Welcome, humans.
Mathematicians have a private inside joke called "The Book." It's the mythical, never-published volume in which God keeps the most elegant proofs ever written, the kind so clean and clever that math professors weep when they read them. Hungarian legend Paul Erdős used to say a proof "from The Book" was the highest compliment a theorem could receive. Most working mathematicians spend their entire career hoping to write one.
Yesterday, GPT-5.4 Pro wrote one. Mathematician Przemek Chojecki used it to crack Erdős problem #1196, a 60-year-old asymptotic primitive set conjecture that had only ever seen partial human progress. The kicker, per Yale mathematician Jared Lichtman? The proof is three pages long, uses a single elegant trick (a von Mangoldt function rearrangement), and bypasses the entire probability "gambit" implicit in every human attempt since Erdős's own 1935 paper. Lichtman called it "a Book Proof." AI just got published in God's notebook. Wild.
Oh, and stay tuned: rumor has it that both Claude Opus 4.7 and OpenAI's Spud should drop today, so today might be a two-model-launch kind of Thursday.
Pretty sure Opus 4.7 is imminent at least, because Claude always acts up right before a new model drops, and this week it’s been ROUGH. Much dumber atm. Messing up my skills, forgetting things… Usually a good sign of a better model coming online… HOPEFULLY, anyway.
Here’s what happened in AI today:
😼 Allbirds, the wool sneaker brand, is renaming itself "NewBird AI" and pivoting to GPU clouds; stock popped 600%.
📰 OpenAI's $852B valuation faces backer scrutiny while VCs flood Anthropic with $800B offers.
📰 A federal court ruled your AI chats have no attorney-client privilege.
🍪 Tasklet shipped 14 new triggers so your cloud agent reacts to Slack, Calendar, and Drive events instantly.
🎓 How to replace your n8n or Zapier workflow with a Claude routine in 5 minutes.
… and a whole lot more that you can read about here.
P.S: Want to reach 675,000 AI-hungry readers? Click here to advertise with us.
P.P.S: Love robots? We’re starting a new robotics newsletter! Sign up early here.

- 😼 2026 in a Nutshell: Allbirds Quit Saving the Planet to Power AI.
- A defining moment for AI compute
- 🎓 AI Skill of the Day: Replace Your n8n / Zapier Workflow With a Claude Routine in 5 Minutes
- 🎙️ New Episode: Peter Wilczynski on the gap nobody's talking about in AI.
- 🍪 Treats to Try
- 📰 Around the Horn
- What would you do with an extra 10 hours every week?
- 🧩 Thursday Trivia
- A Cat’s Commentary
😼 2026 in a Nutshell: Allbirds Quit Saving the Planet to Power AI.
Today in news I literally can’t believe is not an April Fools joke: Allbirds, the eco-conscious wool sneaker brand your yoga instructor wore in 2019 (and this writer wears to this day), has decided humans don't deserve nice shoes anymore. The company executed a $50M convertible facility to pivot to AI compute infrastructure and is renaming itself "NewBird AI".
What's next, Nike starts selling GLP-1s? Side note: is that why Tim Cook keeps buying Nike stock?
Anyway, here's what happened:
Shares popped 600-700% in a single morning.
The actual shoe brand was sold to American Exchange Group for $39M in March. No more shoes for NewBird.
All that remains is the public stock ticker, now worth 7x more once you whisper "AI compute."
The plan: use the $50M to go buy GPUs and resell compute as a GPU-as-a-Service provider.
The deeper irony. Allbirds is a Delaware Public Benefit Corporation and a Certified B Corp since 2016, with environmental conservation written into its corporate charter and a founding mission to "reverse climate change through better business." Now this same company, legally bound to weigh environmental impact in its decisions, is pivoting into one of the most electricity-hungry, water-cooled businesses humans have ever invented. Bro: a B Corp pivoting to GPU cloud is like a vegan restaurant pivoting to foie gras.
Why this matters: This isn't really an Allbirds story; it's a capital markets story. When a sneaker brand can 7x its market cap by uttering "GPU-as-a-Service," the economy is now pricing assets on AI exposure, not fundamentals. Nobody knows what anything is worth. All they know is AI = important. Everything else? Shrug emoji.
Our take: WTF man, we literally wear these shoes. Even though we cover AI every day, we did not sign up for our footwear brand to LARP as a colocation facility outside Reno.
Worth remembering: in December 2017, an iced tea company renamed itself "Long Blockchain Corp." The stock popped 200%, and the SEC later charged its insiders with fraud. The final act? Nasdaq delisted them within 18 months. RIP Allbirds. If I was a Polymarket man (which I am not), I doubt NewBird makes it past 2027. Unless, y’know… AI 2027 happens. Then NewBird might be OnlyBird because we’ll all be yolo’d into the machine future. Wild times, man.

FROM OUR PARTNERS
A defining moment for AI compute
At Arm Everywhere, Arm marked a major milestone for the compute platform: its expansion into production silicon with the Arm AGI CPU. Joined on stage by leaders from Meta and OpenAI, Arm outlined what this move means for the future of AI infrastructure. Built for the rise of agentic AI, this next step brings Arm into data center silicon for the first time—signaling a more integrated approach to performance, efficiency and scale across the AI stack.

