The Race Just Went Public

In partnership with

Opening

Google used I/O to unveil Gemini 3.5 plus Gemini Spark, an always-on agent in the Gemini app that takes action across your connected services (CNBC). OpenAI is targeting a September IPO with the S-1 filing expected this summer (TechCrunch). Anthropic announced a global alliance with KPMG, putting Claude in front of 276,000 employees across 138 countries (Anthropic). Each of the three biggest AI labs made their largest move of the year before lunch on the same day.

The operators paying attention right now are going to build something the people waiting for the dust to settle will have to play catch-up with for the next year.

Turn AI into Your Income Engine

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential
Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background
Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

Today's Signals

Google unveiled Gemini 3.5 and Gemini Spark at I/O. Spark is an always-on personal AI agent in the Gemini app that reasons across connected apps and takes action on your behalf. In beta for Google AI Ultra subscribers and trusted testers starting next week. The 3.5 lineup also ships Omni, a world model that simulates physical environments inside Flash, Gemini App, Flow, and YouTube Shorts. (CNBC)
Anthropic announced a global alliance with KPMG, embedding Claude across the 276,000-person professional services firm. Claude goes into Digital Gateway, the platform KPMG's tax and advisory clients use for billable work, plus PE-focused product offerings and a cybersecurity workflow. (Anthropic)
OpenAI is targeting September for its IPO, according to people familiar with the plans. The company would file its S-1 this summer, following the January close of a $40B funding round. (TechCrunch)
Simon Willison tested Gemini 3.5 Flash: 3x the price of Gemini 3 Flash Preview, 6x the price of 3.1 Flash-Lite. Yet Google is rolling it out across free consumer products. Model ID: gemini-3.5-flash. (Simon Willison)
A hardware builder gave an OpenClaw agent a physical robot body, bypassing the sandboxed agent layer. The agent navigates physical space using the same tool-call loop it uses in software. (Wired)

$20.8B in Redemption Requests. Percent Was Issuing Deals and Paying on Schedule.

Those requests came from non-traded BDC investors in Q1 2026, and most got back roughly half of what they asked for. Moody's U.S. BDC sector outlook: Negative.

On Percent's marketplace that same quarter: new issuances, scheduled payments, 0.44% lifetime net loss rate on asset-based deals since inception.† The difference is structural: concentrated corporate loans with redemption windows that close at manager discretion vs. asset-based finance with 6–24 month deal terms. 14.6% net ABS returns LTM after losses (3/31/26).† Starting at $500.

See how private credit works when it's built for individual investors.

_{Alternative investments are speculative. No assurance can be given that investors will receive a return of their capital. †Past performance is not indicative of future results. Terms apply.}

The Drops

[Repo] campfirein/byterover-cli — portable memory layer for autonomous coding agents. ByteRover gives agents a persistent, structured memory store that survives session restarts, tool swaps, and context resets. Install: npm i -g byterover-cli. 4,800 stars. Your agent picks up where it left off after a context reset instead of starting fresh. (github.com/campfirein/byterover-cli)

[Repo] can1357/oh-my-pi — terminal coding agent with hash-anchored edits, LSP support, Python execution, browser access, and subagents. 5,300 stars. Every file modification gets pinned to a specific hash of the file state, so the agent cannot make a confident edit to a stale version. If you have watched an agent overwrite work it did not know it was overwriting, this is the fix. (github.com/can1357/oh-my-pi)

[Skill] wshobson/agents — multi-agent orchestration collection for Claude Code. 35,700 stars. Covers Haiku/Sonnet hybrid routing, pre-built slash commands, and multi-agent workflow templates you can drop in directly. The evergreen Wednesday pick. (github.com/wshobson/agents)

Real-World Ads, Simple to Run

With AdQuick, executing Out Of Home campaigns is as easy as running digital ads. Plan, deploy, and measure your real-world advertising effortlessly — so your team can scale campaigns and maximize impact without the headaches.

