Every Inc.

Most tech CEOs say they’re “AI-first.” Aaron Levie, CEO of Box, is evolving a 20-year-old cloud pioneer to be one. On the latest AI & I with Dan Shipper, he shares what the transition looks like on the inside—how culture changes, workflows evolve, and mindsets adapt. If you’ve ever wondered what “AI-first” looks like from the inside, this is your blueprint. Watch or listen below.

GPT-5 now beats human experts on almost 50% of real-world tasks. But that doesn’t mean your job is going away. As Dan Shipper writes, every AI breakthrough hides a layer of “smuggled intelligence”: the human judgment, prompting, and feedback that makes it all work. Real progress in AI doesn’t remove people from the loop—it depends on them. Read Dan’s latest piece 👇

OpenAI’s third DevDay wrapped with three main drops. New apps inside ChatGPT. A new agent builder. New models. But this year’s vibe was less “magic show,” more “product launch.” The highlights: • Apps SDK lets partners like Figma, Zillow, Spotify, and Canva build inside ChatGPT • AgentKit helps companies create AI workflows with a visual, no-code builder • GPT-5 Pro, Sora 2, and a new voice model are now in the API “OpenAI is coming for all of the developer-adjacent use cases becoming popular with AI operations,” says Dan Shipper. “It isn’t so much an agent builder as it is a workflow builder, where a business user can define a specific end-to-end process for the AI to follow.” Codex, meanwhile, “announced almost nothing.” “DevDay was a little boring for developers—but extremely exciting if you’re an AI-ops pro,” says Dan. “I’m not convinced this is yet the app-store moment, but I like that they’re continuing to try.” “Generative AI is the closest thing we have to magic,” adds Alex Duffy. “But today felt like pulling back the curtain to explain how it’s done.” 800 million weekly users suggest OpenAI might not need a “wow” moment anymore.

Claude’s getting faster. Sparkle got rebuilt. And AI is learning to use the internet. Here’s what’s inside this week’s Context Window: - Dan Shipper on vibe-checking Claude Sonnet 4.5 - Yash Poojary on turning Sparkle into an AI command center - Rhea Purohit with Alex Rattray on building the “plumbing” that lets AI use the web - Chris Silvestri on making AI write less like AI (and more like you) - Dan Shipper on why intelligence is really pattern recognition - Dr Ashwin Sharma reads on the rise of “accountability sponges”

I recently came across this brilliant blog on the shift from knowledge to allocation economy, written nearly 2 years ago, yet remarkably prescient given today's agentic workforce revolution. Here's the paradigm shift: It's no longer about what you know, but how you allocate intelligence. In the knowledge economy, competitive advantage came from hoarding information and hiring the smartest people. That playbook is becoming obsolete faster than yesterday's tech stack. In the allocation economy, winners are determined by: - Strategic resource deployment: How effectively you orchestrate AI capabilities. - Intelligent task distribution: Knowing when to leverage human creativity vs. AI efficiency. - Invisible optimization: Using AI to silently outpace competitors in speed, cost, and quality. The timing was extraordinary. Dan Shipper wrote this before AI agents were even a blip on most radars. Now, with agentic workflows transforming everything from customer service to software development, his prediction feels like looking into a crystal ball. My take: Within 24 months, market leaders will emerge not from traditional R&D powerhouses, but from companies that excel at AI orchestration. The competitive moat won't be proprietary algorithms; it'll be proprietary allocation strategies. The most successful leaders of the next decade won't be those who understand every technical detail. They'll be the ones who can envision, design, and execute intelligent resource allocation at scale. The question isn't whether this shift is coming; it's whether you're positioning yourself to thrive in an allocation-first world. What's your take? Are you seeing this transition in your industry?

Sora brought Dan Shipper’s latest Every piece to life on screen. The result? Pure cinema. 

“Language models let us see the world as it really is: complex, interconnected, and impossible to reduce to simple rules.”—Dan Shipper That’s the core argument in the latest piece for Every’s Chain of Thought—on the worldview Dan Shipper has developed by writing, coding, and living with AI. Western rationality gave us rockets, medicine, computers, and smartphones. But it also left a blind spot: intuition. Large language models point beyond rules to a new worldview: -Reality as a web, not a chain -Meaning through contrasts, not definitions -Knowledge as participatory, not objective This shift matters for science, business, and creativity. It reminds us that uncertainty isn’t a flaw to eliminate—it’s the feature that drives discovery.

Happening today: Claude Code Camp: non-coder edition—a live session built for writers, founders, marketers, and operators. You’ll learn how to use Claude to: - Create marketing docs + support copy with Nityesh Agarwal - Use slash commands to automate tasks with Kieran Klaassen - Crunch data + draft posts with Katie Parrott - Do research + manage expenses with Dan Shipper 500+ joined our last Camp. Many called it the most practical AI workshop they’ve attended. 📅 Today — Friday, October 3rd, 12 p.m. ET 🎟️ Exclusive to Every subscribers Every subscribers can RSVP here: https://lnkd.in/g8phzibZ Not a subscriber yet? Join today at the same link.

Sometimes the hardest part of building isn’t starting from scratch. It’s inheriting someone else’s idea — with all its complexity. That’s exactly what Yash Poojary did with Sparkle. ✨ What began as a tangled codebase, passed through multiple hands, became the foundation he rebuilt into Sparkle Search — an AI-powered command center for your Mac. Full story below:

APIs are the invisible infrastructure that powers every email you send, every payment you process, and every login you complete. They're the handshakes between programs that make the internet work. Now AI needs its own kind of plumbing. Alex Rattray, founder of Stainless, joined Dan Shipper on the latest episode of AI & I to talk about MCPs (model context protocols)—the servers that let language models plug into software and get work done. The challenge: They don't work as well as they should yet. At Stainless, which builds robust SDKs that turn APIs into developer-friendly tools, Rattray has built MCP connections to Notion and HubSpot. He can ask: "Which interesting customers signed up last week?" The system queries the database, cross-references HubSpot, pulls notes from Notion, and delivers a summary. It's not flawless yet. But it's transforming work that normally requires multiple logins and manual searches. If you're building with AI or just trying to understand how it will integrate into your workflow, this conversation is worth your time.