In partnership with
Sign Up | Advertise | Dispatches | Consulting
Hey Explorers!
For two years AI felt free... This week the invoice showed up.
Uber torched its entire annual AI budget before spring ended, OpenAI started gutting the chatbot that made it famous, and Washington began circling the labs for a cut.
The question, beyond what AI can do, is what it costs, and whether you can prove it paid off.
Here's the signal that matters 👇🏼
🗞️ TECH & TRENDS | The Bill Arrives
🧾 Uber Burned a Year of AI Budget in Four Months. Now It Can't Find the ROI. Uber rolled Claude Code and Cursor out to 5,000 engineers, gamified adoption with a usage leaderboard, and ran through its entire 2026 AI budget by April. Adoption worked perfectly. What nobody could draw was a straight line from all those tokens to anything a customer actually felt, and Microsoft just pulled its own Claude Code licenses over the same math. The lesson for everyone scaling AI: flat-rate pilots hide the bill, usage-based billing mails it to you. (🔗 Link)
🎨 Google Ran Its Whole Flagship Event Through Its Own AI, Then Published the Prompts Google built nearly every creative asset for I/O 2026 with its own stack, the title film, the brand identity, the event music, the speaker cards, even the swag, and then it showed the prompts and the workflow. The playbook is copyable. Feed the model your brand guidelines and a few years of past work, iterate against it, and generate on-brand assets at a scale no studio could match. This is the AI-native creative pipeline in the open, so steal it before your competitors read the same post. (🔗 Link)
🧠 ChatGPT Now Remembers Everything About You, Whether You Tell It To or Not OpenAI rolled out a memory upgrade it calls "dreaming" that learns your preferences, projects, and constraints in the background and carries them into every new chat, no "remember this" required, with a summary page to see and edit what it knows. The convenience is the cover story. The real move is lock-in. An assistant that has quietly accumulated years of context on you is one you will not casually walk away from, and OpenAI is counting on exactly that. (🔗 Link)
💬 OpenAI Is Rebuilding ChatGPT Around Everything Except Chat A senior OpenAI employee put it flatly: chat is dead. The company is rebuilding ChatGPT around agents and Codex, nudging you toward coding tools, image generation, and partner apps from Canva and Booking.com, then planning to drop the prompts entirely once the model can just read your intent. The chat box you built your AI habits around is quietly becoming a doorway to something else, and any strategy that assumes it stays put has an expiration date. (🔗 Link)
🏛️ Washington Wants to Own a Piece of OpenAI, and OpenAI Offered Trump confirmed from Air Force One that the government is in talks to take equity in the major AI labs, and the twist is OpenAI pitched the idea itself, offering to donate shares to seed a "Public Wealth Fund" Altman first floated in early 2025. Bernie Sanders wants to go further and take 50%. When left and right agree the government should own a slice of your AI suppliers, those suppliers have stopped being startups and become strategic infrastructure, and your vendor risk just changed shape. (🔗 Link)
🔒 OpenAI Shipped a Panic Button for the Agent Era OpenAI launched Lockdown Mode, which switches off live browsing, web images, deep research, and agent mode to blunt prompt-injection attacks that hide malicious instructions inside the content your AI reads. OpenAI concedes it is not a complete fix and built it for teams handling sensitive data. The subtext matters for everyone. Every capability you hand an agent is also an attack surface, and the more you let it act on your behalf, the more there is to hijack. (🔗 Link)
AI is about to become a metered utility. Always on, billed by the token, impossible to run a company without. Altman is already saying it out loud, that intelligence gets sold on a meter. The budget that wrecked Uber's year is just the first draft of a line item every company will carry.
So here is the question nobody is costing out yet. When every workflow runs on metered AI, what is your return per token, and can you defend that number to a CFO?
The winners over the next year won’t be the companies using the most AI. They’ll be the ones who know precisely what each dollar of it buys.
That is the work, and it is what we do. Purple Horizons helps leadership teams turn scattered AI spend into a system with a return you can point to. Strategy sprints, executive briefings, buildathons, and hands-on deployment.
Signal over noise. Always.
Ralph & Gianni
Six people doing the work. Your headcount is one.
Your finance close runs in #finance. Stripe and QuickBooks reconciled, runway updated, posted Sunday night without you asking.
Engineering review lands in #eng. Viktor pulled the open PRs, left comments on auth-refactor, flagged a dependency blocking api-pagination.
Campaign brief lands in #growth: Meta CPA up 18%, recommendation to pause broad match, a draft landing page already deployed for the variant test.
You hired him on day zero. He lives in Slack and Microsoft Teams alongside your contractors and investors, connects to 3,000+ tools, pushes back when you ship something dumb.
"Viktor is now an integral team member, and after weeks of use we still feel we haven't uncovered the full potential." Patrick, Director, Yarra Web.
📅 LET’S HANG OUT | Upcoming Events
📍 7/7 South Miami Tech Tuesdays @ Bougies
📍 7/14 Cafecito & Pastelitos @ Chug’s Diner