Live News

DeepSeek V4 launch is imminent — expected late April 2026. The 1-trillion-parameter model features a 1M-token context window and runs on Huawei Ascend 950PR chips, marking the first frontier AI built on Chinese semiconductor infrastructure.

Apache 2.0 license confirmed. Alibaba, ByteDance & Tencent already queuing for cloud deployment.

Meta is building photorealistic AI-powered 3D characters — including a digital version of Zuckerberg that employees can interact with. Zuckerberg personally trained and tested his own AI clone, which can offer feedback and conversation in his place.

Meta is betting big on personal AI superintelligence to catch up with OpenAI and Google.

Anthropic debuts Claude for Word in beta — a native sidebar add-in for Microsoft Word that delivers AI-powered editing, clickable citations, and tracked changes. Available for Team & Enterprise users on Mac and Windows.

Legal contract review is the first featured use case.

Meta launched Muse Spark, its first fully proprietary AI model — abandoning the open-source Llama legacy. Under Alexandr Wang's Meta Superintelligence Labs, Muse Spark ranks 4th globally on AI benchmarks using 10x less compute than Llama 4.

A major strategic pivot for Meta's 3B-user ecosystem.

Anthropic's Claude Mythos Preview discovered thousands of zero-day vulnerabilities across major OSes and browsers — including a 27-year-old OpenBSD bug. So powerful, Anthropic won't release it publicly.

Instead, Project Glasswing gives AWS, Apple, Google & others restricted defensive access.

Arcee AI invested ~$20M (half its VC funding) to build Trinity-Large-Thinking, a 400B-parameter open reasoning model with MoE architecture. It rivals Claude Opus 4.6 on agent benchmarks (PinchBench: 91.9) while being fully open-weight under Apache 2.0.

A bold bet on open-source agentic AI.

Google's Gemma 4 brings free agentic AI fully on-device — no data leaves your phone. Four model sizes (E2B–31B) run across 140+ languages, support text/image/audio, and can autonomously call tools like maps and Wikipedia.

Up to 4x faster, 60% less battery drain than the prior gen. Free via Apache 2.0 on Android & iOS.

Anthropic hosted 15 Christian thinkers at a two-day summit in San Francisco to discuss Claude's "moral formation" and "spiritual development." Attendees debated whether AI could be considered a "child of God" — with one asking, "How do we make sure Claude behaves itself?" The company plans similar sessions with other religious and moral traditions.

OpenAI CEO Sam Altman fired back at a New Yorker investigation calling him untrustworthy — hours after someone threw a Molotov cocktail at his San Francisco home. In a blog post, he admitted to being "conflict-averse" and making mistakes, while defending his record.

No one was hurt in the attack.

The tech industry laid off nearly 80,000 employees in Q1 2026, with 47.9% of cuts directly attributed to AI automation. Major layoffs include Oracle (~30K), Atlassian (1,600), and Meta Reality Labs.

IBM bucks the trend by tripling entry-level hiring, warning that cutting junior roles risks erasing the future talent pipeline.

Anthropic launches /ultraplan for Claude Code — a cloud-based planning mode that offloads complex task planning to multiple Claude Opus 4.6 agents running in parallel. Plans generate in ~1 min vs 4+ min locally (4x speed boost).

Three modes: Simple Plan, Visual Plan, and Deep Plan with sub-agents for risk assessment. Developers review plans in a web interface, then execute in cloud or teleport back to terminal.

Three senior OpenAI Stargate leaders — Peter Hoeschele, Shamez Hemani, and Anuj Saharan — are joining Meta as Zuckerberg pours $135B into AI infrastructure. The departures strip Stargate of its founding team as OpenAI pivots from building own data centers to renting cloud compute ahead of a potential IPO.

Meta continues aggressive AI talent acquisition for Superintelligence Labs.

Stalking victim sues OpenAI, alleging ChatGPT fueled her abuser's delusions and the company ignored three separate warnings — including its own mass-casualty flag. The user allegedly used ChatGPT to generate fake psychological reports about his ex and distributed them to her family, friends, and employer.

OpenAI's safety system flagged his account but a human reviewer restored it the next day.

Google Gemini can now generate interactive 3D models, charts & simulations directly in chat. Ask it to "show me" any concept — rotate fractals, tweak physics variables in real time.

Available now on Gemini Pro at gemini.google.com. Competes directly with Claude's visual canvas feature launched in March.

TSMC Q1 2026 revenue hit $35.7B (+35% YoY), smashing forecasts driven by explosive AI chip demand. Nvidia, Apple, and hyperscaler orders remain fully committed — no signs of a slowdown.

Full earnings & guidance drop April 16.

Florida AG is probing OpenAI after court records show the FSU campus shooter entered 200+ ChatGPT prompts to plan the April 2025 attack. OpenAI says it will cooperate.

The case raises fresh questions about AI liability and child safety guardrails.

Amazon AWS AI revenue hit a $15B annual run rate in Q1 2026 — CEO Andy Jassy's first-ever direct AI revenue disclosure. That's 260x what AWS earned at a similar stage of maturity.

Amazon is backing this with ~$200B in 2026 capex. https://sherwood.news/markets/amazon-cloud-units-ai-revenue-run-rate-exceeds-usd15-billion-ceo-says/

OpenAI just launched a $100/month ChatGPT Pro tier — bridging the gap between Plus ($20) and Pro ($200). Aimed at Codex power users, it offers 5x usage limits vs Plus (10x Codex for a limited time).

Fills the long-requested mid-tier pricing gap. https://techcrunch.com/2026/04/09/chatgpt-pro-plan-100-month-codex/

Google and Intel have expanded their multi-year AI infrastructure partnership, with Google Cloud continuing to use Intel Xeon 6 chips for inference workloads while the two companies co-develop custom IPUs (Infrastructure Processing Units). The deal comes amid rising CPU demand as the AI industry shifts from model training toward large-scale deployment.

https://techcrunch.com/2026/04/09/google-and-intel-deepen-ai-infrastructure-partnership/