Week in Product #487 đ
AI-led layoffs, Anthropic's agentic capabilities, Meta and Google developing OpenClaw-like products, Apple opening up to different AI models, Agentic coding at Uber & more
Hi folks đ
Welcome to a new Week in Product!
đ° The week in figures
$45B: Anthropicâs run rate nears $45B, and its (final?) pre-IPO round could push $50B into a valuation that could reach above $1T (pre-IPO valuation just hit $1.2T on secondary markets)
$7.3B: Chinese DeepSeek is raising $7.3B at a valuation of more than $50B
$6.3B: Long Lake agreed to acquire Amex GBT for $6.3B, explicitly betting that AI will reshape corporate travel
$2B: Moonshot AI, the China-based open-source AI platform, raised $2B at $20B valuation
$1B: Kalshi, the SF-based prediction market platform, closed a new $1B round boosting valuation to $22B
$950M: Sierra raised $950M Series E at a $15.8B valuation for its AI customer-service agents
600M: China's AI usage surpasses U.S. in token consumption. Over 600M Chinese citizens now use GenAI daily, with weekly data usage by Chinese models overtaking U.S. models for the first time
$500M: Periodic Labs is raising $500M at a $7.5B valuation for AI-driven materials discovery
1.1k: Cloudflare cut 1,100 jobs in what CEO Matthew Prince called an AI-first restructuring.
23: Google DeepMind's AlphaEvolve scaled to real-world impact, with the team reporting 23 verified scientific discoveries across chemistry, materials, and applied math in the past quarter
25%: DeepL announced staff cuts of about 25% as machine translation margins collapse under open-weight competition
20%: PayPal has plans to cut around 20% of its staff
17%: Coinbase cuts around 400 jobs, 17% of its staff, in AI-driven restructuring
đ° Whatâs going on
OpenAI announced Codex for Chrome: lets Codex autonomously drive background tabs on Mac and Windows for deep research, CRM data transfer, and admin-console workflows
OpenAI released ChatGPT for Google Sheets, an add-on that embeds natural language AI directly into spreadsheets via a sidebar panel
OpenAI also released three new audio models designed for developers building voice-enabled applications: GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper. OpenAI is explicitly pitching the new Realtime-2 model to product teams, describing it as an âinterface between people and productsâ
Anthropic released three new capabilities for Claude Managed Agents:
âDreamingâ is a scheduled review process that analyzes past agent sessions and memory to identify patterns (things like recurring mistakes, converged workflows, team preferences). It restructures memory to keep it high-signal as agents accumulate experience, enabling self-improvement across sessions
Outcomes define success criteria as a rubric. A separate grader evaluates the agentâs output against your standards in an isolated context window, then surfaces specific gaps if corrections are needed. The agent iterates until the work meets your bar. Internal testing showed 10-point improvements in task success, with +8.4% gains on Word documents and +10.1% on PowerPoint files
Multiagent orchestration lets a lead agent break complex work into parallel tasks and delegate each to specialist agents with custom models, prompts, and tools
Anthropic and OpenAI both launch enterprise AI ventures: Anthropic raises $1.5B with Blackstone and H&F for forward-deployed engineers, while OpenAI raises $4B for similar venture at $10B valuation
Google Docs now lets users set persistent custom instructions for Gemini that automatically apply to all future interactions. Instead of repeating preferences like âuse bullet pointsâ or âkeep it professionalâ in every prompt, users input these rules once in the Gemini side panel
Google is building OpenClaw-like agent Remy, powered by Gemini
Meta is also planning an OpenClaw competitor product for its massive user base
Spotify has released a Command Line Interface tool called âSave to Spotifyâ. It lets you connect tools like Claude Code generate podcasts from any text source and then store it to your library
Apple plans iOS 27 as choose-your-own-adventure of AI models, letting users pick preferred assistants. May offer custom Apple Wallet Passes beyond standard boarding passes & tickets.
