Sommaire
Meta’s race to dominate artificial intelligence is turning into a grind for the people building it.
As CEO Mark Zuckerberg pushes the company to ship new AI models and features faster, trying to keep pace with rivals like OpenAI, Google, and Microsoft, employees are describing an internal climate that feels increasingly punishing, with tighter deadlines, longer weeks, and more visible top-down pressure, according to reporting by French tech outlet 01net.
The complaints aren’t just about working hard. They’re about a culture shift: teams say they’re being asked to move at startup speed while carrying the heavy responsibilities that come with deploying AI at the scale of Facebook, Instagram, and WhatsApp.
A faster clock inside Meta
Meta has spent the past two years reshaping itself, cutting jobs, reorganizing teams, and refocusing on efficiency. Now, the company’s center of gravity has swung toward generative AI, both for internal tools and consumer-facing products.
On the ground, that pivot is showing up as more frequent performance targets, quicker decision-making, and less tolerance for delays, the report says. Employees describe a work tempo that has intensified as leadership tries to prove Meta can compete with the industry’s AI front-runners.
Speed vs. safety, and workers absorb the impact
The central tension is straightforward: Meta wants to ship quickly to stay credible in a market where flashy demos and rapid releases can move stock prices and shape public perception.
But “industrializing” AI, turning research into products used by billions, comes with brutal constraints: data quality, security, privacy protections, model reliability, and guardrails to limit harmful outputs. When those demands collide with aggressive timelines, employees say the burden lands on teams expected to stretch their availability and work longer weeks, sometimes around 60 hours, per the accounts cited by 01net.
Internal competition for GPUs, talent, and budget
Meta’s AI strategy also depends on constant proof: showing measurable gains, rolling out new features across its apps, and demonstrating it has the computing power and research chops to match the leaders.
That creates internal competition among projects for scarce resources, especially high-end GPUs, specialized AI talent, and budget. When priorities shift quickly, teams can feel like they’re rebuilding the same work multiple times, a recipe for burnout and frustration.
Not just a Meta problem, but Meta’s scale raises the stakes
The pressure isn’t unique to Meta. Across the tech industry, the current AI boom has the feel of a gold rush, with compressed schedules and intense demands on engineers, data scientists, and product managers.
What makes Meta different is its size and visibility. Decisions made inside the company don’t just affect a single app, they can ripple across platforms used by a huge share of the world’s internet users. That amplifies reputational risk and increases the need for internal justification when things go wrong.
The retention risk: AI talent can walk
In the short term, Meta’s challenge is keeping its speed without damaging its ability to retain the people who can actually build the future it’s promising.
Top AI workers are scarce, expensive, and highly mobile. A deteriorating workplace climate can trigger departures, sap morale, and break continuity on long-running efforts like model training, infrastructure optimization, and responsible deployment, areas where stability can matter as much as raw technical skill.
A culture built on “move fast” meets a more fragile technology
Meta has long sold itself on rapid execution and constant iteration. But AI systems are more complex, more costly, and more sensitive than the social features that once defined the company’s “move fast” ethos.
The trade-offs, speed versus safety, novelty versus reliability, innovation versus compliance, hit employees first. The accounts highlighted by 01net suggest those choices are increasingly being made in ways that raise internal strain, leaving leadership with a looming decision: keep accelerating, or rebalance before exhaustion becomes a business problem.
https://www.europe-infos.fr/actualites/9006/arret-maladie-plafonne-des-le-1er-septembre-2026-31-jours-au-depart-62-en-prolongation/
