el producto #455 🚀

WhatsApp live translations, Chrome's dev MCP server, Microsoft's agentic tools, From chatbots to AI Agents, Claude<>Figma MCP, How Brex uses AI, Google mood board & more

Hi friends 👋

Happy weekend, and welcome to a new edition of el producto

🎰 The week in figures

$14B: A $14B deal was approved to keep TikTok operating in the U.S. An executive order will require ByteDance to cut its stake in the app’s American business to under 20%, with Oracle, Silver Lake, and Abu Dhabi’s MGX Fund taking major ownership

$10B: xAI is raising $10B at $200B valuation

3B: Instagram hits 3B users

$873M: the top 6 AI coding agents (Cursor, Replit, Lovable, Cognition, Stackblitz & Emergent) are already generating $873M in annual revenue

$100M: Cohere raised $100M at $7B valuation. Partnered with AMD to run AI models on AMD’s GPUs, positioning against OpenAIand Nvidia. Focuses on enterprise needs and hardware diversity amid AI regulatory concerns

$100M: Databricks has signed a $100M multi-year deal with OpenAI to integrate its AI models into its platform, giving business users access to advanced AI tools (that can be used alongside its own AI agent, Agent Bricks) without needing to exit the platform

1M: Klarna hits 1M US card sign-ups in just 11 weeks

85%: Waymo’s robotaxis now 85% safer than human drivers

---

📰 What’s going on

Meta rolls out real-time translation feature on WhatsApp. Translations will initially be available in six languages on Android and 19 languages on iPhone

WhatsApp added QR code-based payments to its Business App, allowing small businesses to share QR codes with customers for direct payments using their preferred method

Google is rolling out Gemini in Chrome to all Workspace users. Gemini in Chrome is Google’s version of an AI browser that can perform most of the tasks of other AI browsers like Dia and Comet, such as summarizing web pages and referencing content across multiple tabs

Google launches an AI-powered mood board app to build Pinterest-like design boards. Their PM Director Jaclyn Konzelmann tells us all about it here

Chrome is getting a development MCP server, which could be helpful for product and engineering teams to debug issues in tools like Cursor and Claude Code

Google also revealed an updated version of its Gemini Live Voice APIs. According to the PMs who worked on the release, the new version is twice as good at successfully calling the right functions, and also has new features that make it sound more natural and fluid, such as pausing in the right place or an improvement in ‘barge-in’ recognition (where a user interrupts the AI voice assistant)

Google Live Search is finally releasing across the US. This new way to search lets you select the “live button” and then start a conversation with your camera switched on to get answers to questions

Microsoft announced new agentic tools to help engineers and product teams tackle tech debt. The new tools include new capabilities in GitHub Copilot, which analyze your codebase, suggest safe migration paths, and build those changes

ChatGPT Pulse is a new feature where each night, ChatGPT researches topics you care about based on your chat history and feedback, then delivers focused visual cards the next morning that you can scan or dive deeper into (only available to Pro on mobile atm, but coming to Plus “soon”)

OpenAI is hiring to bring ads to ChatGPT. OpenAI is on the hunt for someone to oversee all monetization efforts, including bringing advertising to ChatGPT

Perplexity released a search API service this week. Now, other companies can put the company’s search prowess to work in their own context, as Perplexity makes a play to become a platform as well as a service

Perplexity has also launched an Email Assistant. You can add the Perplexity Email Assistant to any email thread, and it will check availability, suggest meeting times, and send calendar invites. It will also read your inbox and automatically label and organize your emails - and even auto-draft responses that match your existing tone and style. The new Assistant is only available for Max users who pay $200/month. 10x the price of alternatives like Superhuman

Amazon’s Q Developer is lagging behind competitors like Cursor and Windsurf, with just $16.3M ARR after a year. Internal demand for external tools is high—even Amazon staff prefer Cursor for speed and features. Amazon is now rethinking its Q branding and product strategy as the AI coding market heats up

Claude has added a new Figma MCP Connector that can view Figma mock-ups at the data level (things like components, design tokens and layout rules) and translate them into production-ready code. Early feedback is mixed, with some saying that the MCP connection only works if Figma is configured perfectly, and others saying that this could mark the end of front-end development as we know it

Figma also now supports MCP on Figma Make. This makes it easier for non-designers to mock up prototypes in Make that match design systems set up in Figma

GitHub has launched a new comprehensive MCP Directory, which lists all of the official MCP servers you can use in your own product or AI app. Each server is backed by its GitHub repository

Neon, the nr.2 social app on the Apple App Store, now pays users to record their phone calls and sells data to AI firms

---

📚 Good reads

How Brex is rebuilding operations around AI agents. Brex is redesigning its product and operations to make AI invisible, automating manual work and shifting roles from task management to agent management. Their internal agent platform lets teams build, test, and deploy AI agents, speeding up workflows and reducing legal friction. The result: faster onboarding, smarter automation, and a new focus on hiring generalists who can design and manage AI-driven processes

From chatbots to AI agents: the next workplace revolution. Aakash Gupta shows us how AI at work is moving fast, from simple chatbots to autonomous agents that handle complex workflows end-to-end. In the next few years (months?), most pros will manage teams of specialized AI agents, multiplying productivity for those who adapt early. Prompt-writing just won’t cut it anymore

From AI hype to real GTM impact: why workflows matter, by Maja Voje. Most companies talk up AI, but only 11% have real, end-to-end AI workflows in GTM. The real challenge isn’t adopting “any AI”—it’s using AI to deliver actual business ROI, starting with rethinking your workflows. Human skills like judgment and stakeholder management still matter most, so let AI handle the busywork while you focus on what moves the needle

Agentic AI is reshaping product design and team roles. AI agents are moving us from static products to dynamic, adaptive systems, changing how teams build and govern software. Designers and PMs now architect systems that generate experiences, not just screens, while balancing speed, quality, and human judgment

How AI is reshaping the Four Fits growth framework. Brian Balfour explains AI is rapidly disrupting every part of the Four Fits—product, market, channel, and model—forcing companies to rethink how they achieve and maintain growth. Product-market fit can now appear or vanish almost overnight, and old distribution channels like SEO are losing their power as new AI-driven channels emerge. To survive and thrive, PMs need to constantly reassess all four fits, not just PMF, and be ready to pivot fast as AI changes the rules

How Claude Code is built: AI-first engineering at Anthropic. Claude Code exploded in usage and revenue by letting Anthropic’s engineers prototype and ship features at lightning speed, with 90% of its code written by the tool itself. The team chose a “model-friendly” tech stack (TypeScript, React, Ink, Yoga, Bun) and built a minimal UI so users feel the raw power of the model. Rapid prototyping, simple architecture, and a flexible permissions system are key so AI-first teams can move fast and scale output

When too many choices make us less free. Anne-Laure Le Cunff explains how endless options can paralyze decision-making, making us feel less free, not more. She explains how “choice overload” leads to anxiety and second-guessing, and offers practical tips: anchor your priorities, narrow options, set deadlines, and settle for “good enough.” True freedom comes from choosing confidently and moving forward, not chasing perfection

---

That’s a wrap for this week! 🌟

I’d love to hear your thoughts—what stood out to you, and how are you thinking about integrating these insights into your Product strategy? Reply to the email or drop a comment on Substack to share your take. And if you found this valuable, forward it to a fellow PM, Product enthusiast, startup founder or entrepreneur who’d enjoy the read

More next week! 👋

Angel