MakerSquare Issue 007
AI for Operators & Founders · July 2026

AI Builder
Brief.

What happened in AI this week, and what it means for your work.
From the Desk

Last week the frontier closed. This week it opened.



The export controls that were holding back the newest AI models got lifted. The latest Claude went worldwide overnight. California signed the largest AI rollout in the history of U.S. state government. And Meta laid off 8,000 people while quietly moving 7,000 others into brand-new AI teams.

The direction is the same one we keep tracking. The tools are arriving faster than the rules, faster than the org charts, and faster than most people's habits. The one thing that stays in your control is whether you're ready to use them when they land on your desk.
Marliis Schneider
Founder & CEO, MakerSquare · Austin, TX
01 · The 3 Things

This week's AI stories, and why they matter
Policy · AI Access

The export controls came off. The newest Claude went global the same week.

On June 30, the Trump administration lifted the export controls that had frozen Anthropic's most advanced models, Claude Fable 5 and Mythos 5. Within a day, Fable 5 was available worldwide again, across the web app, Claude Code, and the API. At the same time, reports surfaced that the White House is in advanced talks with OpenAI, Anthropic, and Google on a voluntary framework for how the next tier of frontier models gets released, and who can access them.
Why it matters
Two weeks ago the story was Washington deciding who gets access first. Now the same models are open to everyone at once. The takeaway isn't that the gate is gone for good. It's that access keeps changing, and the people who built real skills while it was closed are the ones ready the moment it opens. Availability changed overnight. The ability to actually use it doesn't.
Read the story →
 
Government · Deployment

California just made Claude available to every state agency at half price.

On June 29, Governor Gavin Newsom announced a first-of-its-kind partnership putting Anthropic's Claude in front of every California state agency, plus cities and counties, at a 50% discount with free workforce training included. The DMV is using it to cut wait times. The Department of Technology and the state's emergency office are using it for cyber defense. The Department of Health Care Services is using it to better serve Medicaid recipients. It is the largest AI deployment in the history of U.S. state government.
Why it matters for you
When an organization the size of California decides AI is core infrastructure and not a pilot project, the "wait and see" phase is over for a lot of people. Government work is not known for moving early. If the DMV is retraining its staff around Claude this summer, the question for everyone else stops being whether to learn these tools and becomes how fast.
Read the story →
 
Workforce · Layoffs

Meta cut 8,000 jobs and moved 7,000 people into new AI teams in the same restructuring.

Meta began notifying roughly 8,000 employees, about 10% of its workforce, that they were being laid off, framing the cuts as a way to fund its AI push. In the same restructuring, upward of 7,000 workers were redirected into newly created AI-focused teams like Applied AI Engineering. Mark Zuckerberg told staff no further company-wide cuts are expected this year. The bigger picture: AI is now the single most-cited reason in 2026 layoff announcements, named in roughly 22% of all cuts this year.
Why it matters for you
The layoff and the AI hire happened inside one memo. Companies aren't shrinking so much as reshaping, and the people who move into the new AI roles are the ones who already showed they could work with the tools. The seat doesn't disappear. It gets relabeled, and it goes to whoever is ready to fill the new version of it.
Read the story →
02 · Try This Today

One prompt you can use before Friday.

Which parts of my week could an AI tool already handle, and which parts are the judgment that only gets more valuable when the routine work speeds up? The Meta memo is the assignment this week. In one document, the cut and the AI hire happened together. So run the same exercise on yourself before someone runs it for you. Open Claude, Copilot, or whatever AI tool you already have access to, and paste this in. No technical setup required. Takes about 20 minutes.
Copy this prompt
Here are the 10 tasks I spend the most time on each week:
[list them in plain language]

For each one, tell me:
1. Which parts an AI tool could realistically handle today.
2. Which parts still need a human judgment call.
3. What a person who hands the routine parts to AI would spend that freed-up time on instead.
What you get: A clear split between the routine work a tool absorbs and the judgment that becomes more valuable as everything else speeds up. That second list is your anchor. Build there.
03 · Deep Dive

The cut and the hire are the same decision.

Last issue we made the case that companies are cutting jobs based on what they believe AI will do, not what it does today. This week Meta showed the other half of that same coin.

The 8,000 layoffs and the 7,000 people moved into new AI teams weren't two separate events. They were one restructuring. Meta didn't decide it needed fewer people. It decided it needed different people, doing different work, organized differently. The headcount barely moved. What moved was the definition of the work.

That's the pattern underneath the whole year. AI is cited in about 22% of 2026's layoffs, up from a small fraction of that in January. But over the same stretch, the roles that require AI skills are the ones companies are competing hardest to fill. The layoff tracker and the hiring surge are describing the same reorganization from opposite ends.
Your role isn't disappearing. It's being redefined around who can direct these tools and who can't, in memos being written this quarter.
The uncomfortable part is that the redefinition doesn't wait for you to feel ready. California's DMV is retraining its workforce this summer. Meta reshuffled 15,000 people in a single announcement. The gap between "I should learn this" and "the decision already got made" is closing fast.
Myth of the Week
"If AI takes my job, I'll see it coming." Usually you won't, because it doesn't arrive as a robot doing your work. It arrives as a reorg where the new roles are written for people who already work with AI. The signal isn't a machine at your desk. It's a job posting that lists skills you haven't built yet.
04 · Tool Spotlight

NotebookLM.

An AI research assistant that only knows what you give it.
notebooklm.google.com →
Most AI tools answer from the whole internet, which means they can drift, generalize, or make things up. NotebookLM flips that. You upload your own material, reports, meeting notes, a contract, a stack of PDFs, and it answers only from those documents, with citations pointing back to the exact passage.

There's nothing to install and no technical setup. You drag in files, then ask questions in plain language, the same way you'd ask a colleague who had read everything.

One use that pays off immediately: drop in every document tied to a decision you're about to make, a proposal, the last three status updates, the relevant policy, and ask "what's inconsistent across these?" It reads the whole pile in seconds and surfaces the contradictions a person skimming would miss. Because it's grounded in your own material, you can trust the answer enough to act on it.

It also generates a short spoken summary of whatever you upload, which is genuinely useful for reviewing a dense report on a commute.
Best for
Making sense of your own documents
Price
Free
Plus via Google One
Setup time
~3 minutes
05 · What We're Building

Cohort 1 is live.

As this issue lands, Cohort 1 is in the room, building. Eight sessions. Three deployed AI products each, built around their actual businesses, not classroom exercises.

Everything in this week's news points the same direction. Access opened up. Government went all-in. The layoff and the AI hire showed up in the same memo. Cohort 1 is for the people who decided they'd rather be the ones writing the new role than the ones reading about it.

No coding background required. Cohort 2 seats are filling now.
Cohort 2
Aug 3-14
Small cohort
Cohort 3
Oct 12-23
Small cohort
Cohort 4
Nov 9-20
Small cohort
Download the curriculum →

Know someone who should be reading this?

Forward it to a founder or operator who needs the signal without the noise.
Forward this issue →
Free · Every Tuesday · makersquare.ai
MakerSquare
AI Builder Program · Austin, Texas

Recommended for you

© 2026 MakerSquare LLC.
beehiivPowered by beehiiv