🎓 AI Skill of the Day: Replace Your n8n / Zapier Workflow With a Claude Routine in 5 Minutes
Anthropic just shipped Claude Code Routines, and as Nick Saraev walks through in his demo, it is a 1-to-1 replacement for n8n, Make.com, and Zapier. The old way: drag-and-drop nodes, set up credentials, map variables, debug for hours. The new way: describe what you want in plain English, hit save, done.
Here's the 5-step setup for non-technical users:
Go to claude.ai/code/routines and click "New routine."
Name it (e.g., "Morning email triage").
Write the prompt like an SOP. Routines run hands-off, so be more precise than a normal Claude chat. Example: "Pull my Gmail unreads. For each one, check for prior conversations with that contact. Draft a reply. Send me the drafts in Slack."
Pick a trigger: schedule (daily at 5 AM), webhook (a URL that other apps can "ping" to fire the routine automatically; e.g., a new Gmail arrives → routine runs), or API call (an API is how programs talk to each other; "call" means any script or tool you run can kick off the routine on demand from anywhere).
5. Add connectors under Settings → Connectors. Sign in to Gmail, Slack, Notion, etc. (uses OAuth, the same "Log in with Google"-style flow you've clicked through a hundred times) so the routine can read and write in those tools.
Hit "Run now" to test. If it works, walk away. The routine fires forever, no laptop required.
Pro tip from Saraev: already have an n8n or Make workflow? Copy the JSON from your nodes (in n8n: shift-select then Cmd+C), paste into Claude Code, and ask it to "turn this into a Claude routine." It rebuilds the whole thing in 30 seconds.
Total AI beginner? Start here (goes with this video).
Have a specific skill you want to learn? Request it here.

🎙️ New Episode: Peter Wilczynski on the gap nobody's talking about in AI.
Frontier models can write code and crush math olympiads, but they have zero grounded understanding of the physical world. Peter, the Chief Product Officer at Vantor (formerly Maxar), is fixing that. His team built a 3D model of the entire planet at 50cm resolution and made it machine-readable, so AI agents can finally reason about where things actually are in space.
In our latest episode, Grant sits down with Peter to unpack what "spatial intelligence" really means, why it might be the missing piece for AGI, and how the same Rosetta Stone trick that made LLMs work on text could now work on the Earth itself. Plus: predictive world models, GPS-free navigation for drones and AR, and Peter's spicy bet that the physical world is about to become the new navigation layer for a digital world that's gotten too crowded.
🎧 Watch / listen on: YouTube | Spotify | Apple Podcasts

🍪 Treats to Try
Tasklet shipped 14 new event triggers so your cloud agent now responds the instant something happens in Slack, Google Calendar, Drive, Outlook, Telegram, YouTube, Apple Shortcuts, Notion, GitHub, or HubSpot. Try prompts like "brief me on every meeting 15 min before it starts" or "respond when someone in #ask-ai mentions Jarvis" (free tier; paid plans from $35/mo).
Gemini for Mac is Google's brand-new fully-native Swift app that lets you share your screen or local files with Gemini in real time and get help with whatever's on it (100+ features built in 100 days, per Josh Woodward) —free with a Google account.
Gemini 3.1 Flash TTS is Google's new text-to-speech model with scene direction, per-speaker voice control, inline audio tags for emotion and pacing, and 70-language support —free to try in AI Studio's audio playground.
Midjourney V8.1 renders images natively in 2K HD at 3x the speed of V8 and 3x cheaper, with image prompts back, a new Describe tool, moodboards, and srefs restored—subscriptions start at $10/mo.
Tubi became the first streaming service with a native ChatGPT app; install it from the ChatGPT app store and type @Tubi to ask things like "a movie that feels like a fever dream but isn't horror" and get curated picks from 300,000+ titles —free.
ACE-Step 1.5 is a new open-source AI music model that turns any song description into a finished track in seconds; Victor M's free demo on Hugging Face is the fastest way to try it —free.
OpenRouter Video Generation routes one API call to the top video models alongside text, images, audio, embeddings, and rerankers in the same gateway —usage-based pricing per model.

📰 Around the Horn
OpenAI's $852B valuation faces investor scrutiny per FT and Reuters, with one backer telling the FT it requires assuming an IPO valuation of $1.2T or higher. Meanwhile, VCs are reportedly flooding Anthropic with offers as high as $800B.
Apple is shipping its Siri programmers to a multi-week AI coding bootcamp, per The Information; fewer than 200 people, two months before the expected major Siri revamp. Even Apple's voice assistant team needs to vibe-code now.
A federal judge in S.D.N.Y. ruled there is no attorney-client privilege for AI chats in US v. Heppner, prompting nationwide lawyer warnings to stop treating chatbots like trusted confidants when freedom or liability is on the line.
US federal agencies are openly skirting Trump's Anthropic blacklist to test the Mythos cyber model, while Politico Europe reports European cyber agencies have been almost entirely shut out of Project Glasswing.
Google DeepMind released Gemini Robotics-ER 1.6, an upgraded "embodied reasoning" robot brain that now reads complex gauges and sight glasses with 93% accuracy (up from 23% in v1.5), built in partnership with Boston Dynamics for Spot facility inspections; available now via the Gemini API and Google AI Studio.
📺 MUST WATCH: Dwarkesh Patel grilled Jensen Huang on TPU competition, why the US should sell AI chips to China, and NVIDIA's supply chain moat… worth it if you want the CEO's case for his own empire in his own words.

FROM OUR PARTNERS
What would you do with an extra 10 hours every week?
Build AI agents and assistants to multiply your work and life – no tech skills required. The AI Executive System walks you step-by-step through building AI assistants for your professional and personal life — from a frontline "office team" to a strategic advisory council — with live coaching, proven templates, and a peer community behind you.

🧩 Thursday Trivia
You know the drill. One is AI, and one is real. Which is which? Vote below!
A.

B.

Which is AI, and which is real?Which is AI, and which is real? The answer is below, but place your vote to see how your guess everyone else (no cheating now!) |

A Cat’s Commentary


Trivia answer: A is AI movie (very funny video), and B is real movie.
![]() | That’s all for now.
|




