Claude Can Now Control Your Computer

Welcome back apprentices! 👋

Hey {{first_name|friend}},

This week in AI feels a bit like walking into your office… and finding out half your coworkers got replaced overnight — and the other half got promoted to “manager of robots.”

Nothing looks dramatically different at first glance, but something big is clearly shifting behind the scenes.

The tools are getting faster, quieter, and a little more… independent.

And whether you like it or not, they’re starting to take initiative.

_{Read Time: 4 minutes}

🧠 Hope Up, Trust Wobbly. Anthropic just turned its own AI, Claude, into a global interviewer — chatting with 81,000 people across 159 countries in 70 languages—and found we’re equal parts excited and uneasy. People love the idea of AI boosting productivity and freeing up time, but fear it messing things up (a lot), taking jobs, and making us overly dependent. Sentiment isn’t universal either: optimism runs higher in places like India and South America, while the U.S., Europe, and East Asia are more skeptical.

🐮 Silicon Valley Meets the Barnyard. New Zealand startup Halter is reportedly heading toward a $2B valuation, using AI-powered “Cowgorithm” collars to herd cattle with virtual fences and real-time data from 6,000+ signals per minute. Ranchers can now manage livestock from an app, replacing physical fences and manual checks with solar-powered wearables and behavioral insights. Backed by major investor interest (including Peter Thiel’s Founder’s Fund), the company is scaling fast from its farm roots into the U.S. market.

🤖 Your Next Device Might Think Like You. Figure AI founder Brett Adcock is back with Hark, a stealthy new startup building personalized AI devices designed to act as a “new interface to AGI”—and maybe even think ahead of you. Backed by $100M of his own money and a team from Apple, Google, and Tesla, Hark is pairing custom hardware with powerful models (fueled by thousands of NVIDIA GPUs) set to launch this summer. It’s a bold bet in a space where others have struggled—but with serious talent and ambition behind it

Unlock a focused set of AI strategies built to streamline your work and maximize impact. This guide delivers the practical tactics and tools marketers need to start seeing results right away:

Stay ahead of the curve with these top strategies AI helped develop for marketers, built for real-world results.

Download the Free Report

This week in AI feels like the industry collectively decided to stop experimenting and start reorganizing for dominance. 

OpenAI is scrapping Sora to double down on its next big model (“Spud”), Anthropic is turning Claude into a hands-off digital operator, and Meta is building internal AI agents to flatten its org chart. 

Meanwhile, Musk is dreaming (big) with a terawatt-scale chip factory, Microsoft is inching toward independence with stronger models, and ARC-AGI-3 reminds everyone that “real reasoning” is still unsolved. 

Oh — and Reddit is preparing for a future where bots might outnumber humans. 

OpenAI dropping Sora isn’t failure — it’s prioritization. 

Video generation, while flashy, is compute-heavy and commercially unclear compared to foundational models. “Spud” signals a return to core intelligence scaling, and the internal reshuffling (safety + “AGI Deployment”) suggests the company is aligning everything toward faster, broader rollout of powerful systems. 

Translation: fewer side quests, more main storyline.

Claude controlling your desktop is a major shift — from “assistant” to “operator.” 

By combining app integrations, browser use, and fallback screen control, Anthropic is quietly building a general-purpose agent that works across environments. 

The addition of mobile-triggered tasks (“Dispatch”) removes the need to even sit at your computer — this is early infrastructure for ambient computing.

Zuckerberg’s “CEO agent” isn’t just a cool tool — it’s a statement about how companies might run. 

By replacing layers of middle management with AI systems that fetch, reason, and negotiate, Meta is testing a leaner, faster org structure. 

Internal tools like “Second Brain” and agent-to-agent negotiation hint at a future where employees manage AIs as much as they manage people.

Terafab is classic Musk — ambitious to the point of disbelief, but grounded in a real bottleneck: compute scarcity. 

Vertical integration (design → fabrication → deployment) plus space-based AI infrastructure could drastically change cost curves if it works. It’s less about today’s models and more about enabling orders-of-magnitude bigger ones tomorrow.

MAI-Image-2 shows Microsoft is no longer just distributing AI — it’s building competitive models. 

Strong gains in text rendering and visual quality point to real technical progress, while leadership shifts toward frontier research suggest long-term independence from OpenAI is the goal.

Every new ARC benchmark humbles AI — and V3 is no different. 

Sub-1% scores highlight that today’s models still struggle with abstract reasoning and novel problem-solving. 

The real question: will progress come from smarter architectures — or just more compute and training tricks?

Bot labeling is a small fix for a massive trend. 

As AI-generated content scales, platforms face a core challenge: preserving human authenticity without killing anonymity or usability. 

Reddit’s approach — light-touch verification + community control—may be a preview of how the broader web adapts.

AI is becoming infrastructure. Your future tools won’t just help you write or search — they’ll act on your behalf, manage workflows, and even replace layers of decision-making. 

Whatever you do, the shift to agents + personalized AI + compute wars means one thing: knowing how to use AI won’t be enough — you’ll need to know how to delegate to it.

If your to-do list still requires you to open your laptop… you’re already behind the next version of the workflow.

Hey, you didn’t get all this info for nothing — share it! If you know someone who’s diving into AI, help them stay in the loop with this week’s updates.

Sharing is a win-win! Send this to a friend who’s all about tech, and you’ll win a little surprise 👀

Those numbers aren't a CX issue — they're a design issue. Gladly's 2026 Customer Expectations Report breaks down exactly where AI-powered service loses customers, and what the architecture of loyalty-driven CX actually looks like.

Get the report

A playground for your imagination (and low-key prompt skills).

Each send, we give you a customizable DALL·E prompt inspired by a real-world use case — something that could help you in your business or job if you wanted to use it that way. But it’s also just a fun creative experiment.

You tweak it, run it, and send us your favorite. We pick one winner to feature in the next issue.

Bonus: you’re secretly getting better at prompt design. 🤫

Last week, we challenged you to test GPT-4o’s visual generation skills with this prompt.

Here’s the WINNER:

Congrats to Jonathan for his creation!🥳

Want to be featured next? Keep those generations coming!

Inside a sleek, mirrored elevator with soft ambient lighting, a group of people stands quietly as the doors begin to close. Just before they shut, the back wall display lights up with a bold, ultra-clean visual of [your concept] — rendered as a striking, photorealistic graphic or object that feels out of place in such a confined space. The reflections multiply the image infinitely across the mirrored panels, creating depth and visual tension. Subtle colored lighting (deep blues, warm ambers, soft magentas) enhances the scene. Shot from a cinematic eye-level angle with sharp focus, crisp reflections, and high-end realism — capturing a fleeting, powerful moment of attention in an otherwise ordinary setting.

If AI outputs feel inconsistent, it’s usually not the model, it’s missing structure.
We documented the exact 6- Part System we use to get reliable results across ChatGPT, Claude, and Gemini.

It’s a short guide you can finish in under an hour, with plug-and-play prompts + exercises so you actually build the skill and fix the frustrating AI inconsistencies.

Subscriber Price: $10 (normally $19).

_FEEDBACK