Hi friends 👋
Welcome to a new Week in Product!
🎰 The week in figures
$15B: Modal Labs, the AI infrastructure startup, is discussing a new round at a $15B valuation, months after raising $355M
$1B: DeepSeek’s annualized revenue reportedly reached $1B while the company worked to close roughly $7.5B in financing
$1B: TypeSafe AI (maker of Jev) reportedly discussing a $1B+ round at a valuation above $10B, roughly a week after raising $40M at about $200M
$400M: Island, the enterprise browser security startup, raised a $400M round to fund R&D, international expansion, and workforce growth as businesses confront AI-enabled threats
$340M: Heidi, the healthcare AI startup, raised $340M, including a $100M Series C at a $900M valuation
$250M: OpenEvidence, the AI medical search platform, raised $250M at a $15B valuation to expand its physician-focused diagnostic and treatment tools
$250M: Precision Neuroscience, the brain-interface startup, raised a $250M Series D to advance clinical trials, secure further FDA approvals, and prepare its surface-mounted implant for commercialization
$140M: Basecamp Research, the London life sciences startup, raised a $140M Series C
$100M: Numeral, an AI-powered sales tax compliance platform, raised $100M Series C
$100M: Micro1, the AI training-data provider, raised over $100M at a $4B valuation, with two frontier labs and two xAI cofounders participating
500K: Meta’s Muse passed 500K users and 250K daily actives about a week after launch, while xAI’s Grok Bot reached roughly 418K weekly users after its first month. And it’s getting some fun integrations, like GitHub or Expedia
263%: Global AI-glasses shipments jumped 263% YoY in H1 2026, with Meta accounting for 94% of display-less units, according to Counterpoint Research
📰 What’s going on
Google, OpenAI, and Anthropic began forming a frontier AI standards group around independent testing, incident reporting, and auditor standards
ChatGPT Voice can now use plugins like your email, calendar, and Slack. Powered by GPT-6 Astra, Sol, and Luna, it can be used in ChatGPT Work on web and mobile
Anthropic released Opus 5.5, which delivered its strongest results yet on internal safety evaluations and cut operation costs by 40% compared to Opus 5
Anthropic has reportedly pushed its planned IPO from October to November, in part due to the discourse around an “AI slowdown”
Remember Tamagotchi? Meta is launching its smarter cousin, Muse Charm, a pocket-sized device you can talk to, give a job, and let it handle things for you
Meta unveiled mobile and browser tools that turn prompts into Horizon games, then distribute them through Facebook and Instagram
Meta also announced Ray-Ban Meta Gen 3 (available now), Ray-Ban Meta Audio (ships October 13) with open-ear listening and assistant access, and Meta VR Glasses (a much lighter Meta Quest-like device for $1,299, weighing 100 grams, planned for spring 2027). Ray-Ban Display is getting upgrades in navigation, calendar, spatial audio, and Threads features
Amazon has blocked Meta's Muse shopping agent from accessing its platform, just two weeks after the AI assistant launched. The move highlights growing tension over who controls customer relationships as agentic commerce gains traction
Microsoft unveiled a unified Copilot app spanning workplace chat, no-code app creation, and autonomous agents
Apple is prototyping a screenless fitness tracker with a fabric band and sensor module, targeting Whoop and Oura
Google launched Gemini 3.8 Live with Live Avatar, a new enterprise feature that pairs real-time video generation with speech, enabling AI agents with precise lip-syncing and natural expressions across 97 languages
Gemini phone calls are being tested for U.S. Pixel 11 subscribers, letting the agent call businesses, wait on hold, check stock, and make reservations
YouTube launches Gemini-powered conversational editing for Shorts, dynamic thumbnail testing, and live auto-dubbing set for 2027, expanding Shopping affiliates to 35 countries
Grok 4.7 is out. xAI's latest AI model for coding and knowledge work starts at $2/M input tokens, $6/M output, half the price and twice the speed of rivals. 2.1T-parameter model trails Claude & GPT-6 on key benchmarks
Linear built a dedicated bot that turns social media mentions into structured product feedback that can be used to shape new features or iterations of your product
UserTesting released new MCP servers for research recruitment across the UserTesting and User Interviews platforms. Product teams can recruit participants, launch studies and validate AI-generated concepts from inside ChatGPT, Claude, Figma Make or Gemini
🤖 Futurism
China ready to test the TMS-10, a supersonic passenger aircraft able to fly Beijing to Shanghai in 30 minutes at speeds up to Mach 4 (3,069 mph)
Claude agents uncovered a previously unknown CRISPR-like enzyme system in bacteriophages, prompting Anthropic to open an internal biology lab
📚 Good reads
AI isn’t your spaghetti-brain enemy. AI feels chaotic when you manage it reactively, the same way struggling first-time managers handle their teams. Hilary Gridley explains how to treat AI like a direct report. Align early on problem framing and project plans, use strong reference sets to define “what good looks like,” and batch work to avoid constant context switching
What it’s like to work at an AI-native company. Elena Verna explains her experience at Lovable, an AI-native flat org with little meeting time and lots of autonomy (and AI agents). Ownership is blurry, and there are constant reorgs, and that’s exhilarating and exhausting, but sounds like the place to be right now
Top performers aren’t smarter, they work differently. Polina Russell digs into research showing that top performers aren’t necessarily defined by IQ, ambition, or personality, but by 9 learnable work strategies
🎧 Good listen/watch
[Video Podcast] Muse, the friendly AI agent for normal people.
Claire Vo reviews Meta’s Muse, showing how its careful UX, permission model, and activity feed make it feel trustworthy and helpful for everyday family life. I’ve been using it for a couple of time-consuming cases (like sorting out rental car pick-up questions, or researching car insurance), and have been pleasantly surprised! Complement with Julie Zhuo’s mini-review
[Video Podcast] Jev and “system one” models. Jev founder explains their ML, code-first AI that outputs structured decisions (choices, scores, booleans), optimized for intelligence per dollar and reliability. A long and insightful podcast that will tell you everything you need to know about Jev. Complement with this note by Jaya Gupta explaining why Jev is to be taken seriously:
[Video Podcast] AI fluency at Zapier. Aakash Gupta talks to Zapier CEO about their AI fluency rubric, adopting an AI mindset, and redefining how PMs are evaluated. They now run fully agent‑driven operating models, and the podcast explains how to build a personal “agent stack”
🧑💻 Worth learning
Getting the most out of Opus 5.5 in Claude and Claude Code. Opus 5.5 is built to handle long, multi-step work with less hand-holding. Define what “done” looks like and let it run. For PMs, think of it as a dependable long-run collaborator
Using Jev (a lot of Jev this week, but hang on with me, it’s worth it). Aakash Gupta explains what Jev is, how to set it up, and practical cases/examples to get it to work
🔧 Products to try
OpenMuse lets you self-host a Muse-style personal agent with a persistent browser, files, optional Linux workspace, Gmail, Calendar, and background tasks; free
Kimi Browser Extension gives the Kimi AI model a Chrome/Edge sidebar that can navigate, click, fill forms, extract information, and record repetitive browser flows as reusable skills; free tier available
🧍Personal updates
After two amazing years in New York, my family is moving back to California! And (of course) I had to make a relocation planner app to help us find the ideal place to move to based on personal preferences. Sharing the GitHub repo in case it helps anyone with upcoming moves.
If you happen to be around the Bay Area and want to connect, feel free to ping!
That’s all for this week.
Feel free to drop your comments/questions/feedback. Would love to hear what you’d like to see more about in WiP, so I can make it better for you.
Have a great start to your week!
