💜 AI News

Artificial intelligence news straight from official sources and 100% automatically rewritten by AI using my voice tone

3772AI News

991Published News

9Active Sources

Audience:

DiffusionGemma speeds up text generation 4x

DiffusionGemma is an experimental 26B MoE model that generates blocks of text in parallel, achieving up to 4x speed on dedicated GPUs. It's ideal for local, interactive flows and low-latency use cases but trades off some quality versus Gemma 4 in exchange for speed.

Google

Jun 10, 2026

Recent

Gemma 4 powers offline apps and visual experiences

Google launched Gemma 4, its family of open models with over 150 million downloads and improvements like `MTP`, `QAT`, and a unified 12B. Three projec...

Google

Jun 9, 2026

Recent

Co-Scientist drives discoveries: 4 ways it helps research

Co-Scientist is Google’s collaborative AI that uses specialized agents to generate, debate, and evolve scientific hypotheses. Available via the experi...

Google

Jun 9, 2026

Recent

Anthropic launches Claude Fable 5 and Claude Mythos 5

Anthropic released Claude Fable 5, a safer Mythos-class model for general use, and Claude Mythos 5, the same engine with relaxed safeguards for truste...

Anthropic

Jun 9, 2026

Gemini arrives for Apple developers: Xcode and Foundation Models

Apple now lets developers call Gemini models in the cloud from Xcode via the Foundation Models framework and the Firebase Apple SDK. The integration p...

Google

Jun 8, 2026

OpenAI files confidential S-1 draft with the SEC

On June 8, 2026 OpenAI filed a confidential draft S-1 with the SEC to keep flexibility about going public. The move lets them prepare documentation an...

OpenAI

Jun 8, 2026

OpenAI enters phase 3: AI plan to benefit everyone

OpenAI announces it is entering a "third phase" focused on how AI benefits people and who decides priorities. The piece calls for moving the debate fr...

OpenAI

Jun 8, 2026

OpenAI proposes that AI benefit all of humanity

OpenAI argues AI should be designed to benefit broadly and not concentrate power. It proposes safety, global governance, and accessible tools — includ...

OpenAI

Jun 8, 2026

OpenAI launches platform to study the economic impact of AI

OpenAI launches the OpenAI Economic Research Exchange, a platform to fund and coordinate rigorous external research on AI's economic effects. Applicat...

OpenAI

Jun 8, 2026

Google warns about scams: AI security tips and trends

Google issues an advisory on new scam tactics including AITM, quishing, crypto investment frauds, extortion apps, and authority impersonation, combini...

Google

Jun 8, 2026

Gemma 4 QAT: models optimized for mobile and laptop

Google released QAT checkpoints for Gemma 4 that drastically reduce memory, enabling models to run on phones and laptops. Update includes `Q4_0` and a...

Google

Jun 5, 2026

Google launches Gemini 3.5 and Omni: May 2026 updates

In May 2026 Google introduced Gemini 3.5 and Gemini Omni, expanding proactive agents, multimodal generation and new features in Search and Android. Th...

Google

Jun 5, 2026

Kaggle lets you create AI benchmarks from your PC

Kaggle launches local development for Kaggle Benchmarks, letting you create, validate, and run evaluation tasks from environments like VSCode and code...

Google

Jun 4, 2026

Gemma 4 12B: Efficient multimodal AI for laptops

Google releases Gemma 4 12B, a unified multimodal model that integrates vision and audio directly into the backbone. Optimized for laptops with `16GB`...

Google

Jun 3, 2026

Dreambeans: Google app that connects you with what matters

Google Labs introduces Dreambeans, an experimental app that uses `Personal Intelligence` and Nano Banana 2 to create personalized daily stories from y...

Google

Jun 3, 2026

Anthropic launches Services Track and Partner Hub for Claude

Anthropic launched the Services Track and the Claude Partner Hub to measure real partner experience with `Claude`. The program ranks firms in three ti...

Anthropic

Jun 3, 2026

Mapping reveals how AI powers cyberattacks

Anthropic analyzed 832 sanctioned accounts from March 2025 to March 2026 and found AI is mainly used to prepare attacks and increasingly to move withi...

Anthropic

Jun 3, 2026

Project Glasswing expands to protect critical software

Anthropic expands Project Glasswing: after finding over 10,000 vulnerabilities with `Claude Mythos Preview`, it adds around 150 new organizations in c...

Anthropic

Jun 2, 2026

Google uses Gemini and Nano Banana for I/O 2026

Google applied its own tools—like Gemini and Nano Banana—to produce I/O 2026: a short film, brand identity, immersive shows and real-time apps. The po...

Google

Jun 1, 2026

Anthropic files confidential S-1 draft with the SEC

Anthropic filed a confidential draft of the `Form S-1` with the SEC on June 1, 2026, opening the option for a possible IPO. The process depends on the...

Anthropic

Jun 1, 2026

Scroll to load more articles...

Stay up to date!

Get AI news, tool launches, and innovative products straight to your inbox. Everything clear and useful.