Happy Tuesday ⚡️
The Pentagon drama sent Claude to #1 on the US App Store last week—51% more users in seven days, and Anthropic's servers felt every one of them. Fourteen incidents in eight days. Opus noticeably nerfed—responses shorter, reasoning shallower, the thing that made it worth paying for quietly dialed back. Usage limits hitting way faster than usual.
The why: Cursor's internal analysis found Anthropic's $200/month Claude Code plan now consumes up to $5,000 in compute—2.5x what it was last year. When you're eating 25x on subscriber costs and a surge hits, you start throttling.
Meanwhile OpenAI dropped GPT-5.4 in Codex—1M context window, native computer use, unified model—and it's been ripping. (brb writing a Codex guide.)
Today, we're talking about:
The most underrated hire in the AI era
How to turn Claude Cowork into an actual executive assistant
Meta just bought the social network for AI agents
Anthropic's job map and who's on it
The prompt that audits your entire AI brain
The Most Underrated Hire Right Now Isn't an Engineer
When building was hard, engineering was the bottleneck. The status hierarchy reflected that, and it made sense. Building is no longer hard. Which means the variance in outcomes has shifted almost entirely to judgment: what to build, how to sequence it, and how to talk about it.
The hire that matters most right now isn't a product manager. It's something closer to a product thinker—someone with an intuitive grasp of the product as it exists, where it's soft, where it sings, and how to iterate it toward something sharper. This person has to hold in their head where the product should be two years from now and work backwards from that.
And the story matters as much as the thing. Internally, it organizes the team around a shared model of why. Externally, it shapes the interpretive frame users bring to their first experience. You can't retrofit narrative onto a product and expect it to land—it has to be load-bearing from the start.
The rarest version of this person sits at the intersection of culture and deep technology. Genuinely bilingual. They know what's technically possible and they know which cultural currents are real vs. ephemeral. That combo is what separates products that feel inevitable from products that feel assembled.
Before the "this person has always been valuable" clap-back: correct. The point is they might now be the most important person in the room, and almost no org is hiring like that's true. When engineering was the constraint, a great product thinker got overruled by shipping timelines. That constraint is gone. Give that person a team of AI agents and a clear north star and they out-build any eng-heavy org from three years ago.
The Anthropic labor report from last week is a useful parallel. The gap between what AI can do and what most orgs are deploying it on is enormous—because most orgs don't have anyone whose full-time job is holding the vision and sequencing toward it. They have managers and executors, not thinkers. That's the real bottleneck now.
Need help building AI into your engineering and growth workflows?
Tenex is the team behind this awesome newsletter. We embed with your team to design, build, and ship AI systems that actually work—from agentic engineering pipelines to AI-powered growth engines.
Cowork Is Generic Until You Build It a Brain
Most people open Claude Cowork, start from scratch every time, and wonder why the output feels a little off. That's not a Claude problem. That's a setup problem.
The insight from JJ Englert's guide—which blew up on X this week and turned into a full video walkthrough—is that Cowork out of the box is a good assistant. With the right context structure, it becomes an executive-level partner. The difference is what Claude knows before you say a word.
Here's what the setup looks like:
1. build the folder structure. Create a workspace folder that acts as Claude's long-term memory. Inside it: about-me.md, brand-voice.md, working-preferences.md, a Current Projects/ folder with individual project files, and a memory/ directory with subfolders for people, projects, and context. This is the brain. Claude reads it at the start of every session.
2. install the right plugins. The Memory plugin gives Claude a two-tier context system that persists across sessions—and it comes bundled with the Productivity plugin, which handles task tracking and daily updates. Install Memory and you get both. Worth doing before anything else.
3. connect your tools. Slack, Gmail, Google Calendar, Notion. The more tools connected, the less Claude has to guess. Meeting prep, email replies, and task management all improve because Claude is pulling real data instead of making things up.
The setup prompt below does the heavy lifting. Paste it into Cowork and Claude will interview you phase by phase—building each file, showing it to you before saving, and not moving on until you approve.
JJ also dropped a video walkthrough and a full copy-paste guide on GitHub. that you can grab here.
Human-verified prompt
You are going to help me set up my Claude Cowork workspace so that every
future session starts with full context about who I am, what I do, and
how I work. We're building a "brain" that makes you useful from the
first message.
You're going to interview me in phases. Ask questions, then build the
files based on my answers. Don't rush. Don't assume. Ask before you build.
Phase 0: Recommend the Productivity and Memory plugins, then help me
connect my daily tools: Slack, Gmail, Google Calendar, Notion.
Phase 1: Interview me to create about-me.md — work, background,
content channels, professional values, positioning.
Phase 2: Analyze existing content (or interview me) to create
brand-voice.md — voice rules, tone by context, dos and don'ts.
Phase 3: Create working-preferences.md — what I want help with daily,
communication style, format preferences, safety rules.
Phase 4–8: Content strategy, team members, active projects, memory
system (CLAUDE.md hot cache), and skill files for recurring output.
Rules: One phase at a time. Show each file before saving. Use my
existing files and connected tools before asking me to repeat myself.
Lowercase, hyphens, .md format. Save everything to my workspace folder.
Start with Phase 0.
Meta just bought the social network where only AI agents can post — Moltbook launched in January as an experimental "third space" for AI agents. Humans can read. Only agents can write. Meta acquired it this morning, with founders Matt Schlicht and Ben Parr joining Meta Superintelligence Labs on March 16. No price disclosed. The infrastructure for agent-to-agent communication just got a $1.4T company behind it. Axios
Anthropic mapped which jobs are next — 17-page labor market report from two economists using actual Claude usage data. AI can cover 75% of a programmer's tasks now. Hiring for workers aged 22–25 in exposed fields is quietly slowing. The mass layoffs haven't started—but the replenishment is stopping. Fortune
Audit your own Claude sessions before you automate anything — Chintan Turakhia's prompt: "scrape all of my Claude sessions on this computer. give me a breakdown of all the things I do, things worth making into skills vs plugins vs agents vs claude.md." Run it. You'll find stuff you forgot you were doing. Tweet
Intercom's Fin crossed $100M ARR. On its own. — Eoghan McCabe just wrote one of the more honest essays from a SaaS CEO in a while. Intercom is at $400M ARR. Fin does a million resolutions a week—the equivalent of 6,533 humans, by his math. Growing 3.5x. Read it
"AI spikes on intelligence. Humans spike on judgment." — Julien Bek's frame on the Anthropic labor report. The gap between what AI is capable of and how organizations deploy it is the whole game right now. Tweet
OpenAI's hardware lead quit over the Pentagon deal — Caitlin Kalinowski, who ran OpenAI's robotics team, walked out last week. The deal Sam Altman later called "opportunistic and sloppy" has now cost the company one of its more respected execs. TechCrunch
Open roles:
Newsletter Writer (yup, you’ll write this thing)
AI Strategist
Talent Acquisition Lead
Technical Recruiter
Forward Deployed Engineer
Applied AI Engineer
Engagement Manager
Salary ranges vary by role and experience. Additional comp based on output. Must be NY-based.