Visit AdQuick.com

The Stack

Exa's MCP server gives your agent live web search and full-page crawling through a single MCP connection. Wire it in once and any tool call in your agent can query the live web or pull a complete URL. 4,445 stars, maintained by the Exa team.

The reason this matters for Claude Code agents: Claude's built-in search is prompt-level. Exa MCP is tool-level. The agent decides when to search, what to search for, and what to do with the result. You do not wire a search function by hand. Install: npx exa-mcp-server. One API key in your MCP config. There is a .claude-plugin directory in the repo for the one-click install path.

(github.com/exa-labs/exa-mcp-server)

The Onboard

1. Request Gemini Spark beta access this week. Spark rolls out to Google AI Ultra subscribers and trusted testers starting next week. If you have an AI Ultra subscription, request access through the Gemini app settings. Spark reasons across Gmail, Calendar, Docs, and YouTube without prompt-by-prompt instruction. It is the first persistent personal agent shipping at consumer scale from a major lab. (CNBC)

2. ByteRover CLI drops into any Claude Code workflow in one step: npm i -g byterover-cli, then brv init in your project directory. ByteRover creates a .byterover/ memory store your agent reads and writes across sessions. Your agent picks up where it left off after a context reset. (github.com/campfirein/byterover-cli)

3. In CI and shared pipelines, skip --dangerously-skip-permissions. Put an explicit tool whitelist in your CLAUDE.md instead. List every tool your agent needs by name. Claude Code reads the file at startup and limits itself to that set. A runaway subagent cannot reach tools it was never supposed to touch. (Anthropic docs)

The Frame

All three frontier AI labs made their year's largest play this week. The plays were different.

Google went wide. Gemini 3.5 ships across free consumer products. Spark moves from concept to beta access in weeks. Omni, a world model that simulates physical environments, lands inside Flow and YouTube Shorts the same week. That is a flood, not a launch.

OpenAI is filing. A September IPO target puts the S-1 window in front of us this summer, faster than most growth-stage companies move. The read on that: OpenAI thinks the revenue can hold up under public-market scrutiny, and the AI agent market is more durable than the people who dismissed it in 2024 thought it was.

Anthropic went deep on enterprise. The KPMG alliance puts Claude inside Digital Gateway, the platform 276,000 tax and advisory professionals use for billable work in 138 countries. The two firms are also co-building products for PE portfolio companies, with a cybersecurity integration on top. That is not what a pilot looks like.

The operators reading these three moves as one signal will ship faster than the ones still waiting for the dust to settle. The signal: the AI race is now happening on three layers at once. Product, public market, enterprise. The floor moved again today.

Builder's Brief

Friday's kit is the Solo Operator Stack: a Command Center dashboard plus five named agents (Sales, Content, Ops, Support, Comms). The specific thing that surprises operators when they ship it: the entire stack runs for $80-$180 per month at solo scale. That includes Claude API, hosting, and every third-party integration the agents need. Compare that to $26 per seat for Notion AI plus $20 per seat for Linear plus the rest of the SaaS stack people pay for to fake agentic capability, and the math stops being close.

The kit ships the full Command Center deploy in ten numbered steps (clone to Vercel, wire your data sources, done) plus the five agent configs you drop into Claude Code one at a time. The operating layer that ties them together, the part Monday's brief flagged as missing in most setups, is what makes this a stack instead of five disconnected chat windows.

Full breakdown Friday for Operator Access subscribers.

Unlock Operator Access

Before You Go

Google flooded the product layer with Gemini 3.5 and Spark. OpenAI accelerated toward the public market. Anthropic locked in a 276,000-employee enterprise integration. Which of those three is playing the longest game?

Forward this to one person building with AI agents this week. They will want to know about Gemini Spark.

Get your referral link and track who you've sent this to: recs.page/the-aigent

See you Thursday.