Apple's camera-equipped AirPods enter early production testing. Prototypes with low-resolution visual sensors for Siri queries (cooking guidance, directions) now in design validation, with longer stems and cloud-indicator LED
Amazonâs new âBedrock AgentCore paymentsâ preview allows AI agents to âinstantly access and pay for what they use, such as web content, APIs, MCP servers, and other agents.â Weâre this close to being able to charge agents per-read fees to access content, and shake up the online content market for the rest of time
đ Good reads
The AI 100. CB Insights shares an interactive database with the most promising AI startups of 2026
AI godfather Yann LeCun pushes back on AI doom and hype. LeCun says CEOs exaggerate AIâs power and risk. He argues AI wonât wipe out 20% of jobs, but will turn most of us into âbossesâ of AI agents, making strategy and critical thinking more valuable. For PMs, his take is that this wave looks a lot like past tech revolutions: powerful new tools, not an imminent existential threat
Why nobody has nailed the personal AI agent yet. Peter Yang reviews OpenClaw, Hermes, Claude Code, Codex, and Gemini against ten must-have âAI chief of staffâ capabilities, and finds that none fully delivers. Each nails different pieces â from OpenClawâs flexibility to Hermesâ reliability, Claudeâs personality, Codexâs desktop experience, and Geminiâs Workspace access â but all stumble on reliability, coverage, or platform gaps. His takeaway for PMs: pick one or two agents that fit your workflows today, ship real work with them, and avoid endlessly migrating in search of a perfect stack that doesnât exist yet
Agent-native Product Management is changing PM work fast. Marcus Moretti explains how Product Management is starting to run everything through AI agents, especially Claude Code. He introduces two core âcompound engineeringâ skills: /ce-strategy to co-write and refresh a product strategy, and /ce:product-pulse to generate daily, founder-style health reports from your data stack
đ§ Good listen/watch
Uberâs CEO on agentic coding and why product people are making changes to Uberâs codebase. âNow, theyâre just going in and they are vibe coding, and an engineer is going to review the code, but essentially the product person is going directly into the code base, so to speak, or going directly with an agent into the code base.â Uberâs CEO on replacing policies with outcomes, why agentic coding is changing who writes the code, and the bet that there wonât be one physical AI model to rule them all
Why agency will matter more than skills in the AI era. Max Schoening (Head of Product at Notion) argues that as AI makes the first 10% of any project basically free, the key skill for PMs becomes agency, the willingness to tinker, prototype, and change the world around you. In a world where software volume is exploding but quality isnât, PMs who build taste, get their hands dirty with AI, and obsess over reliability and shipping will be the ones who stand out
đ§âđ» Worth learning
Cognitive Bias Index is an interactive reference for identifying and mitigating cognitive biases; quickly browse or filter dozens of documented biases by context and tag to understand how they arise, how they distort judgment, and how to avoid them; a free web tool curated by Jon Yablonski
30 Claude Code tips to seriously level up your workflow. Hannah Stulberg shares 30 tricks for using Claude Code as a real workbench. She shows how to customize Cursor, see and control what Claude is doing, and feed it cleaner context from docs, code, and the web. Also looks into longer-running, multi-device workflows (loops, mobile, remote control) so Claude can help you ship product work faster
Live coding session. In this live stream, Boris Cherny (Head of Claude Code) and Jarred Sumner (creator of Bun) pair up to build stuff with Claude Code using their everyday workflow. A practical tour of what âAI-nativeâ coding looks like in 2026, and how PMs can think about workflows
đ§ Products to try
Perplexity released Personal Computer for Mac, a desktop app that browses, takes actions, and runs research workflows on your behalf locally; free during beta
GPT-5.5 Instant is OpenAI's new ChatGPT default that delivers 52% fewer hallucinated claims on medicine/law/finance prompts and 30% more concise responses, free to try for all ChatGPT users
Saperly is the first phone carrier built for AI agents; provision a real number in seconds via any MCP-compatible agent for unified calling, SMS, and stable caller ID with audit trails; first number free for 30 days, then $2.50/month + usage
Thatâs a wrap for this week.
Feel free to drop your comments/questions/feedback. Would love to hear what youâd like to see more about in WiP, so I can make it better for you.
Have a great weekend!
