In partnership with
Welcome Automaters, 👋
On June 24, (yesterday) OpenAI finally pulled back the curtain on its very first custom-built AI chip named Jalapeño. Hot take? We think the name is absolutely immaculate.
OpenAI built this new piece of tech in a tight partnership with semiconductor giant Broadcom. But here’s the ultimate, mind-bending part: OpenAI’s own AI models actually helped design it.
So What Does Jalapeño Actually Do?
One specific job: inference. If you are wondering what that means in plain English, inference is the exact, high-stakes moment when an AI model actually processes your prompt and spits a response back to you. Every single time you type a question into ChatGPT, inference is happening under the hood.
Right now, that entire global process runs on insanely expensive general-purpose graphics chips made by Nvidia. OpenAI has been burning through those Nvidia chips like popcorn at a summer blockbuster, which is a massive financial headache.
Here’s what the new chip brings:
Power Efficiency: OpenAI's early tests are already showing significantly better performance per unit of power compared to current market options.
The Sweet Spot: The chip is heavily optimized for real-time coding models, which require lightning-fast, continuous responses.
Why does this matter?
Because up until yesterday, OpenAI was completely dependent on Nvidia’s hardware. Now those GPUs are not only eye-wateringly expensive, they’re also in ridiculously high demand with massive waitlists.
Every other massive tech titan, from Google to Amazon, built their own internal custom chips years ago to escape this exact bottleneck. OpenAI is just now joining the custom silicon club, fashionably late, but clearly dressed to impress.
OpenAI President Greg Brockman hopped on the company’s internal podcast to spill the strategic tea behind the project, according to him: The goal was to find specific workloads that existing chips on the market were not serving well, and then build something purpose-made to absolutely crush them.
While the chip is still technically in its internal testing phase, the ultimate corporate direction here is crystal clear. OpenAI doesn't just want to build cool software anymore. They want to completely own the entire tech stack from top to bottom: the AI model, the massive data centers, and the raw silicon underneath it all.
However Nvidia isn't out just yet, in fact according to Techcrunch, the heavy lifting of pre-training will likely stay on Nvidia's turf for now, but inference is where the bills pile up daily, and cutting those costs, even slightly, is like finding a leak in a very expensive pipe.
Real quick: We go much deeper on YouTube
So go over there, hit Subscribe, tap that notification bell and come hang with us where the real conversations happen. You see, subscribing keeps you plugged in so you get the right links, the right videos, right when they drop.
Here's what we have for you today
😱 Anthropic Accuses Alibaba of Stealing Claude AI Capabilities in Record-Breaking Attack
Anthropic has officially gone nuclear on Chinese tech giant Alibaba.
According to a bombshell report from Reuters, the U.S. AI powerhouse sent a formal letter straight to the White House. The accusation? They’re claiming Alibaba orchestrated the largest, most coordinated AI model capability theft in Anthropic's entire history.
The sneaky method used is a technical process called "distillation." If you want the gossip in plain English, distillation is essentially a high-tech shortcut:
A company takes a cheaper, weaker model and trains it directly on the premium outputs of a competitor's frontier model.
It effectively copies the algorithmic "brains" of an advanced AI without having to spend hundreds of millions of dollars on research and development.
It’s pure, industrial-scale intellectual property cloning.
Now, Anthropic didn't just guess here. They brought the actual receipts. They claim the data-scraping campaign ran from April 22 to June 5, 2026, and racked up some absolutely jaw-dropping metrics:
The Scale: The operators used a massive network of nearly 25,000 fraudulent accounts.
The Volume: They successfully forced Claude into more than 28.8 million rapid-fire exchanges.
The Target: The entire operation was reportedly carried out by units affiliated with Alibaba and its internal AI laboratory, Alibaba Qwen. The ultimate goal was to aggressively accelerate China's domestic capabilities to match Anthropic's unreleased, top-tier model, Mythos Preview.
And let’s be real: Alibaba is definitely not the first lab caught with its hand in the cookie jar. Back in February, Anthropic revealed that Chinese AI startup DeepSeek, Moonshot AI (clocking 3.4 million queries), and MiniMax (topping 13 million) had all run similar extraction campaigns. But this new Alibaba heist completely blows those previous records out of the water!
The timing of this White House briefing is politically spicy too. Just two days after Anthropic sounded the alarm, the U.S. Commerce Department clamped down hard, imposing massive export restrictions on Anthropic's advanced Mythos and Fable architectures over fears of Chinese military exploitation. This forced Anthropic to pull the plug and disable global access to those models entirely.
So that export control restriction might have a reasonable explanation after all.
