Business Latest | Meta’s New Reality: Record High Profits. Record Low Morale by Paresh Dave, Lauren Goode, Steven Levy, Zoë Schiffer

AI generated summary, Read the full article for complete information.

Meta is preparing to cut about 10 % of its workforce—roughly 8,000 jobs—adding to the 25,000 layoffs announced over the past four years, and employees describe the atmosphere as “historically low” morale, with unhappiness widespread except among senior executives. The dip in morale stems from a combination of factors: shrinking compensation (share‑based raises cut for two consecutive years), widening pay gaps, costly AI initiatives, mandatory installation of tracking software that records U.S. staff’s keyboard and mouse activity to train internal AI models, and recent legal setbacks that highlighted the company’s product harms. Workers are protesting the AI surveillance, some are trying to form a UK union with United Tech & Allied Workers, and many hope to be laid off to collect severance and health benefits. While Meta’s ad business continues to generate strong profits—nearly $27 billion in Q1—the company is simultaneously spending billions on AI talent and infrastructure, creating tension between high‑level financial performance and the everyday experience of employees who feel pressured, undervalued, and uncertain about their future.

Read more: https://www.wired.com/story/meta-layoffs-bad-vibes-mark-zuckerberg-ai/

#Meta #AndrewBosworth #UnitedTech #AlliedWorkers #ModelCapability #AppliedAI #aipilled #business #business_artificialintelligence

Meta’s New Reality: Record High Profits. Record Low Morale

Next week, Meta is cutting about 10 percent of its staff. WIRED spoke with more than a dozen current and former employees about what it's like inside a company where "everyone is unhappy."

WIRED

Dan McAteer (@daniel_mac8)

GPT-5.4가 추론 도중에 메시지를 보내 추론 궤적을 중간에 바꿀 수 있는 기능을 제공한다고 보고됨. 작성자는 이전 모델에 없던 전혀 새로운 모델 역량이라고 평가하며, 동작 원리로는 모델이 여러 청크로 작업하고 상태를 유지하는 'stateful orchestration loop' 방식일 것이라는 이론을 제시함.

https://x.com/daniel_mac8/status/2031028862897865008

#gpt5.4 #llm #modelcapability #stateful

Dan McAteer (@daniel_mac8) on X

GPT-5.4 allows you to send a message to change the reasoning trajectory mid-reasoning. Afaik, this is a brand new AI model capability that no previous model provides. Theory for how it works: stateful orchestration loop, where the model does work in chunks and holds state.

X (formerly Twitter)