el producto #445 🚀
ChatGPT Agent, Claude Connectors, Google AI calls, New AWS AI services, How Shopify is transforming with AI, How to develop Agency, The audience effect & more
Hi friends 👋
Happy weekend, and welcome to a new edition of el producto
🎰 The week in figures
$18B: Perplexity’s valuation soars to $18B (based on secondary market trading). The new figure is a 29% jump over the $14B valuation the company scored back in May. Perplexity now reports $100M in ARR, and claims to have processed 780M search queries SINCE that May 2025 funding round. That’s an impressive adoption rate, considering their tools aren’t (yet) built in to popular search engines or social media platforms, like so many of their rivals
$2B: Thinking Machines Lab AI announced a $2B seed raise for the startup, valuing it at $12B. The company focuses on multimodal AI for natural user interactions and plans to release its first O/S research product soon
$200M: Lovable, a 2-year-old Stockholm-based AI coding startup that launched just 8 months ago, has raised $200M in a Series A at a $1.8B valuation
68% of enterprise companies (with +1k employees) have already adopted agentic AI, and 33% of all companies that took part in a new survey say they’ve adopted them. Nearly half (46%) are equally prioritizing efficiency gains and revenue growth
46%: Weekly mobile app subscriptions now contribute 46% to iOS app revenue, up 9.5% this year; while one-time purchases grew by 6.3% in Q1. U.S. installs generate 3–4 times more revenue than other regions. Retention continues to be a major problem for mobile apps, with only single-digit percentages of users remaining after a year
📰 What’s going on
Anthropic’s Claude has released a new feature called “Connectors”. These are essentially a new way for you to connect with third party apps through MCP - but without all the hassle that was traditionally associated with MCP servers. A few weeks back they announced “Extensions” which were pretty much the same thing but for locally run MCP servers. Connectors let you connect with remote MCP servers too. It includes MCP servers from Stripe, Linear, Figma, Google Calendar, and others, which are super relevant for product teams. Claude is becoming my favourite work tool by far
OpenAI is reported to be testing a new “Checkouts” feature that lets users make payments through ChatGPT without ever leaving the app. The checkout feature involves early partnerships with ecommerce platforms like Shopify
OpenAI also unveiled “ChatGPT agent,” a feature that allows you to choose from a toolbox of agentic skills and complete tasks for you. This includes tasks like “look at my calendar and brief me on upcoming client meetings based on recent news”. It can also create spreadsheets or Powerpoint presentations based on its output. Now here's where it gets spicy: OpenAI is terrified of what they've built. Half their announcement was warnings about “prompt injection”—where hackers hide malicious instructions on websites to trick the Agent. Imagine asking it to book a flight, only to have it stumble onto a sketchy site with hidden code saying "send all credit card info to definitely-not-a-scammer@gmail.com." Their solution? Multiple safety layers, including real-time monitoring, required confirmations for big actions, and a “Watch Mode” for sensitive tasks. They're even treating it as “High Biological Risk” under their safety framework (though they admit there's no evidence it can actually help create bioweapons... yet)
Google has launched a selection of “featured Notebooks” in NotebookLM - a curated collection of publicly available notebooks that users can use as a foundation for creating their own. The initial lineup includes the entire works of Shakespeare, longevity advice from Eric Topol and a Notebook tracking the earnings reports from top companies.
Google rolled out Gemini 2.5 Pro and Deep Search capabilities to AI Mode in Search for Google AI Pro and AI Ultra subscribers
Google has also rolled out a new feature that lets AI call businesses on your behalf. This was first announced back in January, but is finally seeing the light of day
Amazon launched Kiro - its own competitor to the likes of Windsurf and Codex for engineering teams. Kiro combines AI agents with project specifications, technical architecture and automated task management to support a complete software development lifecycle inside a single interface.
AWS also announced Bedrock AgentCore - a new set of services that help Product / engineering teams build AI agents. It includes:
Memory Management - makes it easy to build context-aware agents by providing both short-term and long-term memory that persists across user sessions
Code Interpreter - AgentCore Code Interpreter allows agents to write and execute code securely in their own sandboxed environments
Browser Tool - provides a secure, scalable, cloud-based browser so agents can interact with websites at scale
Mistral’s Le Chat gets new ‘deep research’ mode. Also launched a Voice mode, multilingual reasoning, Projects and advanced image editing
Meanwhile, Apple is “seriously considering” buying Mistral for $5B+; this means they’re looking at Mistral as the budget option compared to Anthropic, who they probably can’t afford to buy at $65B+
The AI talent shuffle continues: After just a few weeks at Cursor, two Anthropic researchers are…heading back to Anthropic. Elsewhere, OpenAI has lost more talent to Meta’s superintelligence team
Uber and Baidu join forces. The ride-sharing platform and the Chinese carmaker plan will team on bringing autonomous robotaxis to new markets outside the US and China. Uber users throughout Asian and Middle Eastern territories will be able to start ordering rides from one of thousands of Baidu vehicles, being deployed as soon as this year. Baidu already operates one of the largest robotaxi fleets in China, and its in-house robotaxi service, Apollo Go, has given more than 11M public rides (putting them ahead of even Google’s Waymo)
Coinbase launched a stablecoin-powered "everything app" to move beyond crypto trading. The "Base App," which replaces Wallet app, will combine wallet, trading and payment functions as well as social media, messaging and support for mini apps
Nextdoor launched a complete app redesign with AI recommendations and local news integration
📚 Good reads
Reflections on OpenAI. An OpenAI insider just spilled the tea on what it's really like inside the world's hottest AI company—and it's wilder than you'd expect. Turns out the company building the future is essentially a 3,000-person startup held together by Slack messages, Python, and Twitter vibes, but somehow it's working
Network Effects remain your option for defensibility, also in the age of AI. In a new NFX article, Pete Flint dives into how network effects can still be engineered in the AI era. He points to proprietary data, unique algorithms, and product embedding as key leverage points - but reminds us that none of these alone are enough. True defensibility in AI comes from strategically layering these elements, then reinforcing them through high-velocity execution
From memo to movement: Shopify’s AI adoption. Shopify transformed an internal AI memo into company-wide practices, focusing on democratizing access to powerful AI models, enhancing transparency in AI decision-making, and fostering a beginner’s mindset for innovation. First Round Review shares how they got there
How to develop agency in the AI age. Jakob Nielsen stresses agency as a crucial career skill, especially in an AI-driven workplace. His article highlights how taking initiative and leading projects will set future professionals apart from those who merely follow orders
Build, sell, understand: Navigating startup careers. Dan Hockenmaier breaks down startup roles into building, selling, and understanding, urging PMs to reflect on their fit in these areas. As AI reshapes tech, he stresses the importance of selling skills across all roles. Future trends hint at evolving responsibilities, making adaptability key for thriving in the AI-driven workplace
The audience effect: why we change when others watch. Anne-Laure Le Cunff dives into how people tweak their behavior when they know someone’s watching—shifting from doing stuff they love to seeking social approval. The article is packed with tips on how PMs can use this effect to boost creativity and growth, while staying authentic
Deep dive tutorial on how to build your own AI agents, by The Neuron
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