Google is destroying its own business - and still winning
Plus: Slack launches a "new AI era", Linear's dashboards and are product managers the new bottle necks?
Hi product people đ,
This week, Google delivered an impressive earnings call that told a curious tale of a company that is both destroying itself and winning at the same time. Their CEO took their earnings call this week as an opportunity to position Alphabet as the company best placed strategically to win across each of the important AI dimensions. Is he right?
Plus, major new releases from Slack, GitHub and Linear - and a cautionary tale from a Replit user whose AI agent made a âcatastrophicâ mistake that forced their CEO to make an apology.
Have a great weekend ahead!
Rich
Watch on YouTube | Follow on Substack Notes
Key reads and resources for product teams
Knowledge Series - Software 3.0: What product teams need to know
In this Knowledge Series, weâll dig into some of the core concepts and expand upon these to understand how product teams can prepare for this new era. From updating API docs to accommodate AI Agents and non-engineers building prototypes through to building features with âpartial autonomyâ that allow AI and users to work together, Software 3.0 upends many of the ways of building products that we previously took for granted.
AI skills - How to write effective AI prompts for high quality output
Explore 5 of the most powerful types of prompts and how these can be used at work with leading tools like ChatGPT, Claude and Perplexity at work. Updated this week to include a new guide to metaprompting. (Department of Product)
Case Study - How Notion built the new version of their MCP server
Notionâs hosted MCP server allows AI tools (like Cursor and Claude) to interact with Notion workspaces using natural language, skipping the need for complex API integrations. In this piece, their engineering teams explain how they built the second iteration of their product. An insightful read if youâre planning to introduce MCP into your own product stack. (Notion)
The Product Triad explained - The product triad (also called the product trio or three-legged-stool model) is likely the most common pattern for organizing product teams in Agile organizations (NN group)
Why development leaders are investing in design - Leaders who rated design as âextremely importantâ were five times more likely to say their last project far exceeded expectations. (Figma blog)
How to build vibe coded apps with Lovable - a handy guide full of best practices from Lovableâs own designer ( Felix Haas )
New product features and innovation this week
First up, GitHub has launched its new vibe coding app, Spark. Itâs GitHubâs answer to competitors like Loveable, Replit and Figma Make and it allows users to build personal vibe coded apps in a few seconds. At first glance, this does just appear to be a copycat app but since Spark is built directly into the GitHub platform, it actually builds entire mini apps with backends that are fully functional, not just the frontend UI.
Speaking on X, Satya Nadella says heâs loved playing with Spark so far but itâs going to take a lot to stop Lovable. This week, they were officially crowned the fastest growing company in history, reaching $100 million in record time. Their CEO says that, incredibly, a massive 10% of the new websites created on the internet last month were built with Lovable.
MeanwhileâŚ
Slack has unveiled a bunch of new AI features in what it is calling âa new era of AI in Slackâ. The updates include a new jargon translator to provide instant explanation for acronyms and project names, writing tools for Canvas and enterprise search which allows users to search across a companyâs entire knowledge base from apps including Google Drive, Confluence, Teams and Slack channels. The search feature ships with a cleaner interface for selecting which file types you want to include, too. As well as this, the company teased a few upcoming features that are currently in development - and one of these sounds a little anxiety-inducing. Itâs a new feature that will automatically generate action items that are added to your profile whenever someone mentions your name relating to a task that youâre responsible for.
The browser wars are continuing with Dia and Perplexity Comet both working on new features that close the gaps in differences between them. Dia is working on agentic actions and Perplexity is getting shortcuts like Dia skills which you can use to perform pre-defined actions quickly. The Browser Company launched a gallery of AI skills you can browse. One of the most recent includes a skill for assessing whether someone is the right fit for a role. Perplexity also published its monthly changelog outlining its major releases which is worth a read.
Linear has released a new Dashboard. Dashboards can be filtered by team or scope, and users can drill into insights to view underlying issues and take action directly. In typical Linear fashion, as part of the release, they produced a beautifully directed video which explains the thought process behind the new feature.
In other newsâŚ
OpenAIâs new CEO for Applications, former Instacart CEO Fidji Simo has outlined her vision for the future of AI and how the company intends to create it. The piece covers AIâs impact on health, economics and product development. âA single person can now brainstorm, prototype, market, and launch a product with tools they control themselves even if theyâve never written a line of code.â she says. Her 9 year old daughter vibe coded an app to use to plan for her birthday - and this will be transformative. The whole piece is worth a read to get a sense of where OpenAI might be headed but there are few details on the exact applications OpenAI is working on.
One version of the future is where we all wear devices that listen to everything we do. This week, Amazon acquired a new startup called Bee - a wearable device that listens and records everything a user does throughout the day. After the failures of devices like Humane and Rabbit, one does wonder why this concept is still gaining traction. But, with OpenAI busy finessing their own new hardware product, maybe Amazon execs know something about OpenAIâs upcoming new device that the rest of us donât.
Are product managers the new bottlenecks?
Tools you can use
Commitify - AI agents that call your phone. This is actually a pretty fun proposition that could boost your productivity. Youâll get real world phone calls from AI personas like CEOs, best friends or âZen Mastersâ who encourage you to stay productive and help solve problems.
Plumb - lets you build your own AI workflows to use at work - and then monetize them by charging others for access.
Traversal - troubleshoots and fixes issues with your product in production. It works like an AI engineer on call and recently raised $48 million in seed funding.
Growdoro - a pomodoro timer app that grows a beautiful looking garden every time you successfully complete a pomodoro.
đ Product data and trends to stay informed
AI continues to destroy traditional SEO. A new study from Pew shows that Google users who encounter an AI summary are less likely to click on links than those who donât. Just 8% of users who saw a page with an AI summary clicked on a link vs 15% who didnât. 26% of users ended their browser session after seeing the AI summary vs just 16% who saw pages without the summary.
Google still dominates the market, for now at least. An estimated 5.6% of US search traffic went to AI alternatives like ChatGPT or Perplexity. But, thatâs doubled year on year. Googleâs CEO took this weekâs earnings call as an opportunity to flex Googleâs position as an AI leader. Search revenue hit a record $54.2 billion in Q2, up 12% year-over-year, beating analyst expectations and Googleâs history of adapting to industry shifts (e.g., acquiring Android, paying Apple for default search) suggests itâs pretty well-positioned to defend its dominance.
Hereâs the strategy diagram Pichai used to demonstrate Googleâs strategic strength, arguing that this has been long in the making:
AI companion usage is exploding. 72% of teens in the US have tried interacting with one at least once and 33% say they find conversations with AI companions more stimulating than with humans. Slightly terrifying but a sign of things to come perhaps.
10% of the new websites created on the internet last month were built with Lovable according to their CEO.
Demand for Agentic AI skills is starting to rise for new roles in product teams. Thereâs been a +985% increase in job postings related to agentic AI from 2023 to 2024. The trend is seen across all product team jobs including Software engineers, data scientists, researchers and product managers. McKinsey Technology Trends report 2025 in full.
But if you are considering introducing AI agents into your workflows, hereâs a cautionary tale worth noting. An AI coding agent at Replit went rogue last week, deleting a companyâs database and then lied about it.
Paid subscribers get the full DoP Substack including: The Knowledge Series for sharpening your tech skills, AI tutorials for putting AI into practice at work and DoP Deep dive reports to learn from the worldâs top tech companies.