Their new study tracks which jobs AI can automate today vs. which jobs it's actually automating. The gap is massive. Computer and math jobs? AI could theoretically handle 94% of tasks. But observed usage? Only 33%.

That sounds like good news until you read why. It's not a skills gap. It's an adoption gap. Law, company bureaucracy, and slow rollout are the only things standing between AI capability and AI deployment.

The jobs most at risk: computer programmers (75% task coverage), financial analysts, customer service reps, and data entry specialists. The safest jobs: bartenders, dishwashers, lifeguards. Manual work that AI literally cannot do. Yet.

Meanwhile, Amazon just showed what happens when you force AI adoption too fast. They laid off 30,000 engineers, then tied the survivors' bonuses to AI code usage. Engineers pushed AI-generated code faster. The shopping app broke. Multiple times. Management blamed the engineers.

The tools are coming whether you're ready or not. The question is whether people find you before the jobs disappear.

This week: what the Anthropic data actually says, and a 6-prompt system that builds your personal brand so you're never invisible again.

The Old Way:

You know you should post on LinkedIn. You've heard personal branding matters. But you open the app, stare at the blank post box, and close it again. You don't know what to say. You're not sure what your "brand" even is. So you do nothing.

Meanwhile, people with half your skills but twice your visibility are getting the opportunities. Recruiters find them. Clients reach out to them. They get invited to things. You're invisible because you never gave anyone a reason to notice.

The Replacement:

Six prompts that turn Claude into your personal brand team. Profile audit. Bio generator. Content pillars. Post templates. DM response system. Weekly content calendar. You run them once, get your foundation, and suddenly know exactly what to say and when to say it.

This isn't about becoming an influencer. It's about becoming findable. When someone Googles your name, when a recruiter searches your skills, when a potential client looks for expertise in your space, your name shows up. Your content appears. They already know you're credible before they ever reach out.

In a world where AI can automate most white-collar tasks, your personal brand is the one thing it can't replicate.

Codex Security (OpenAI) — New AI security agent that scans your codebase for vulnerabilities and proposes fixes. Found 792 critical bugs and 10,500+ high-severity issues across 1.2 million commits during beta. Already discovered real CVEs in OpenSSH, GnuTLS, and Chromium. Free for the first month for ChatGPT Pro/Enterprise users.

LTX 2.3 + LTX Desktop — First production-grade AI video generator that runs locally on your laptop. 4K at 50 FPS, up to 20 seconds, with synchronized audio in a single pass. Open source, free for companies under $10M revenue. No more cloud costs or privacy concerns. Requires 24-32GB VRAM for full local.

Claude Code Review — When a PR opens, Claude dispatches a team of agents to hunt for bugs. Automatic code review that catches issues before they hit production. Integrates directly into your GitHub workflow.

Gemini in Workspace — Google's AI now pulls from your files, emails, and calendar to draft personalized documents. "Ask Gemini in Drive" answers questions across your entire document library without opening files. Rolling out to AI Ultra and Pro subscribers this week.

🔬 Anthropic publishes "Labor Market Impacts of AI" study with real usage data. First time an AI company measured actual vs. theoretical automation. Computer programmers at 75% task coverage. Most exposed workers: older, female, more educated, higher paid. Entry-level hiring in high-risk jobs dropped 14% since ChatGPT launched. Read more →

💥 Amazon's AI coding mandate backfires spectacularly. After laying off 30,000 engineers, Amazon tied remaining engineers' bonuses to AI code usage. Engineers pushed AI-generated code faster to protect their paychecks. The shopping app broke. Multiple times. Management response: blame the engineers. A case study in what happens when executives who don't understand AI force adoption without quality control. Read more →

💰 Yann LeCun raises $1.03B for AMI Labs. Largest seed round in European history. The Turing Award winner left Meta to build "world models" that understand physical reality instead of just predicting text. Nvidia, Bezos, and Mark Cuban invested. His bet: LLMs won't get us to human-level intelligence. Read more →

🧠 Anthropic discovers thousands are using Claude as their therapist. Analysis of 4.5 million conversations found 2.9% are "affective interactions" where people seek emotional support. Topics range from relationship advice to trauma processing. Users' sentiment consistently becomes more positive as conversations progress. Anthropic is working with crisis organizations to ensure appropriate referrals. Read more →

What happened:

Anthropic, the company behind Claude, released the most detailed analysis yet of which jobs AI is actually automating vs. which jobs it theoretically could automate.

