AI Apple quietly registers genai.apple.com subdomain ahead of WWDC26 Top Stories
Pinned News
AI Apple quietly registers genai.apple.com subdomain ahead of WWDC26 
Developer Context Is a Budget — Not a Trash Can 
OpenAI Codex: More Than a Coding Assistant, It's a Work Automation System 
Anthropic Why Claude Code Is Ditching Markdown for HTML 
Anthropic Elon Musk shuts down xAI, hands its 220,000 GPUs to Claude 
AI Musk's lawyers may have just blown open the xAI-OpenAI probe 
OpenAI The decade of OpenAI: ambition, conflict, and betrayal News,Images,Videos
Most Recent
DeepSeek V4 sparks a new US AI fight: banned chips or open innovation?
A former White House official says DeepSeek's V4 model is behind and likely using restricted chips. A Silicon Valley CEO argues it's a genuine open-source breakthrough. The battle over China's AI capabilities is heating up again.
Doubled training data to 33T is why DeepSeek V4 is late, says DeepMind researcher
DeepSeek V4's technical report reveals training instability after doubling training data to 33T tokens. A DeepMind researcher says this explains the delay, calling the fixes 'band-aids.'
Xiaomi AI lead reveals 1 trillion parameter model, says AGI is two years away
In a rare interview, Xiaomi's head of large language models reveals a 1-trillion-parameter model, a no-deadline team culture, and a prediction that AGI is just two years away.
DeepSeek V4-Pro tops GPT-5.4 on coding, but lags in knowledge and long context
DeepSeek's open-source V4-Pro-Max set a record on Codeforces, beating GPT-5.4, but fell short on knowledge and long-context benchmarks against Gemini and Opus.
DeepSeek V4 makes 1M context standard across all models
DeepSeek's new V4 series makes 1M-token context standard across all tiers, with V4-Pro matching top closed-source models in reasoning and openly acknowledging gaps against Anthropic's Opus 4.6.
DeepSeek drops V4 preview
DeepSeek open-sources its V4 preview under MIT license: 1.6 trillion parameters, 1M token context, and memory usage slashed to a tenth of V3.2.
Xiaomi's new TTS model sings, clones voices, and takes mood requests
Xiaomi's new TTS models can sing, clone voices from seconds of audio, and adjust emotion via natural language — all free during the public beta.
Claude agents get cross-session memory, Rakuten errors drop 97%
Anthropic adds cross-session memory to Claude Managed Agents, letting agents learn across conversations. Rakuten reports 97% fewer first-pass errors, 27% lower costs, and 34% lower latency.
Anthropic finally explains why Claude got dumber — and says it's fixed
Anthropic blames three compounding changes — a default thinking duration tweak, a caching bug, and an overzealous system prompt — for Claude's recent intelligence drop, all fixed as of April 20.
OpenAI offers $25,000 for a single prompt that breaks GPT-5.5's biosecurity
OpenAI is offering $25,000 to anyone who can jailbreak GPT-5.5's biosecurity protections with a single prompt. The challenge runs through July.