
{{first_name | Hey}}, welcome back.
Today’s issue is about Huang playing with viral robot ducks, Moonshot’s new AI model breaking into the top four, and Mistral’s push to challenge Big Tech with open-weight models and sovereign AI.
These are today's updates.
Huang’s viral video with Microduck robots
Moonshot’s Kimi K3.1 Pro cracks the top four AI models
Mistral bets on open-weight models to challenge Big Tech
Tools, jobs, resources, and last issue’s winning prompt ⬇️
Spotlight: Anthropic
Moving AI beyond chat and into everyday work takes more than giving employees access to a tool. Anthropic's guide to deploying Claude Cowork explores where to start, how to structure a pilot, and how to scale adoption across business functions, with practical examples from its own teams and customers.
It helps you with:
How to choose your first use case and evaluate a pilot
The five levels of Claude Cowork adoption, from chat Q&A to department-wide plugins
How finance, legal, sales, and product teams use Cowork
A roadmap for expanding from small teams to organization-wide deployment
Lessons from Thomson Reuters, Zapier, and Jamf
Start with Claude Cowork
Top News
A video of Jensen Huang playing with Microducks at GTC Berlin went viral, where the 25 cm robots sat, napped and somersaulted on voice command. The $399 robot from Pollen Robotics have been making waves for their super cute avatars since their launch. It has 15 motors, a camera and lidar; walks, picks up objects, gets back up after a fall and can even roller-skate. Hugging Face’s Clem Delangue said preorders are nearing 20,000.
Works best for developers and hobbyists, since owners can train new behaviors in simulation with reinforcement learning, load them onto the robot and share them. People are also plugging in a controller and lining them up for races and football. Note that only the software is open source, not the hardware designs. Delivery is expected before Christmas. For more details and preorder link visit the Microduck page.
Moonshot released Kimi K3.1 Pro on Friday, a post-trained version of Kimi K3 with a 1M-token context window. It ranks fourth on the Artificial Analysis Intelligence Index with a score of 55, behind Claude Opus 5.5 at 58. API pricing is $2 per million input tokens and $10 per million output tokens. Artificial Analysis also found that Kimi K3.1 Pro’s AA-Omniscience Index rose to 30 from 20 for K3, mostly because the hallucination rate fell to 44% from 53%
It is ideal for teams running coding agents or finance workflows that want a cheaper option, since pricing is a third below K3. Note that weights are due by November 4, and Moonshot’s own benchmark claims are unverified. Find more details on Kimi K3.1 Pro capabilities here.
Mistral is pitching open-weight models and sovereign AI as its answer to US and Chinese rivals. It raised €3 billion at a valuation above €21 billion in September in a round led by Samsung. CEO Arthur Mensch said this month that its newest model beats Chinese models on cybersecurity.
It is ideal for European and regulated-industry teams that want control over deployment, IP and costs, since Mistral’s open-weight models can run on their own infrastructure. Note that Mensch did not say which Chinese models or what measure he used. Watch the interview.
Don’t miss this: Grok published the first seven minutes of its AI-made Odyssey, as Episode 1 of a series made with Grok Imagine, Wonder Studios and director Dev Walker. Elon Musk had promised a full-length, “historically accurate” movie three months ago. It’s a good track for people exploring the AI video’s live action capabilities and its pros and cons. You can watch Episode 1 here.
Signals
Tools
Customer.io: Turn customer data into personalized experiences (Startups get 12 months free)
Canva Craft: Turn prompts into design-ready images you can edit
Ami AI: An AI GTM agent that works as Lovable for lead generation
Resources
Poll
Choose a prompt you want us to publish next
Market
Funding
Icon raised $200M to scale robotic construction systems
Tenstorrent raised $693M to advance computing technology and scale its hardware & software ecosystem
Roles in AI
AEO & GEO Marketing Manager at Stripe
Principle Engineer at Snowflake
Tutorial
Learning faster isn't necessarily about being naturally gifted. In this 8 min video, learning expert Justin Sung introduces the Confusion Compass Method, a technique for turning confusion into targeted questions that help you understand new concepts more efficiently.
Prompt of the Day
AI Meeting Agenda Builder
When to use this?
Use it to turn rough notes, project updates, and strategic priorities into time-boxed discussions with clear objectives, decision points, risk assessments, and actionable follow-ups.
Role and Purpose
You are an AI-powered Meeting Agenda Builder designed to help enterprise AI leaders, mid-level executives, department heads, and cross-functional teams plan productive, focused, and outcome-driven meetings.
Your job is to...
{Click on the button to view full prompt}Get more such prompts in the Prompting Playbook (free for you)
Stay curious, {{first_name | reader}}
PS. If you missed yesterday’s issue, you can find it here.


