Practical AI news for people running businesses. Today, Alibaba's new flagship Qwen model arrives at a lower price, a new standard tells AI agents what they're allowed to do at a business, and a Microsoft 365 outage that lasted most of a day shows what happens when an alarm only has one way to reach you. North Wayne, who writes First Opinion, joins us on the Qwen story. Plus three quick hits, and what each one means for you.
Today's stories
- 00:21 Alibaba makes Qwen4-Max, its most capable model, generally available on QwenCloud after a preview that began November 11. It is a 2.4 trillion parameter mixture-of-experts model with about 71 billion active per token, image and video input and a one million token context. API pricing is $1.50 per million input tokens and $4.50 per million output. Alibaba says it ranks 6th on Code Arena and, on its own Work360 benchmark, cuts the average cost per task by about 60% versus Qwen3.8-Max. Smaller open-weight Qwen4 models are promised soon. North Wayne, who writes First Opinion, on why the saving is versus Alibaba's own older model, where your data goes, and how to test it side by side. (Alibaba Cloud, QwenCloud, TechNode)
- 02:39 Sierra releases version 1.0 of its Personal Agent Protocol, Poppy, a standard that lets a company tell a visiting AI agent who it is dealing with and what it may do, with customers signing in and choosing what their agent can reach. Shopify has switched it on for every store, Meta's Muse looks for it first on every site, and more than 100 companies have signed up. Mercury and Turkish Airlines are live, with payments and upgrades still approved by the customer. The specification is still marked as a draft. If you sell on Shopify, check what your store now tells agents; if not, decide your agent rule on purpose. (Sierra, The Next Web)
- 04:20 A Microsoft 365 outage that began on Thursday, November 12 spread to more than 70 countries and kept Outlook email down for about a day, with Teams and Copilot also hit. Microsoft says it was not a cyberattack or AI, and its report blames a latent defect in a gateway fix rolled out since July, set off by extra load and reduced capacity. On Friday a faulty chlorine dosing system at a theme park hotel pool in Germany sent its warning by email, which never arrived; 15 people were checked by first responders and four went to hospital. Give the alerts that matter a second channel, and write down the manual fallback for the AI tools and automations your team relies on. (WIRED, MyBroadband, Notebookcheck, WDR)
Quick hits
- A petition to reinstate Jensen Huang as Nvidia's CEO passes 160,000 signatures, with Elon Musk among the backers. The Financial Times reports Huang has discussed a return and the board sounded out Arm CEO Rene Haas; Nvidia says it has not formally engaged any candidates. Nvidia reports earnings on Tuesday. (CNBC, CNBC (investor letter))
- The Federal Reserve's November Financial Stability Report finds AI the most cited risk to the financial system, named by 87% of market contacts, up from 50% in the spring, over valuations, cybersecurity and complex AI financing. (Federal Reserve, CNBC)
- LangChain raises a $300 million Series C at a $3 billion valuation, led by CapitalG. It says 7,000 customers use LangSmith, and pitches it as not locking businesses into any one model. (TechCrunch)
---
AI. For the people who can’t afford to guess
AI Squawk is a daily AI news briefing for business leaders, produced by VarOps, the resident AI company. For more in-depth reads built on real data and hard-earned lessons from the trenches, sign up for the VarOps newsletters at VarOps.com ›
More from VarOps:
Weitere Episoden von „AI Squawk“



Verpasse keine Episode von “AI Squawk” und abonniere ihn in der kostenlosen GetPodcast App.








