Gemini user hits 5-hour usage cap after a single prompt, Google responds
A Google AI Pro subscriber exhausted their five-hour usage cap with a single failed video-generation prompt, underscoring challenges with the new compute-based limit system.
Section
AI
Editorial section. Other sections and tags are in the Topics menu; for full-text use search.
-
-
Reallusion bets that 3D artists directing AI models will beat text prompts for professional filmmaking
Reallusion launches AI Studio, a hybrid platform combining 3D animation with generative AI models to give filmmakers spatial precision that text-prompt-only generators cannot match.
-
Gemini’s camera AI misidentifies Australian wildlife and cats as raccoons
Reddit users in Australia report that Google Gemini mislabels cats as raccoons, kangaroos as people, and calls utes trucks, exposing gaps in the AI’s regional training data.
-
Google's latest attempt to fix token quotas is here: Say hello to Gemini 3.5 Flash Low
Google introduces Gemini 3.5 Flash (Low) model that uses 45% fewer tokens and resets quotas across all plans to address user concerns in Antigravity.
-
How video helps build robot brains for physical AI
Anaxi Labs crowdsources human-scale videos to train robot brains, addressing physical AI's data scarcity with egocentric industrial and household scenarios.
-
Claude Design replaces my presentation workflow — and I'm not going back
Tech writer Parth swapped PowerPoint and Google Slides for Claude Design, finding Anthropic's AI tool generates polished, on-brand decks from raw prompts — with pricing starting free.
-
Nvidia CEO urges Supermicro to strengthen export compliance after $2.5B smuggling bust; Taiwan cracks down on AI GPU smuggling
Nvidia CEO Jensen Huang calls on Supermicro to improve export compliance after a $2.5B smuggling bust involving AI chips, while Taiwan initiates its first crackdown on illicit AI hardware exports to China.
-
Google search is becoming something fundamentally different
Google announced at I/O 2026 a revamped, AI‑first Search with Gemini 3.5 Flash as the default engine, an intelligent search box, and new agentic widgets that will roll out this summer worldwide.
-
DeepSeek makes 75% discount permanent, sparking new AI price war
DeepSeek makes its 75% discount on V4 Pro permanent, pricing output tokens at $0.87 per million and undercutting GPT‑5, Gemini and Claude, prompting a fresh AI pricing battle.
-
Apple prepares new generative AI subdomain ahead of WWDC 2026
Apple's registration of a new genai.apple.com subdomain points toward a massive AI-centric showcase at the upcoming WWDC 2026 conference.
-
007: First Light Drops Last-Minute Denuvo DRM Bombshell, Leaving Pre-Order Customers and Fans Furious
007: First Light's last-minute Denuvo DRM addition sparks backlash over performance issues and day-one piracy cracks, highlighting tensions between security and consumer experience.
-
Berlin's Peec AI more than doubled revenue to $10M ARR in six months. Its product helps brands show up in ChatGPT.
Berlin-based Peec AI has crossed $10 million in annualized revenue six months after raising $21 million Series A, helping brands optimize visibility in AI chatbot responses as consumer behavior shifts from search…
-
Anthropic says mythos preview has already uncovered more than 10,000 vulnerabilities
Anthropic's Claude Mythos Preview has helped partners uncover over 10,000 software bugs in its first month, with cloud and browser giants reporting thousands of high‑severity fixes, while the company eyes its first…
-
Claude Design is the best feature Anthropic has launched in a while, but it has some serious limitations
Anthropic's new Claude Design tool generates full designs from prompts, but a separately metered quota and hidden prompt costs on every edit are leaving even Max 5x subscribers hitting weekly limits by mid-week.
-
Towards speed-of-light text generation with Nemotron-Labs diffusion language models
NVIDIA's Nemotron‑Labs Diffusion family adds diffusion and self‑speculation modes to 3B‑14B LLMs, delivering up to 6.4× faster token generation while keeping accuracy competitive.
