ChatGPT reached 1 billion global monthly active users in May, roughly three years after launch, based on Sensor Tower data. The app is growing 62% year over year, widening its lead over rival AI chat apps.
9:28 PMPYMNTS10 stories from 5 sources
AI scales into platforms as Microsoft builds its own stack
Day’s Recap
Supporting Articles
Microsoft unveiled MAI-Thinking-1, a new flagship in-house reasoning model, alongside other internal models at Build 2026. The launch extends Microsoft’s move away from exclusive reliance on OpenAI after the companies revised their partnership terms.
2:12 PMThe VergeMicrosoft is launching Scout, an always-on AI assistant built on OpenClaw that works across Microsoft 365 apps like Outlook, OneDrive, and Teams. It is positioned as an assignable virtual assistant for employees, handling tasks such as calendar organization, expense reporting, and drafting emails.
2:00 PMThe Verge9:28 PMPYMNTS
ChatGPT Hits 1 Billion Users Faster Than Any App in History
Summary
ChatGPT reached 1 billion global monthly active users in May, roughly three years after launch, based on Sensor Tower data. The app is growing 62% year over year, widening its lead over rival AI chat apps.
Why it matters
A billion monthly users turns ChatGPT into a de facto interface layer for information and software, reshaping competition, monetization, and policy around consumer AI.
2:12 PMThe Verge
Microsoft’s first advanced reasoning AI is here
Summary
Microsoft unveiled MAI-Thinking-1, a new flagship in-house reasoning model, alongside other internal models at Build 2026. The launch extends Microsoft’s move away from exclusive reliance on OpenAI after the companies revised their partnership terms.
Why it matters
Owning a flagship reasoning model lets Microsoft set its own roadmap and economics for AI products instead of inheriting them from OpenAI.
2:00 PMThe Verge
Microsoft Scout is a new AI personal assistant built on OpenClaw
Summary
Microsoft is launching Scout, an always-on AI assistant built on OpenClaw that works across Microsoft 365 apps like Outlook, OneDrive, and Teams. It is positioned as an assignable virtual assistant for employees, handling tasks such as calendar organization, expense reporting, and drafting emails.
Why it matters
An always-on Microsoft 365 agent could make AI delegation a default workflow layer, forcing enterprises to resolve permissioning and compliance at scale.
Other Developments
A curated list of other prominent stories from this day.
Microsoft released Adaptive Spec-driven Scoring for Evaluation and Regression Testing, an open source framework that lets developers generate AI evaluations from plain-language specs. The tool aims to make it easier to define, run, and repeat behavior tests as models and prompts change.
3:02 PMTechCrunchPerplexity says it is chip agnostic and appeared with Intel at Computex to unveil what it calls a hybrid local/server agentic inference orchestrator. The company framed the system as a way to run agentic workloads across on-device and server infrastructure without being tied to a single hardware vendor.
1:28 PMBloomberg MarketsOpenAI released six job-specific plug-ins inside the Codex app for tasks like data analytics, creative production, sales, product design, equity investing, and investment banking. Each tool packages integrations, instructions, and context so Codex can behave more like a role-focused worker than a general assistant.
12:00 PMTechCrunchAnthropic is expanding Project Glasswing to provide Claude Mythos and a security vulnerability workflow to 150 organizations across more than 15 countries. The program targets critical infrastructure operators in power, water, healthcare, and communications where disruptions could impact up to 100 million people.
10:44 AMTechCrunchGemini Spark demonstrates a trip-planning workflow that goes beyond chat by autonomously researching options, assembling an itinerary, and executing multi-step tasks with minimal user input. The experience feels unusually capable and smooth compared with prior AI travel demos, while also raising alarms about how much initiative the system can take on a user’s behalf.
8:38 AMThe VergeZeroDrift raised $10 million to build an AI compliance layer that sits between models and end users to detect risky outputs and substitute safer responses. The product targets regulated use cases where a model can create legal, privacy, or policy violations at the point of interaction.
