Good morning {{first_name| Artificial intelligence lovers}}. Google’s announcements at I/O gave us a clear picture of the company’s direction: a comprehensive approach to making life easier with AI.
We sat down with Google CEO Sundar Pichai to understand what this shift means for creators, engineers, and everyday users – touching on everything from YouTube to agent programming, and why today’s AI will look like a flip phone in three years.
-
Exclusive insights from Sundar Pichai at I/O 2026
-
The latest wave of Codex upgrades for OpenAI
-
Create a native CLI for the proxy from any website
-
California moves to protect workers affected by artificial intelligence
-
4 new tools for AI, community workflow, and more
Rundown: we He sat down With Google CEO Sundar Pichai on I/O 2026 For an exclusive interview about the company’s AI push, its efforts to empower YouTube creators, users, and engineers, and what AI will look like three years from now.
-
Models like the Omni enable creators to express themselves better, but YT will still be creator-first, while maintaining its “human-to-human” connection, Pichai said.
-
When asked why an everyday user switched to Gemini, Pichai pointed to the seamless integration into everyday life, especially with agents handling tasks.
-
He added that agents will work 24/7 across devices, and will be very common three years from now, when today’s technology seems as primitive as flip phones.
-
Pichai said engineers will have a team of agents, and the measure of success will not be code written by AI, but rather coding an agent, to handle long-running tasks.
Why it matters: The full interview covers more about Google’s AI efforts, Pichai’s advice for the younger generation, and where he believes the human presence will always be important. But the broad outlines are clear: the shift is sharp, and getting used to these tools will be key – for creators, programmers, and everyone else.
Along with Algolia
Rundown: Just like any powerhouse, an AI-powered research group needs a solid foundation on which to grow. If you’re not sure where to start or where to go next, Algolia has got you covered with a new white paper that shows you how to build a doc that stays productive from day one.
-
Assimilation, enrichment and hybrid indexing for better retrieval
-
RAG interfaces that connect your data to LLMs
-
Monitoring and governance using secure API keys
-
Manage the lifecycle as your research needs evolve
Obinay
Rundown: OpenAI decreased Another wave of Codex upgrades – this time improving the proxy assistant with the ability to attach application windows, a new target mode, use locked computer, and advanced annotations to work on the website without interfering.
-
Appshots allows Mac users to attach any open application window (screenshot, text, content) to a Codex thread with a simple click. Command-command He presses.
-
Goal Mode-now available in the Codex application, IDE extension, and CLI-allows users to set a goal and let Codex work toward it for hours (or even days).
-
Use Locked PC, when enabled, allows Codex (running via a second device) to use desktop applications, even after your Mac is locked with its screen off.
-
Advanced annotation mode allows users to describe directly to the manuscript what they want to change in their web page, and instantly previews the results.
Why it matters: OpenAI continues its efforts with Codex, giving developers not only more usability, but also useful features that improve how they work with the agent assistant and the context it can access. The company is betting that these efforts will help it close the gap with Anthropic, and maintain its advantage over xAI and Google.
Artificial intelligence training
Rundown: In this guide, you’ll learn how to use Printing Press to download proxy-friendly tools for hard-to-use APIs like Google Flights or ESPN. You will also learn how to create your own CLI tools to access any website using an Application Programming Interface (API).
-
Go to Printing press Browse examples or read github Install notes. Now, install the starter pack: npx -y @mvanhorn/printing-press install starter-pack
-
Tell your agent (Codex/Claude Code/Hermes): “Examine the launch toolkit. List the tools, read their help output, and try a safe, read-only command from ESPN or Flight-goat. Tell me which tool worked and which command you ran.”
-
Have your dealer install the printing press binary using: go to install github.com/mvanhorn/cli-printing-press/v4/cmd/printing-press@latest
-
Now the agent can create its own CLI tools. Start with a site that contains a lot of public data. “Use the printing press to create read-only commands for [WEBSITE]. Suggest useful commands first, then build and test them.
Pro Tip: If your agent is having difficulty using a specific MCP or integration, you can ask them to create their own orders using Printing Press. You just need an API key.
Provided by AWS Marketplace
Rundown: Multi-agent systems fail in predictable ways without a coordination layer-state is lost between steps, no human logging on critical decisions, and silent failures when one agent goes down. The AWS workshop on May 26 shows you how to build a control plane that prevents this.
-
Manage state and retry using AWS Step Functions
-
Coordinate thinking and use of tools through Amazon Bedrock agents
-
Add human approval gates that pause execution until logout
-
Handle multi-hop pipeline failures gracefully with Amazon MWAA
Artificial intelligence protection
Image source: Images 2.0 / Rundown
Rundown: California Governor Gavin Newsom I fell Executive order directs state agencies to study and develop policies on how to protect workers from job losses due to artificial intelligence – one day after Meta Demobilization 8 thousand employees to compensate for artificial intelligence investments.
-
The order directs agencies to explore policies such as severance standards, equity compensation, worker ownership models, and universal equity capital.
-
Within 90 days, the state will launch a dashboard to track the impact of AI on jobs, and within 180 days, agencies will display WARN Act updates for faster layoff alerts.
-
By October 15, the state will review how unions negotiate AI adoption, modernize workforce training, and explore ways to direct AI revenue toward public benefit.
Why it matters: California is home to 33 of the world’s top 50 AI companies, and is now the first state to formally examine what the technology might mean for workers and the economy. It’s hard to ignore the timing: it’s over 70 thousand jobs It has already disappeared in 2026, and the industry expects further reductions in the future as AI adoption accelerates.
-
🤖 Scrunch – Learn how AI interprets your website, performs a free audit, and unlocks the new way to reach customers*
-
🦞 NanoClaw – A secure, lightweight, open source alternative to OpenClaw
-
💻 Antigravity – Google’s next generation agent development platform
Figure AI co-founder Brett Adcock Hark He grew up $700 million at a $6 billion valuation to build its own personal intelligence, using original models, software, and hardware.
Anthropic He is It is said It is in talks to use Microsoft’s purpose-built Maia AI chips, following similar deals with Google for its TPUs and Amazon for its Trainium chips.
Agent AI Getting Started Guide He is Explore optionsincluding an increase of about $1 billion, in compliance with China’s request to unbundle Meta’s $2 billion acquisition, Bloomberg reported.
SpaceX IPO prospectus open Which Anthropic is paying the company $1.25 billion a month through 2029 for access to computing capacity via Colossus and Colossus II.
Microsoft and EY Announce New $1 billion partnership, Microsoft pairs its AI engineers with 400,000 EY consultants to accelerate clients’ AI projects.
In each newsletter we show how the reader is using AI to work smarter, save time or make life easier.
Today’s workflow comes from the reader Loic R. In London:
“I tore the ACL in my knee years ago while playing soccer. As a runner, I have never done long-distance running. But in one month, I will be participating in a 5-hour multi-sport adventure race. To train for this, I asked ChatGPT to create a custom 4-week training program taking into account my past injuries and current strengths.
After each running session, I use audio to record what I’m doing and how my body feels. ChatGPT then adjusts my program based on the time remaining before the race and how to maximize it while minimizing the risk of injuries. A personal trainer in my pocket for a very specific challenge.“
How do you use artificial intelligence? Tell us here.
That’s all for today!
Before you go, we’d love to know what you thought of today’s newsletter to help us improve The Rundown experience for you.
Rowan, Joey, Zack, Shubham and Jennifer – the humans behind The Rundown
