AI News Roundup — April 24, 2026
OpenAI released GPT-5.5 with stronger coding and research capabilities, while Meta announced a 10% global workforce cut to redirect resources toward AI. Anthropic's Mythos model sparked a cybersecurity emergency after reports of an 83% exploit success rate, Tesla tripled its AI capex to $25 billion, and Tencent and Alibaba entered talks to invest in DeepSeek at a $20B+ valuation.
OpenAI Releases GPT-5.5 with Stronger Coding, Computer Use, and Research
OpenAI announced GPT-5.5, its latest model featuring improved coding ability, computer use, and deep research capabilities. The model is rolling out immediately to ChatGPT Plus, Pro, Business, and Enterprise users, as well as Codex — bringing OpenAI one step closer to its vision of an AI "super app" that handles complex real-world tasks end to end.
Meta Announces 10% Global Workforce Reduction to Accelerate AI Focus
Meta plans to cut approximately 8,000 employees — 10% of its global workforce — and will not fill 6,000 open roles, redirecting the savings toward artificial intelligence. The first wave of layoffs begins May 20 with additional cuts expected later in the year, according to an internal employee memo from leadership.
Anthropic's Mythos AI Can 'Hack Nearly Anything' — 83% Exploit Success Rate Alarms Security Community
Former national cyber director Kemba Walden warned that Anthropic's Mythos can discover flaws in "virtually any and every operating system, browser, or other software product," achieving an 83% success rate in exploit creation on the first attempt. Security experts convened in Washington to debate emergency AI security standards, as global financial institutions raised concern about the scale and speed of potential AI-powered cyberattacks.
Microsoft Integrates Claude Mythos Preview into Secure Coding Framework
Microsoft announced it will embed Anthropic's Claude Mythos Preview into its secure coding framework as part of a broad cybersecurity push. The stated goal is to leverage frontier AI for enhanced threat detection and vulnerability response — a notable move given the simultaneous alarm in the security community about Mythos's offensive hacking capabilities.
Tesla Raises 2026 Capex to $25B — Triple Its Prior Budget — for AI and Robotics
Tesla sharply increased its 2026 capital expenditure guidance to $25 billion, up from $8.5 billion in 2025, directing funds toward AI training infrastructure, custom chip design, a new semiconductor research fab in Austin, and robotaxi operations. Shares fell on the announcement as investors weighed the enormous spend increase against near-term profitability concerns.
Tencent and Alibaba in Talks to Invest in DeepSeek at $20B+ Valuation
Chinese tech giants Tencent and Alibaba are in discussions to invest in DeepSeek, the AI startup whose efficient open-source models sent shockwaves through the industry earlier this year. A deal would value DeepSeek above $20 billion, signaling that China's AI ecosystem is consolidating around its homegrown breakout stars.
Google Gemini Enterprise Agent Platform Launches with 200+ Models in Model Garden
Google officially launched its Gemini Enterprise Agent Platform, giving enterprises access to more than 200 models via Model Garden alongside Gemini 3.1 Pro, 3.1 Flash Image, Lyria 3, and the open-weight Gemma 4. The platform is purpose-built to make building, scaling, and governing production-grade agentic AI pipelines a first-class enterprise capability.