8:32 AMTechCrunchUtopai Studios released version 2.0 of PAI, its generative video storytelling platform, less than three months after the product debuted. PAI is positioned as an end to end creative tool that lets users iterate on video at each production step to create visual stories with minimal traditional production resources.
8:00 AMVariety3:02 PMTechCrunch
New Microsoft tool lets devs spin up AI behavior tests using text descriptions
Summary
Microsoft released Adaptive Spec-driven Scoring for Evaluation and Regression Testing, an open source framework that lets developers generate AI evaluations from plain-language specs. The tool aims to make it easier to define, run, and repeat behavior tests as models and prompts change.
Why it matters
Better, cheaper evals make AI deployments safer to scale and harder to ship without measurable guardrails.
1:28 PMBloomberg Markets
Perplexity Is 'Chip Agnostic,' Says CEO
Summary
Perplexity says it is chip agnostic and appeared with Intel at Computex to unveil what it calls a hybrid local/server agentic inference orchestrator. The company framed the system as a way to run agentic workloads across on-device and server infrastructure without being tied to a single hardware vendor.
Why it matters
If orchestration makes model execution portable across chips, it weakens hardware lock-in and reshapes pricing and deployment choices for agentic AI.
12:00 PMTechCrunch
OpenAI launches new Codex tools for white-collar work
Summary
OpenAI released six job-specific plug-ins inside the Codex app for tasks like data analytics, creative production, sales, product design, equity investing, and investment banking. Each tool packages integrations, instructions, and context so Codex can behave more like a role-focused worker than a general assistant.
Why it matters
Role-packaged AI tools accelerate enterprise deployment by selling outcomes, but they also concentrate risk where errors carry financial and legal consequences.
10:44 AMTechCrunch
Anthropic scales Claude Mythos to critical infrastructure in 15+ countries
Summary
Anthropic is expanding Project Glasswing to provide Claude Mythos and a security vulnerability workflow to 150 organizations across more than 15 countries. The program targets critical infrastructure operators in power, water, healthcare, and communications where disruptions could impact up to 100 million people.
Why it matters
If AI meaningfully accelerates vuln discovery and triage in critical infrastructure, it can reduce systemic cyber risk, but it also raises new governance and security requirements for high-consequence environments.
8:38 AMThe Verge
Gemini Spark is the most impressive and terrifying AI experience I’ve had yet
Summary
Gemini Spark demonstrates a trip-planning workflow that goes beyond chat by autonomously researching options, assembling an itinerary, and executing multi-step tasks with minimal user input. The experience feels unusually capable and smooth compared with prior AI travel demos, while also raising alarms about how much initiative the system can take on a user’s behalf.
Why it matters
If agentic AI becomes reliable enough to run real transactions, the next fight moves from model quality to control, accountability, and safety rails.
8:32 AMTechCrunch
ZeroDrift raises $10M to protect AI models from themselves
Summary
ZeroDrift raised $10 million to build an AI compliance layer that sits between models and end users to detect risky outputs and substitute safer responses. The product targets regulated use cases where a model can create legal, privacy, or policy violations at the point of interaction.
Why it matters
Runtime compliance middleware is emerging as a must buy category as companies deploy AI into regulated workflows.
8:00 AMVariety
AI Firm Utopai Studios Launches Enhanced 2.0 Version of Generative Video Platform
Summary
Utopai Studios released version 2.0 of PAI, its generative video storytelling platform, less than three months after the product debuted. PAI is positioned as an end to end creative tool that lets users iterate on video at each production step to create visual stories with minimal traditional production resources.
Why it matters
Faster iteration in generative video tools shifts leverage toward teams that can produce more content with fewer people, reshaping creative economics and competition.
Make it yours
Build Your First Pass.
Pick your topics, set your cadence, and receive your personalized First Pass in your inbox. It’s that simple!