Meta reshapes products and staff

New platform signals include WhatsApp device‑linking warnings and Facebook alerts for suspicious friend requests, features Meta described as part of an AI‑driven anti‑scam package. (about.fb.com) Meta’s trust metric disclosures show the company removed 159 million scam ads in 2025, reported banning 12.1 million ad items in India and said over 93% of those removals were proactive. (about.fb.com) A coordinated law‑enforcement operation that Meta said it supported disabled more than 150,000 accounts tied to Southeast‑Asia scam centres and led to 21 arrests, with partners that included the Royal Thai Police and U.S. agencies. (about.fb.com) The next‑gen model codenamed “Avocado” was pushed from March to at least May after internal testing placed its performance between Google’s Gemini 2.5 and Gemini 3, according to reporting that cited people familiar with the matter. (money.usnews.com) Company leadership has discussed a temporary license to run Google’s Gemini inside some Meta products while Avocado is reworked, a stopgap option various outlets reported in coverage of the delay. (computing.net) Internal planning documents and Reuters reporting show Meta is weighing workforce cuts that could hit about 20% of its roughly 78,865‑person global staff — a scale that would exceed the tens of thousands laid off during its 2022–2023 restructuring. (cnbc.com) Meta has also been buying licensed news content to feed its AI: the company announced broader international publisher integration for Meta AI and struck a multiyear licensing pact with News Corp worth up to $50 million a year. (about.fb.com) Investors and analysts tracked those moves against Meta’s big infrastructure plan after Q4 guidance showed 2026 capital expenditures of $115–$135 billion, and shares slid in the days after the Avocado delay reports. (investor.atmeta.com)