el producto #426 🚀
Perplexity's AI phone, How AI is changing Product Management, Anthropic raises $3.5B, New iPad Air, Smarter Siri delayed, Scaling Figma, Using Claude for work & more
Hi folks 👋
Welcome to a new edition of el producto
🎰 The week in figures
$3.5B: Anthropic raises $3.5B at a $61.5B valuation. The company recently released new models, including one tuned for coding, that we were well received. The investment underscores that while some in the market feel that open-source AI models will win the day, there are still lots of believers in the closed-source approach
$400M: Plaid, the financial data aggregation platform, is finalizing a secondary share sale up to $400M at a $6B valuation
$175M: Quantexa, a decision intelligence platform for financial services and the public sector, raised a $175M Series F at a $2.6B valuation
$150M: Ramp, the corporate card and expense management platform, secured a $13B valuation in a $150M secondary deal
$140M: Darwinbox raises $140M. Darwinbox offers a suite or HR services including payroll and onboarding, putting it in competition with companies like Gusto, Deel, and Rippling
$100M: Feeling lost trying to keep tabs on all the different startups working on AGI? It’s time to add one more to the mix. This time it’s Reflection AI, which is coming out of stealth with more than $100M raised and a valuation of more than $500M
$20k: OpenAI’s upcoming specialised AI agents are reportedly set to cost up to $20,000 a month. A “high income” knowledge worker could cost $2,000 a month with a software developer agent set to cost $10,000 a month and a PhD-level researcher priced at $20,000 a month. If one developer agent can eventually do the work of a team of 5 developers, OpenAI can anchor its pricing in this way to sell it to corporates. OpenAI lost $5B last year and its transition to a for-profit can only really happen if it makes a profit, after all. But $20,000 a month does seem a little absurd for a PhD-level AI researcher
80%: SEO is dying. 80% of users now rely on AI summaries at least 40% of the time. The rise of so-called zero-click searches has long been on the horizon but it looks like we’re finally starting to see some real world data on this
1/4: A quarter of Y Combinator’s latest batch now have 95% of their codebases written by AI. These are highly technical founders who, just a year ago, would have built everything from scratch—now, AI handles most of the work while humans focus on debugging
📰 What’s going on
Apple unveils new iPad Air. The company’s latest, lightest tablets are available in 11-inch and 13-inch options for $599 and $799, respectively. They’re upgrading from the M2 to the M3 chip, which the company suggests doubles the processing speed over the classic M1 Air. According to Apple, AI tasks run 60% faster than the previous edition of iPad Airs. The external Magic Keyboard accessory also got a long-awaited upgrade. Orders are now open; the first devices ship on March 12
Apple seems to be struggling to integrate Apple Intelligence features into Siri and will not be releasing the highly anticipated “modernized and more conversational” version—which Apple has touted as being more natural and human-like—until 2027
AI mode is officially coming to Google Search. Google says that AI mode is particularly helpful for questions that need further exploration, comparisons and reasoning. In the preview, it looks as though users can ask an initial question and then follow it up with further questions, with traditional search results appearing in a side panel that can be expanded. It also appears that AI Mode takes the first slot in the navigation, with “All” appearing second. If this were to be replicated outside of this experiment, it would mark a significant shift in how Search as we know it works - with potentially catastrophic consequences for products that rely on SEO
Gemini’s Live Mode will soon get live video and screen sharing capabilities that will make it possible to co-browser sites together and get recommendations
YouTube is preparing a relaunch of its TV app that will promote more alternative streaming services. At the moment, third party content is available in the movie and TV tab but the new change will see it featured on the homepage, too. YouTube is the most watched streaming platform in the US but is treading a fine line here; if the homepage is polluted by upsells for third party providers, it risks alienating users who simply want to watch their favorite content
Amazon is reportedly developing a new AI model that has advanced reasoning capabilities—similar to OpenAI’s o-3mini and DeepSeek’s R1—which allows it to “think through” problems and queries, making it more reliable in areas like math, coding, and science
Perplexity announced it has partnered with Deutsche Telekom to develop an AI phone (costing less than $1,000) that will allow users to use their voice to complete tasks—like booking taxis or doing shopping—without having to switch between apps. The device, which they’ll reveal this year and start selling (to Europe first) in 2026, will run “Magenta AI” which will give users access to a suite of AI tools from Perplexity and also from Google Cloud AI, ElevenLabs, Picsart
Zapier launched Agents, which enables users to create custom AI assistants that integrate with over 7,000 apps, automating tasks and workflows. These agents can access live business data and operate on command or autonomously
Salesforce has announced a new iteration of its AI agent product, AgentForce 2dx. The new iteration is designed to shift away from chat-based conversational interfaces and to allow agents to work in the background instead. A new API and no-code templates are released as part of the update
Google co-founder Larry Page has a new company, Dynatomics, which plans to use LLMs to create “highly optimized” designs for Products, and then build them. Much hay has been thrown over the use of AI in robotics, biology and other domains. This is a natural place for the technology to find application
📚 Good reads
Lessons on Product scaling and storytelling from Figma’s CPO. Figma’s CPO shares how the design juggernaut has thoughtfully expanded its Product suite, building new tools for new personas that have become just as beloved as the core product. Yuhki Yamashita has a simple way to gauge if a new product will make someone go, “I want that!” before launch day: Can its value be distilled in a screenshot? This useful test presses product teams for simplicity. The more time you spend iterating on a product, he points out, the harder it is to step back and evaluate how well it speaks for itself.
How PMs can leverage AI further by calling ChatGPT API. Vladimir Kalmykov shares a great intro to setting up ChatGPT API, and a few simple use cases to try it out
[Podcast] Anthropic’s Chief Product Officer on AI and Product strategy This interview with Anthropic's CPO Mike Krieger explores where value will be created in AI, how foundation models are evolving, and the future of Product development in an AI-driven world. Some of the most impactful predictions he makes include:
Models will differentiate, not converge - foundation models will develop increasingly distinctive characteristics and specialties rather than becoming interchangeable commodities
Software development will transform fundamentally - developers will shift from writing code to becoming "delegators" who manage AI systems, focusing on high-level direction and review
Product development requires new approaches - teams must design for non-deterministic systems, creating "scaffolds" around AI capabilities while developing new testing and evaluation methodologies
How to use Claude for work. I recently shifted much of my GenAI usage from ChatGPT to Claude. Often I even ask Claude to generate a prompt that I can then use in ChatGPT DeepResearch or Operator. Happy to get some validation seeing that Aakash Gupta did the same shift! Here he explains why, and how to get the most value out of Claude
How AI is quietly changing Product Management. Reforge’s Brian Balfour, Saun Clowes and Fareed Mosavat explore how AI is reshaping the role of PMs by enhancing their capabilities rather than replacing them. AI tools are now handling tasks like data analysis, user research synthesis, and even drafting product specs, allowing PMs to focus on higher-level strategic work such as defining vision and prioritizing features. Complement with Mary Cagan’s perspective, which also envisions AI as a teammate, and with this podcast chat between Aakash Gupta and Claire Vo, in which they ancitipace the future of PMs falling in two categories:
The Experience Manager → Focused on UX, prototyping, AI-driven workflows, and interaction design
The Commercial PM → Deeply involved in revenue, go-to-market strategy, and monetization
Top 100 GenAI consumer apps. A16z published a new report on the top 100 GenAI consumer apps by traffic and mobile downloads
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 this 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
See you next week! 👋
Angel