Happy Saturday Fam! And for those in the US Happy Memorial Day!

Big week. Elon Musk's $134B lawsuit against OpenAI collapsed in under two hours of jury deliberation. Google bet its entire I/O keynote on putting Gemini on your face. And a single customer ordered 18,000 cups of water at a Taco Bell drive-thru and broke the AI taking the order. Let's get into it.

In this edition:

Own a tool that you'd like to see on our site or newsletter? Please Submit a tool or reach us by replying to this email or at [email protected]

Google spent its keynote showing Gemini executing tasks on your behalf, and Anthropic's PwC deal is putting Claude agents in front of hundreds of thousands of professionals. The clearest near-term place that "AI that acts, not just chats" pays for itself? Inbound sales. Our collection breaks down seven AI SDR tools that qualify leads, draft replies, book meetings, and keep your pipeline warm while your reps sleep.

Explore the Collection →

Google I/O 2026 was, functionally, a DeepMind keynote. The Nobel laureate and DeepMind CEO has spent a decade arguing that AI should be a tool for understanding the world, not just chatting about it — and this week that thesis showed up as hardware. Gemini now powers Google's first consumer intelligent eyewear, sees through a camera, answers about what's in front of you, and runs agentic tasks across Android. Hassabis has been clear he sees assistants like this as a step toward AI that genuinely perceives context rather than just predicting text. Whether you want a camera on your face is another question — but the man who brought you AlphaFold is now, unmistakably, shipping consumer products.

Two stories defined the week, and together they tell you exactly where the industry's center of gravity sits right now.

Monday — the verdict. A federal jury in Oakland took less than two hours to dismiss every claim in Elon Musk's lawsuit against Sam Altman and OpenAI. The advisory jury found Musk's "breach of charitable trust" case was filed outside the three-year statute of limitations; Judge Yvonne Gonzalez Rogers immediately adopted the verdict. The court never ruled on whether OpenAI actually betrayed its nonprofit mission — it didn't have to. Musk called it a "calendar technicality" on X and vowed to appeal. After three weeks of testimony, ugly cross-examinations, and read-aloud private messages, the highest-profile fight in AI ended on a deadline math problem.

Tuesday–Wednesday — the glasses. Google used nearly its entire I/O keynote to make one argument: Gemini shouldn't live in an app, it should live on your body. The headline reveal was intelligent eyewear — Gemini-powered audio glasses built with Samsung, Warby Parker, and Gentle Monster, running Android XR, with cameras, speakers, and mics but no in-lens display. Tap the frame or say "Hey Google" and Gemini answers questions about whatever you're looking at, or executes tasks for you. They work with iPhone out of the box — the first Android XR glasses to ship cross-platform — and arrive this fall. Google paired it with a major Gemini app upgrade ("Gemini Spark"), Android 17, and more detail on the Googlebook laptops replacing Chromebooks.

Here's the throughline. A week after Anthropic and OpenAI argued over who's allowed to deploy dangerous models, Google answered a different question: where does AI live? Its bet is everywhere — your laptop (Googlebook), your phone (Gemini Intelligence), and now your face. The interface war has officially moved off the screen.

And the money keeps validating it. Anthropic's round — at least $30B at a $900B+ valuation — is expected to close by month's end, which would push it past OpenAI's $852B March valuation for the first time. The same week, Anthropic announced a sweeping PwC alliance (rolling Claude Code and Cowork out to hundreds of thousands of professionals) and a $200M Gates Foundation partnership. The eye-watering footnote: Anthropic is reportedly paying xAI $1.25 billion a month just for compute. The labs aren't competing on models anymore. They're competing on distribution, real estate, and raw electricity.

Taco Bell vs. 18,000 cups of water.

A customer pulled up to a Taco Bell drive-thru, ordered eighteen thousand water cups, and the AI ordering system — deployed across 500+ locations — dutifully tried to process it before melting down. The clip went viral, hit the front page of Hacker News, and Taco Bell's CTO publicly admitted the drive-thru AI "might need to go away" at busy locations. Parent company Yum Brands is still working with Nvidia on the tech, but for now: humans take over when the lines get long.

The lesson every AI product team relearns monthly: your system will be tested by the single most adversarial 19-year-old in a four-county radius, and he has all night.

Our AI Graveyard sits at 166 entries and counting — 55 of them dead in 2026 alone, the worst year on record. This week's featured plots lean toward the "acquired & folded in" outcome — the quieter way a tool dies:

Browse the full graveyard →

Google's intelligent eyewear is the most exciting and most unsettling reveal of the week. Audio-first glasses with a camera and Gemini listening means the privacy question is no longer "what does the company collect about me" — it's "what does it collect about everyone standing near me who never opted in."

Google Glass died in 2013 partly over exactly this; the term "Glasshole" entered the dictionary. A decade later the tech is better, the assistant is genuinely useful, and the social contract still hasn't been written. Worth watching this fall: whether Google ships a visible recording indicator, how bystander consent gets handled, and whether "Hey Google, what am I looking at" quietly becomes "Google always knows what you're looking at." The consent-credit-compensation framework Spotify and UMG built for AI music this week is a useful contrast — someone in music decided the rules up front. Nobody has for the glasses yet.

We don't think AI eyewear is bad — we think the norms should arrive before the hardware does, not after.

Before you commit to any plan — a launch, a hire, a big purchase — run this. It's the single best way to surface the failure you're not seeing.

Works on everything from product launches to "should we lease this office." The "cheap insurance" line is where the real value is.

A rotating shelf of the AI books, podcasts, and courses we'd point a friend to:

Got a resource we should feature? Reply and tell us.

This week gave us a real one. PwC announced it's rolling Claude Code and Cowork out to hundreds of thousands of its professionals and certifying 30,000 in the US on Claude — one of the largest single deployments of agentic AI inside a professional-services firm to date. It's a useful signal for everyone watching from smaller orgs: the bet isn't "AI replaces consultants," it's "consultants who wield agents out-deliver the ones who don't." If a 360,000-person firm is standardizing on it, the "wait and see" window is closing.

Have an AI win — big or small — at your company? Reply and we may feature it here.