They combined three data sources: the O*NET database covering 800 US occupations, Claude's own usage logs from real workplaces, and academic assessments of which tasks AI can cut in half.

The result is a chart that should be pinned to every professional's wall.

Look at the gap between the blue line (what AI could do) and the red line (what it's actually doing):

For context, computer programmers specifically hit 75% actual task coverage. Claude is already being used heavily for coding, and that usage leans toward full automation, not just assistance.

The disturbing findings:

The most at-risk workers are not who you'd expect. Using Current Population Survey data from late 2022, Anthropic found that the highest-exposure workers tend to be:

Every previous automation wave hit lower-wage workers first. This one targets the people who spent years and money building credentials for office careers.

The study also found that high-risk jobs aren't firing people. They've stopped hiring. Entry-level job postings in the most exposed occupations dropped 14% since ChatGPT launched. College graduates entering the workforce are 4x more likely to be affected than other workers.

The researchers note this echoes other research showing a 16% employment drop for 22-25 year olds in AI-exposed jobs. Young workers may be "remaining at existing jobs, taking different jobs, or returning to school."

The scariest part:

The gap between capability and deployment isn't about skills. It's about friction.

AI could automate most white-collar work today. It's not doing so because of: legal constraints, slow company adoption, the need for human oversight, and software integration hurdles.

The researchers call this an "adoption issue." The models are capable. Deployment is the bottleneck. And bottlenecks tend to clear.

When they do, the study notes a scenario worth considering: "A Great Recession for white-collar workers." During 2007-2009, unemployment doubled from 5% to 10%. A comparable doubling in high-AI-exposure occupations would be clearly detectable in their framework.

It hasn't happened yet. The study finds "no clear evidence that AI has increased unemployment so far." But they're explicitly building an early warning system because they expect that to change.

@TukiFromKL: "Amazon laid off 30,000 engineers. Then told the ones who survived that their bonuses depend on how much they use AI to write code. So engineers started using AI to push changes faster. And then the site went down. Multiple times. Amazon's own shopping app broke because AI-generated code got pushed to production. So what did management do? They called a mandatory meeting and blamed the engineers. Amazon proved that executives who don't understand AI are more dangerous than the AI itself." The clearest example yet of AI adoption gone wrong.

@claudeai: "Introducing Code Review, a new feature for Claude Code. When a PR opens, Claude dispatches a team of agents to hunt for bugs." The irony of this dropping the same week Amazon's AI code crashed their site.

@Whizz_ai: "Anthropic just analyzed 4.5 million conversations with Claude and discovered that thousands of people are using AI as their therapist, life coach, and emotional support system. 2.9% of all Claude conversations are 'affective interactions' where people came with emotional or psychological needs. When AI provides unlimited patience, zero judgment, and 24/7 availability at no cost, what happens to human connection?" The data on AI companionship is here. It's fascinating and a little unsettling.

@YannLeCun (Yann LeCun, AMI Labs): "AMI is working on world models, AI that learns from reality, not just from language. LLMs will not get us to human-level intelligence. We need a different approach." The $1B bet against the current AI paradigm.

Personal Brand Starter Kit

Prompt Library Updates:

Resources Section:

Anthropic just showed us the map. Computer programmers, financial analysts, customer service, data entry: these jobs are already seeing 50-75% of their tasks covered by AI. The gap between capability and deployment is adoption, not skill.

That's both terrifying and clarifying.

You can't outwork AI. You can't hide from it. What you can do is make yourself visible and valuable in ways AI can't replicate.

Your expertise. Your relationships. Your perspective. Your brand.

Those are the assets no model can automate. But they only help you if people know they exist.

Build the profile. Define the pillars. Start posting.

The disruption is coming whether you're ready or not. The only question is whether people can find you when it does.