Agents are already doing real work, robots already have serious owners, capital has already chosen its industry, and companies that didn’t adapt are already laying people off.
I filtered out the noise for you
This week didn’t have a single dominant headline. Instead, there were five signals pointing in the same direction: AI has stopped being just a productivity tool and has become a structural decision. Who gets hired, who gets laid off, where capital flows, and who builds robots—it all follows the same logic. If you don’t see it this week, you’ll see it in your financial results next year.
1: Microsoft launched Copilot Cowork with Anthropic
Microsoft launched Copilot Cowork, an agent that executes complex, multi-step tasks across the Microsoft 365 ecosystem, built on Anthropic’s technology—not OpenAI’s. This is despite the $13 billion Microsoft has invested in OpenAI. Link to the article.
Anthropic’s original Claude Cowork had already triggered a $285 billion selloff in enterprise software stocks, as investors realized an AI agent could replace entire SaaS categories. Microsoft responded by adopting that exact technology to defend its productivity business.
AI that acts has already beaten AI that chats—even inside OpenAI’s own house.
2: Mass layoffs are only increasing
Oracle is considering cutting between 20,000 and 30,000 employees to generate $8–10 billion in cash flow for AI infrastructure. Meta plans to lay off 20% of its workforce (around 16,000 people) to redirect capital toward its $600 billion AI roadmap. Atlassian has already acted: 1,600 employees, 10% of its workforce, were cut to “self-fund AI investment.” (The HR Digest)
This isn’t a post-pandemic layoff cycle. Oracle isn’t cutting due to poor performance—its remaining revenue obligations grew 433% year over year. They’re swapping people for infrastructure. This is a financial decision, not an operational one. Full article here.
Work didn’t disappear—it concentrated into fewer hands that know how to use AI.
3: Uber’s founder launched an AI startup
Travis Kalanick launched Atoms, an industrial robotics startup for kitchens, mining, and transportation. It’s absorbing CloudKitchens into Atoms and acquiring Pronto, the autonomous vehicle startup for industrial sites.
The sectors it targets—kitchens, mines, freight—have controlled environments, repetitive tasks, and high labor costs. They’re the ideal market for specialized robots with immediate ROI. Full article here.
AI that moves things in the real world is already more valuable than AI that generates text.
For sales teams tired of cold leads, slow customer responses, and manual processes, Dapta is the ultimate tool.
Dapta is the leading platform for creating AI sales agents specifically designed to increase inbound lead conversion. Respond to your leads in less than a minute with voice AI and WhatsApp that converts.
If you want your team to sell more while AI handles the complex stuff, you have to try it.
4: Capital keeps flowing into AI
Legora, a Swedish legal AI startup, raised $550M in a Series D, tripling its valuation to $5.55 billion (Bloomberg) in less than a year. Meanwhile, Moonshot AI (the Chinese startup behind the Kimi agent) is looking to raise $1B at an $18 billion valuation. Two massive rounds, two different geographies, one clear signal.
80% of legal tasks are already within reach of current models, yet real adoption is only around 15% (Menlo Ventures). That’s not a problem—it’s the size of the market still to be captured. Investors know it, and they’re entering before the curve accelerates. Full article
If you’re not building in AI today, you’re building in the wrong industry.
5: Anthropic’s marketing team is one person
Anthropic’s entire growth marketing operation (social media, campaigns, PR, global analytics) was run by a single person for 10 months. Austin Lau used Claude Code to automate what previously required a team of specialists, reducing a 30-minute ad process to 30 seconds.
This is at a 2,500-employee company that ran a Super Bowl campaign. Full story here.
The new standard isn’t “how many people are on your team,” it’s how much output each person generates. Managers still sizing teams with 2019 logic are overinvesting in headcount and underinvesting in tools.
A team of 10 using AI well now outperforms a team of 50 that doesn’t.
Bonus: Lovable reached $400M ARR with 146 employees
Lovable, the Swedish vibe coding platform, crossed $400M in ARR in February 2026. In the last month alone, it added $100M in revenue—with just 146 employees. For context: in July 2025 it was at $100M ARR; in November $200M; in January $300M. Full article.
The game-changing metric: $2.77M ARR per employee, already surpassing Gartner’s projection that new unicorns would reach $2M per employee by 2030. The benchmark for “success” has been rewritten—what the market values now is revenue velocity, not the number itself.
Speed is the strategy. The global market is the minimum viable market.
To wrap up
Agents doing real work, robots in kitchens and mines, teams of one operating like twenty, companies laying off thousands to buy GPUs—everything points in the same direction.
The question I’m left with this week: are we heading into an orderly transition where AI creates new roles as fast as it eliminates old ones—or are we normalizing a concentration of economic power in the hands of those who already have access to the best tools?