Meanwhile, in other hilariously hypocritical tech news: companies are already rationing the exact same AI tools they once literally begged their employees to use.
According to a fantastic new exposé by TechCrunch, the corporate trend known as "tokenmaxxing" is officially dead. Consulting giant Accenture is frantically trying to stop its massive workforce from burning through corporate AI token reserves on basic, lazy tasks like converting PDFs into PowerPoint slides.
This is especially hilarious because just a few months ago, Accenture leadership was practically threatening staff, telling employees they risked missing out on major promotions if they didn't aggressively integrate AI into their daily workloads.
Leaked audio from an internal meeting reveals Accenture's agentic AI strategy lead admitting that their corporate AI spend has become "very unpredictable." Senior leadership is now actively panicking and questioning whether they’re getting a single dollar of actual business value from their massive computational investments.
So first, the rule was: use AI or get left behind. Now, the rule is: not like that, it's costing us a fortune! The corporate whiplash is truly spectacular.
But what's your take? Is AI actually pulling its weight where you work, or does it feel like an expensive magic trick so far? Hit reply and let us know. We read every single one of your replies.
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.
🧱 Around The AI Block
🦾 The 12 rules of agentic AI for successful enterprise transformation.
😬 Claude Mythos broke into classified US systems in hours.
🤖 US government reportedly urging Meta to share its AI models.
🚶♀️ AI researchers continue to leave Google for its rivals.
👩💻 Facebook rolls out an AI companion app for creators.
🕵️♂️ Google quietly starts mining your search images for AI training.
👌 Figma adds code layers, support for animations, and more AI features in new update.
🛠️ Trending Tools
For the Language Learners: CoffeeCat AI uses "Nano Banana" (a high-speed visual/audio model) to act as a live conversation partner. It watches your pronunciation and provides real-time feedback through a natural voice chat, making it feel like you're grabbing coffee with a local.
For the Digital Organizers: Mem0 gives your AI a "long-term memory." It’s a layer that helps your personal AI agents remember your preferences, past projects, and specific ways you like things done across different apps, so you never have to repeat yourself twice.
For the Power Developers: Windsurf is for the "Heavy Use" crowd. While other tools are just assistants, Windsurf is your "AI Squad." It features a flow-state engine called Cascade that understands your entire codebase, not just the file you’re in. It acts as a team of engineers that can strategically implement large-scale features while you act as the architect.
For the Automation Architects: Gumloop allows you to build complex, multi-step AI agents that can scrape websites, summarize data, and push it to your internal tools—all without writing a single line of code. It’s the ultimate "stealth speed" tool for operations.
🤖 AI Workout Of The Day: How to Create Stunning 3D Animations from Text and Videos
While the big studios were busy arguing over render farms, AI just handed the keys to the kingdom to... well, everyone.
We’re talking "Pixar-level" visuals from a text prompt. If you’ve got an imagination and ten fingers (okay, one finger for clicking), you’re officially an animator.
The MVP: Krikey AI
If you want to go from "zero to hero" without a manual the size of a dictionary, Krikey AI is the runaway winner. It’s specifically designed to be so simple a 10-year-old could run a mini-studio from their tablet.
How to Build Your Masterpiece:
Hit Krikey.ai and choose a 3D avatar from the library or build your own from scratch
Type a prompt like "doing a backflip" into Quick Animate or upload a video of yourself to have the AI copy your moves.
Type your script into the Magic Studio to generate a voiceover that’s perfectly lip-synced to your character’s mouth.
Personalize the Scene: Pick a 3D background, tweak the lighting, adjust cinematic camera angles, and add text effects.
Hit export. Download it as an MP4 for the 'Gram or a GIF for the group chat or socials and you’re done.
Pro Tip:
Krikey AI adds AI Lip-Sync dialogue in 20+ languages.
For more cinematic, "moody" clips, creators are also flocking to Luma Dream Machine.
The Cost: Most tools offer free credits daily, so you can play before you pay!
💡 Prompts to try:
Prompt: “A sleek superhero in a metallic suit flying down from the sky, landing on one knee in a powerful 'hero pose,' then slowly standing up and dusting off their shoulder with a smirk.”
Prompt: “A small, fluffy monster doing a high-speed breakdance routine, ending with a clumsy trip and a goofy wave to the camera while sitting on the floor.”
Prompt: “A stylish 3D character performing a smooth 'moonwalk' across the stage, followed by a celebratory backflip and a peace sign pointed at the lens.”Is this your AI Workout of the Week (WoW)? Cast your vote!
That's all we've got for you today.
Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇
What'd you think of today's email?
Your feedback means a lot to us and helps improve the quality of our newsletter.
