Who Gets AI First?

Welcome back apprentices! 👋

Hey {{first_name|friend}},

Remember when the hardest part about trying new tech was remembering your password? 

Those days are fading fast. 😅 

Now it's about who gets the keys, who has to wait outside, and why that matters more than you might think. 

Let's dive in.

_{Read Time: 4 minutes}

🧪 The Gap Is Closing Fast. AI agents are getting noticeably better at real freelance jobs, with Anthropic’s Fable 5 becoming the first model to complete 16.1% of tested projects at professional quality — about 6× better than the top score from less than a year ago. The benchmark covered real-world work like design, animation, and floor plans, all judged by humans. That's impressive progress, but it also means AI still falls short on roughly 5 out of every 6 professional tasks.

🕵️ Who Owns Intelligence? Anthropic claims Alibaba carried out the largest known "distillation" attack yet, allegedly using 28.8 million Claude conversations to help improve its own models through thousands of fake accounts. Distillation itself is a standard AI technique, but Anthropic argues there's a big difference between training on your own work and harvesting a competitor's. As AI gets more valuable, expect the race to be fought in courtrooms as much as in labs.

⚡ The Fastest Boom in Tech History. Generative AI is now growing 3× faster than any previous technology, jumping from $110B in revenue last year toward a projected $175B this year. Every price cut fuels even more demand, companies keep talking about it, and data centers are set to drive over half of new U.S. electricity demand by 2030. The hype is loud — but the biggest economic impact may still be ahead.

Howard Dahl spent decades building the machines that feed America. His family invented the Bobcat skid steer. The air drills planting nearly every commodity crop globally? Those too. Now Dahl is manufacturing weed-cutting robots for Greenfield Robotics out of his Fargo factory, and he wrote his own check on top of it. 

Greenfield's current fleet is sold out, with over $1 million in total revenue and robots in the field since 2020. Chipotle’s venture arm and KingsCrowd Capital are also on board. The robots slice weeds with centimeter precision, replacing herbicides linked to environmental damage and rising health concerns among farmers. 

Greenfield is now in Test the Waters under Reg A+. Reserving shares today locks in a 5% bonus that can grow to 20% the week the round opens to the public.

Reserve Your Shares

_{Greenfield Robotics is Testing The Waters under tier 2 of Regulation A. No money or other consideration is being solicited, and if sent in response will not be accepted. No offer to buy the securities can be accepted and no part of the purchase price can be received until the offering statement filed by the company with the SEC has been qualified by the SEC. Any such offer may be withdrawn or revoked, without obligation or commitment of any kind, at any time before notice of acceptance given after the date of qualification. An indication of interest involves no obligation or commitment of any kind. “Reserving” shares is simply an indication of interest. There is no binding commitment for investors that reserve shares in this manner to ultimately invest and purchase the shares reserved of the company, or to purchase any shares of the company whatsoever.}

This week, the conversation was all about who gets to use AI first. 

The U.S. government is now playing an active role in how frontier AI reaches the public, OpenAI's newest flagship launched behind government-approved access, Anthropic reopened its most powerful model under tighter restrictions, and Meta quietly showed that reading words directly from brain signals without surgery is becoming surprisingly realistic. 

AI is still moving incredibly fast — but now it's being shaped as much by regulators and safety concerns as by engineering breakthroughs.

The biggest surprise wasn't GPT-5.6 itself — it was how it launched.

According to reporting, the Trump administration asked OpenAI to delay a public rollout and first make GPT-5.6 available only to a small group of vetted organizations while additional safety evaluations were completed. Sam Altman reportedly agreed, describing it internally as the quickest path toward a broader release rather than the company's preferred long-term strategy.

The model family itself looks impressive. Sol is OpenAI's new flagship with deeper reasoning and an "Ultra" mode that can coordinate multiple subagents simultaneously for difficult tasks. Terra aims to match GPT-5.5 while costing roughly half as much, and Luna focuses on speed and affordability.

The more interesting story, though, isn't another benchmark win.

It's the precedent.

For decades, software companies largely decided when to launch products. Frontier AI may be entering an era where governments increasingly influence when, how, and to whom the most capable systems become available.

That could become one of the defining trends of the next few years.

Anthropic had a busy week.

First came Sonnet 5, a stronger version of its popular mid-tier model with noticeably better reasoning, coding, and long-running agent capabilities. It brings many features that previously belonged only to Anthropic's premium models into a much cheaper package.

Then came Fable 5's return.

After spending 18 days offline following cybersecurity concerns, the model is back with stricter protections that Anthropic says now block the problematic behavior more than 99% of the time. Those safeguards may occasionally reject legitimate coding requests, but the company says most everyday programming remains unaffected.

Perhaps the most important development isn't technical at all.

As part of the reopening, future frontier Anthropic models will also provide pre-release access to the U.S. government before wider deployment.

Taken together with OpenAI's rollout, this no longer looks like an isolated decision.

It looks like an emerging pattern.

Meta's announcement sounds like science fiction, but it's grounded in careful research.

Its updated Brain2Qwerty system can now decode entire words and their meaning from non-invasive brain scans instead of only recognizing individual characters. That's a major leap from previous systems, reaching an average 61% word accuracy, with one participant hitting 78%.

The process still requires someone to spend hours inside a large brain scanner, so this isn't becoming tomorrow's smartphone feature.

What makes it important is that Meta also released the research openly.

That means universities and researchers worldwide can improve the technology instead of starting from scratch.

If future improvements continue without requiring brain surgery, this could eventually transform communication for people who have lost the ability to speak or type.

The latest news are all about who controls access to intelligence.

Governments are becoming active participants in frontier AI launches, companies are accepting new safety requirements before releasing their best systems, and entirely new interfaces — like brain-computer communication — are steadily moving from research labs toward practical use.

For most people, the immediate impact won't be that AI suddenly becomes dramatically more useful overnight.

Instead, you'll likely notice something more subtle: powerful capabilities arriving gradually, behind waiting lists, regional restrictions, enterprise partnerships, and increasing layers of oversight.

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 👀

Gmail tracks you. Proton doesn’t. Get private email that puts your data — and your privacy — first.

Ditch the Gmail data grab

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 Ryan for this creation🥳

Want to be featured next? Keep those generations coming!

Inside a bright, modern design lab, a large worktable displays [your object] recreated dozens of times — each version made from a completely different material. One is carved from marble, another woven from fabric, another made of translucent glass, polished chrome, recycled plastic, warm wood, brushed titanium, or glowing resin. Every version has the exact same shape but feels entirely different because of its material. Sunlight streams through tall windows, creating colorful reflections, shadows, and textures across the table. Ultra-detailed, photorealistic realism, rich colors, premium product photography, sharp focus, cinematic depth of field.

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