ON AIR
Listen to the update
ACN · AI Coding News, the podcast. New every weekday.

ON AIR
ACN · AI Coding News, the podcast. New every weekday.
Sponsor
Update: Anthropic starts trading on the Nasdaq. Meta staff say its AI agent exposed private data. And Google builds a copy of the web to test its agents. From AI Coding News, this is your midday AI news update for Thursday, November 19th.
Our top story: Anthropic has started trading on the Nasdaq. On Wednesday, November 18th, we reported its initial public offering priced at $83 a share. CNBC reports the stock opened at $111.50. That gives Anthropic a market value at the open of about two trillion dollars, behind only five companies. At the opening bell, CEO Dario Amodei said Anthropic will keep pacing how it trains and releases AI. For teams building on Claude, its maker now answers to public shareholders.
Next, Meta. Reuters reports that Meta employees have raised concerns about the company's AI agent, Muse. They say it mixes up conversations across accounts and exposes private data. Meta did not immediately respond to TechCrunch. Employees described childcare requests and bank balances that weren't theirs. One called it an extreme privacy violation. Muse has more than 10 million downloads since launching in September. So teams linking agents to user accounts now have a reported case of data crossing between users.
When we come right back: Google's copy of the web for testing AI agents. Your coding agents say they're done. Are they? Levr is the control plane that verifies and gates your coding agents' work against the plan. Levr holds the plans, issues, tests and gates for every agent doing the work. And work only moves forward when the gates are verified, not when the agent claims it's done. Levr. Try it free at levr.one. Welcome back.
And finally, Google. Google says its Chrome team has built a mirror dimension: a synthetic copy of the web for testing AI agents. The team says testing agents on the live web can be irresponsible, since agents can cause harm. Its fake sites carry prompt injection: hidden instructions in web content. Google says the system makes over thirty times the attack tests one engineer could by hand. It means Chrome's agents meet those attacks on a copy of the web first.
That's your midday update for Thursday, November 19th. Sources are linked in the description. ACN is brought to you by Levr, the control plane that holds your coding agents' work to the plan, at levr.one.