Week in Product #483 🚀
GTM is the new moat, Agentic development at Spotify, Claude Mythos: too powerful to be safe, New ChatGPT tier, YouTube avatars, Google Finance, Foldable iPhone, New Meta models, & more
Hi folks 👋
Welcome to a new Week in Product!
🎰 The week in figures
$30B: Anthropic triples revenue to $30B. Jumps from ~$9B at year-end 2025 to $30B ARR in 4 months, with 1,000+ enterprise customers now spending $1M+ annually
$450M: Perplexity hits $450M in ARR, growing at more than double the rate of the previous quarter. The pivot away from search and toward Computer (Perplexity’s agentic workspace), along with a shift to a use-based pricing model, has given the company a major boost. Their user base reportedly now exceeds 100M
$130M: Spain’s Xoople raised a $130M Series B to build a satellite data business designed to generate better Earth data for AI models
📰 What’s going on
Claude Cowork is now officially available on all paid plans after a few weeks in PReview. The official release comes with new features, like the ability to control which tools Cowork has access to, as well as analytics features. In one shared example, Zapier connected Cowork to Slack, Jira and their internal databases to surface engineering bottlenecks and build a custom dashboard that their PMs and designers could copy
Anthropic launched Claude Managed Agents, a new suite of APIs that lets product teams ship and manage AI agents using Anthropic’s own infrastructure, rather than hosting them yourself
Anthropic told Claude Code subscribers they can no longer use their subscription limits for third-party tools like OpenClaw, moving them to a separate pay-as-you-go plan. The policy will roll out to all third-party harnesses soon
Anthropic has announced Project Glasswing, a cybersecurity initiative that brings together Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks to secure the world’s most critical software using a powerful new AI model. The initiative is built around Claude Mythos Preview, an unreleased frontier model that has already found thousands of high-severity zero-day vulnerabilities, including some in every major operating system and web browser. Claude Mythos Preview won’t be made publicly available due to safety concerns, but its capabilities reveal how AI models can now surpass most humans at finding and exploiting software vulnerabilities
OpenAI has launched a new $100/month ChatGPT Pro tier. The mid-tier plan sits between the $20/month Plus option and the $200/month top tier, and targets developers who rely heavily on Codex for coding. It offers 5x more Codex capacity than Plus
OpenAI, Anthropic, and Google moved to crack down on AI model copying in China as concerns over distillation and illicit extraction intensified. This phrasing is an inference based on recent reporting that OpenAI and Anthropic had already accused Chinese labs such as DeepSeek, MiniMax, and Moonshot of extracting model outputs for distillation
Meta’s new Superintelligence team just released its first product: the AI model Muse Spark. It’s going to take over powering the Meta AI chatbot, and it’s a closed model (Meta is keeping the design and code to itself). That’s a strategic pivot for Meta AI, which has long focused on its Llama family of open-source models
Meta is also developing open-source versions of next frontier models, Avocado (LLM) and Mango (multimedia), to get public releases alongside proprietary versions
Meta started using AI token usage as a key employee performance metric
The Google Gemini app can now transform your questions and complex concepts into customizable interactive visualizations directly in your chat. You can initiate the new feature by selecting “Pro” as your model and asking Gemini to “help me visualize’. Once the first version is built, you can then iterate upon it, much like you can do with assets built in Google Studio
YouTube is introducing a new feature that will make it easier for creators to build a clone of themselves. That avatar can then generate new Shorts from text prompts (up to 8 seconds) or be dropped into existing videos. All avatar videos will be clearly flagged as AI-generated
GitHub is seeing a surge in traffic driven by autonomous AI coding agents, which are hammering its APIs and infrastructure. The influx is boosting usage and relevance for Microsoft, but it’s also causing service degradation and outages that impact human developers
X rolled out automatic post translation and AI photo editing powered by Grok
Perplexity Computer integrates Plaid to link all financial accounts. You can build custom budgeting tools, ask finance questions, track spending, debt and investments through conversational AI interface
iPhone Fold leaks reveal wide-screen design. Production delays push launch timeline as Apple refines foldable prototype
Atlassian announced a new set of features that lets product teams use the data they already have in places like Confluence and “remix” these into different assets. Users can select any content on a Confluence page and instantly transform it into a visual format optimized for how someone needs to consume it
Adobe launched Student Spaces, a free Acrobat study tool that turns documents and notes into flashcards, quizzes, podcasts, mind maps, and presentations
📚 Good reads
What agentic-first development looks like at Spotify. Spotify’s chief architect, Niklas Gustavsson, chats with Anthropic’s David Soria Parra (co-creator of MCP) and Christian Ryan (applied AI lead) on what agentic-first development looks like at scale
GTM is the new moat. Jaclyn Konzelmann argues that since AI makes building cheap, the real challenge is not “can we ship?,” but “who exactly are we building for, and how do they experience it?.” GTM and product are effectively merging into one continuous, context-aware experience
How to use AI without losing your mind. Dan Hockenmaier argues that AI helps when you either automate big, repetitive execution work or use it as a thought partner for complex, exploratory problems, but not as a shortcut to avoid thinking. Chasing every new tool will slow you down, so be intentional about where AI fits in your workflow, and lean on it for execution and structured research, but keep the critical thinking to yourself (for now…)
The illusion of clarity. We often feel we get a concept just because it’s familiar, easy to recall, or a Google away, but that’s not real understanding. The moment you try to explain something step by step in plain language, you expose the gaps in your mental model. As PMs, regularly pressure‑testing our understanding by writing or teaching is a simple way to catch this illusion early
🧑💻 Worth learning
Planning for uncertainty with Product Roadmaps. Teresa Torres walks through why old-school feature roadmaps keep failing everyone (product, sales, marketing). She argues that Now/Next/Later works best when you change what each column contains: concrete solutions in “Now,” customer opportunities in “Next,” and outcomes in “Later”
🔧 Products to try
Google Finance Beta lets you ask a market question like “why is Nvidia up?” and get live data, news, and charts in one place
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 week ahead!
