Week in Product #475 🚀
Uber groceries, Gemini to import chats from comps, Thinking together, The A-native blueprint, Claude Cowork guide, AI-powered discovery, More Apple products & more
Hi friends 👋
Welcome to a new Week in Product!
🎰 The week in figures
$380B: Anthropic’s new valuation after closing a $30B Series G, the second-largest private tech raise ever
$140B: Stripe’s expected valuation in a new tender offer, a $30B+ jump from last year and their highest ever. Still no IPO
$935M: Apptronik, the Austin humanoid robot startup raised $935M in a Series A at $5.3B to partner with Google DeepMind to develop robots using embodied AI
$315M: Runway, the AI video startup, raised $315M at a $5.3B valuation to build what it calls "world models" that understand physics and spatial reasoning
$200M+: Total committed by AI companies to the 2026 midterms. Anthropic put $20M into pro-regulation; OpenAI’s Greg Brockman put $25M into anti-regulation
1,000+ tok/s: OpenAI’s GPT-5.3-Codex-Spark speed on Cerebras chips, their first production deployment outside Nvidia
50+ features: Shipped by Spotify in 2025, while their best engineers haven’t written a line of code since December
💰 Q4 earnings
Spotify: net income surged by 94% to €2.2B from €1.1B in 2024. Total revenues jumped by 10% to €17.19B from €15.67B YoY
Robinhood: reported Q4 revenue of $1.28B, falling short of expectations due to a downturn in crypto trading, but predictions, transaction and Gold revenue showed growth potential
Zillow: posted a $3M profit in Q4, reversing a $52M loss from the previous year. Revenue surged 18% to $654M, fueled by sales, rental growth
📰 What’s going on
Apple's upcoming lineup. Expect the iPhone 17e with an A19 chip at $599, new iPads with improved processors, a budget MacBook, and AI everywhere. Apple will open CarPlay to AI Assistants and the AirPods Pro 4 may include cameras
Apple delayed its major Siri overhaul again due to severe testing issues and bugs, pushing the full Google Gemini-powered integration to late 2026 instead of its original March launch
YouTube TV's new plans. 10+ new customizable streaming packages launching in the US. Include: $64.99 sports, $71.99 sports + news, and $54.99 entertainment
Uber Eats introduces an AI cart assistant to streamline grocery shopping via text or images. Partnering with retailers like Safeway and Kroger, it aims to outpace DoorDash and Instacart with enhanced delivery services
Google released a major upgrade to Gemini 3 Deep Think. 48.4% on Humanity’s Last Exam, 84.6% on ARC-AGI-2 tests. Scientists are using it to find flaws in peer-reviewed papers. Available to Google AI Ultra subscribers and select researchers via API
Gemini is testing "Import AI Chats" to let you transfer conversation history from ChatGPT, Claude, and Copilot; professionals reportedly lose 5-10 hours weekly re-entering context when switching AI tools
ByteDance’s Seedance 2.0 is making AI video competitors look obsolete. Native audio, lipsynced speech, 2K resolution, 20-second coherent clips. One creator made a 2-minute cinematic fight scene for $60. Currently China-only; rumored international launch Feb 24
OpenAI started testing ads in ChatGPT for free and Go tier users. Ads are labeled as sponsored and don’t influence answers. Opt out by upgrading to Pro/Plus. Anthropic responded with Super Bowl ads mocking the decision (!!)
MiniMax M2.5 matches Claude Opus on coding benchmarks at 10-20x lower cost. Open-weight, agent-native, $1/hour to run continuously. Chinese model going head-to-head with American closed labs
Jony Ive’s screenless AI device patent leaked. Built with OpenAI, ditches screens entirely for voice-first design. Could launch later this year
A researcher used 16 parallel Claude agents to build a complete C compiler from scratch. It compiles the Linux kernel. Agents coordinated using Git and plain text files — no human oversight needed. Completed in two weeks what normally takes teams years
Stripe deployed "Minions," a fleet of homegrown, autonomous AI coding agents that now generate over 1,000 pull requests per week
📚 Good reads
Thinking together. Anne-Laure Le Cunff argues that our best ideas often come from collective curiosity: openly wondering, doubting, and experimenting with others instead of trying to be the lone expert. She offers simple practices like mapping what you don’t know together, questioning assumptions out loud, walking meetings, and sharing tiny experiments. For PMs, it’s a reminder that learning in public and inviting diverse perspectives leads to better decisions and less untested thinking
The 5-level blueprint to become an AI-native PM. This piece argues that AI is collapsing the “how” of work, so PMs win by getting very clear on the “what” and then letting AI handle execution. It lays out 5 levels, from basic Q&A to running teams of agents, and claims most value sits in level 2 (voice, meeting notes, and custom-context workspaces) and level 3 (prototyping before you spec)
Magic tricks, moats, and the three-body problem of AI networks. Casey Winters explains how AI products are repeating mobile’s magic trick pattern: viral growth, shallow retention, easy to replicate. Consumer networks need audience + network size + value prop to align. Most optimize for speed, but winners optimize for structure
If you want to get hired as an AI PM, go build something, by Jaclyn Konzelmann. The bar is raised for everyone, and successful PMs are upskilling too. Embrace the Product Builder role, or risk falling behind
We stopped forcing the subscription model on our users. Don’t trap users in subscription models. Elena Verna explains how Lovable introduced credit top-ups after user complaints. 20% markup on top-ups won: engagement up, retention improved 7%, 20%+ of bookings now from top-ups
🎧 Good to listen/watch
[Podcast] Multi-lens thinking. John Cutler argues that great product leaders don’t just install frameworks; they design models that fit their company’s reality and help people make better decisions. He stresses shifting between multiple lenses (behavior, culture, politics) instead of clinging to one “right” view, and making frameworks both easy to grasp and deep enough to stand up to real-world pressure. The episode also challenges the copy‑paste trap: internal product‑market fit, minimal viable consistency, and messy, swinging transformations beat blindly copying Silicon Valley “best practices”
[Podcast] Engineers are becoming sorcerers, by Lenny Rachitsky (w/ Sherwin Wu, OpenAI). 95% of OpenAI engineers use Codex; many manage 10-20 parallel agents. PR review times: 10-15 min → 2-3 min
🧑💻 Worth learning
Full guide to Claude Cowork. Ben Al walks through Claude Cowork as a kind of “AI teammate” that can plan, act on your local files, hit APIs via MCP/connectors, browse, and execute code in one workspace. He explains how “skills” bundle prompts, tools, and knowledge sources into reusable workflows that feel closer to flexible, human-in-the-loop automations than rigid n8n/make.com flows or one-off custom GPTs
How to do AI-powered discovery (step-by-step with live demo), by Aakash Gupta (w/ Caitlin Sullivan). Complete workflow for survey and interview analysis with Claude Projects. Survey analysis drops from 30-60 min to 3-5 min per interview. Includes exact prompts and frameworks for building discovery agents
🔧 Products to try
Claude Code /fast mode: New features: multi-repo sessions, better git visualization, slash commands. /fast toggles faster Opus 4.6 responses for rapid iteration. PM use case: Speed up debugging cycles, manage multiple repos in one session
HeyGen video dubbing: Dub any video into a new language while preserving original voice and lip sync. PM use case: Localize product demos, training videos, customer testimonials without re-recording
Shipper: create full Chrome extensions from a single prompt using Claude, handling browser compatibility, privacy policies, and Web Store autofill for about $0.11 per extension. PM use case: build internal tools and POCs fast
That’s a wrap 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 weekend.
