<p><strong>TLDR:</strong> Meta is laying off 8,000 employees and cancelling 6,000 open roles while committing $115–135B to AI infrastructure this year. Microsoft launched its first-ever voluntary retirement buyout, targeting roughly 8,750 US employees using a “Rule of 70” formula. OpenAI lost its CPO, Sora creator, and enterprise CTO on the same day it shut down Sora. The broader 2026 tech layoff wave has now hit over 96,000 workers — about 40% more than the same period last year.</p>
<hr>
<h2>Meta cuts 8,000 jobs while cancelling 6,000 open roles</h2>
<p>Meta is laying off 10% of its workforce starting May 20, with notifications going out after details leaked ahead of the official memo. The company is simultaneously cancelling 6,000 open job postings — and yet still has somewhere between 600 and 1,000 roles open right now. The cuts are structural, not performance-based, reorganizing teams into AI-focused pods as Meta redirects spend toward AI infrastructure. One detail worth sitting with: Meta is also deploying a program called the Model Capability Initiative that captures employee keystrokes, mouse movements, and screenshots to train AI agents, with no opt-out option. Internal reaction has been sharp — developers who’ve never spoken negatively about the company before are now looking for exits.</p>
<h2>Microsoft’s first-ever voluntary retirement buyout</h2>
<p>Microsoft launched a voluntary retirement program for the first time in its 51-year history, using a “Rule of 70” formula: age plus years of service must equal 70 or more. The program targets employees at senior director level and below, potentially affecting up to 8,750 people — about 7% of US headcount. This is genuinely new territory: voluntary buyouts have historically been a tool for manufacturers and utilities, not software companies. The formula disproportionately targets long-tenured, higher-salaried employees — people who’ve accumulated both institutional knowledge and expensive compensation packages. That’s not a coincidence. Meanwhile, Microsoft reported $81.3B in quarterly revenue and is spending $37.5B per quarter on AI infrastructure. The financial picture isn’t distress — it’s deliberate reallocation.</p>
<h2>OpenAI loses CPO, Sora creator, and enterprise CTO in one day</h2>
<p>Three senior OpenAI executives departed on the same day: chief product officer Kevin Weil, Sora creator Bill Peebles, and enterprise CTO Srinivas Narayanan. The exits coincide with OpenAI shutting down Sora — reportedly costing around $1 million per day in compute — and folding its OpenAI for Science team into other groups. Only 2 of 11 co-founders remain. The leadership team is now built almost entirely around revenue and operational execution, and OpenAI is pulling back from consumer moonshots in favor of enterprise AI products. The financials explain the tension: $25B in annualized revenue against projected losses of $14B this year.</p>
<h2>Google commits up to $40B to Anthropic as Claude dominates enterprise</h2>
<p>Google is investing up to $40B in Anthropic — $10B cash at a $350B valuation plus $30B conditional on performance targets — alongside 5 gigawatts of compute over five years. The deal comes as Anthropic’s ARR surged from $1B to $30B in 15 months, with Claude holding 32% of the enterprise LLM API market. Amazon has simultaneously committed up to $33B to Anthropic, meaning both cloud giants are competing to be the primary infrastructure provider for the same company. Google’s investment is structured to avoid antitrust review (capped at 15% ownership, no board seats) while ensuring Anthropic’s compute spend flows through Google’s TPU infrastructure. Internally, Google DeepMind has formed a strike team to close the gap with Claude in coding, where Claude Code already generates $2.5B ARR.</p>
<hr>
<h2>Also notable</h2>
<ul>
<li><strong>Tokenmaxxing at Meta, Microsoft, Salesforce:</strong> Meta’s internal “Claudeonomics” leaderboard tracked token usage across 85,000 employees, resulting in 60.2 trillion tokens burned in 30 days — potentially worth $900M at retail API prices. The leaderboard was taken down after media coverage, but similar dashboards with minimum spend targets exist at Microsoft and Salesforce, driving engineers to ask AI pointless questions to hit metrics.</li>
<li><strong>“Great Stay” reshaping hiring:</strong> 62% of workers now prioritize long-term job security over higher pay, job-switching appetite is at a five-year low, and the public sector is seeing a rare opportunity to attract tech talent as private sector instability drives candidates toward stable government roles.</li>
<li><strong>Junior developer squeeze:</strong> Entry-level tech job postings are down 34% since 2022, South African fintech junior salaries dropped 26% year-over-year, and 48% of tech leaders are prioritizing senior hires. The pipeline gap this creates won’t show up for 10–15 years — when today’s seniors retire and there’s nobody behind them.</li>
<li><strong>METR AI task horizon study:</strong> Researchers found AI’s 50%-task-completion time horizon has doubled every 7 months since 2019, reaching 110 minutes with o3. Extrapolating the trend puts month-long software tasks within AI reach by mid-2029, though the 80% reliability horizon is 4–6x shorter.</li>
<li><strong>Vibe coding reckoning:</strong> Multiple studies document rising technical debt (code cloning up 4x, refactoring collapsed from 25% to ~3% of changes), a METR study finding AI actually increased task completion time by 19% despite developers believing it helped, and only 33% of developers trusting AI output accuracy — down from 43%.</li>
<li><strong>Meta employee surveillance:</strong> Meta’s Model Capability Initiative captures mouse movements, clicks, keystrokes, and screenshots from US employees’ work computers to train AI agents. EU and UK employees are exempt due to data protection laws requiring explicit consent.</li>
<li><strong>LinkedIn CEO steps down:</strong> Ryan Roslansky handed the reins to COO Dan Shapero after six years, during which LinkedIn grew from 700M to 1.3B members and revenue more than doubled from $8B to over $17B annually.</li>
<li><strong>Thinking Machines Lab talent war:</strong> Meta hired five TML founding members including co-founder Andrew Tulloch, whose compensation package is reportedly worth $1.5B over six years. TML has also drawn talent from OpenAI, Anthropic, Waymo, and Cognition, and recently signed a multibillion-dollar cloud deal with Google.</li>
<li><strong>Only 5% of firms seeing AI ROI:</strong> New research shows only 5% of organizations are seeing measurable returns from AI investments, with just 1% considered mature in AI deployment. A separate Gartner survey found 50% of executives plan to abandon workforce reduction goals tied to AI by 2027, as AI agents currently complete only 24% of assigned tasks.</li>
<li><strong>GitHub has had no CEO for nearly a year:</strong> The author frames this as the result of internal Microsoft political battles rather than an oversight, with the implication that a business unit is bleeding out as a side effect of corporate politics.</li>
</ul>


Career growth, hiring, engineering culture, and developer work life.

Career Digest

A wave of major tech industry moves: Meta is laying off 8,000 employees and cancelling 6,000 open roles while redirecting spend to AI infrastructure, and controversially deploying keystroke/screenshot surveillance to train AI agents. Microsoft launched its first-ever voluntary retirement buyout using a 'Rule of 70' formula targeting ~8,750 senior US employees. OpenAI lost its CPO, Sora creator, and enterprise CTO on the same day it shut down Sora, signaling a pivot from consumer moonshots to enterprise revenue. Google committed up to $40B to Anthropic as Claude captures 32% of enterprise LLM API market. The broader 2026 tech layoff wave has now hit over 96,000 workers. Additional signals include a junior developer hiring collapse, AI productivity studies showing mixed or negative results, only 5% of firms seeing AI ROI, and a 'Great Stay' reshaping workforce dynamics.

Meta and Microsoft cut tens of thousands of jobs, OpenAI loses three executives in one day