month paid fellowship for people at the very start of their careers. They train you to use Claude..."
View post on X ↗"Anthropic confirmed this. 'Claude Code is sending info like timezone, proxy and possible AI Lab connections into the system prompt in ways Chinese users can't notice.' imagine the backlash if a Chinese AI lab silently would target American users. this seems to be stopped now."
View post on X ↗"GPT 5.6 is banned because Fable got banned. As expected, Fable got banned for outputting the same response as Kimi and Haiku! In other words - for no reason at all! This entire drama is absolutely senseless and puts the US back significantly"
View post on X ↗"The US government has removed export controls on Anthropic's Fable 5 AI model, clearing it for wider distribution after the startup resolved the Trump administration's safety concerns."
View post on X ↗"xAI's model calendar is aggressive. Grok 4.5 is already in private beta at SpaceX and Tesla. The 1.5T model is expected in July. 2T is expected in August. The 2T model gets bigger data scope, more scale, and a stronger training recipe across every area."
View post on X ↗"Happy Governance Day to all who celebrate! 👮🏽 We know it's that time of the deployment cycle when we have thousands of people gathered up at 9 am to learn about AI Governance at #EPPC26 😂 We've been through this before though! Remember we told everyone how to build [AI governance]"
View post on X ↗"maybe AI doesn't make ideas matter more, it makes weak ideas die faster. when building is cheap, there are fewer excuses left between 'sounds good' and 'does anyone actually care?'"
View post on X ↗