AI News

Companies are excited about autonomous AI agents, but they often overlook the hidden costs. Find out why agentic AI can unexpectedly drain your budget and how to optimize tokens using models like Gemini 3.5 Flash.

British giant Capita is pivoting from traditional outsourcing toward "agentic AI". With a Snowflake partnership and AWS leadership, it's building infrastructure that enables AI not just to respond, but to act independently. What does this mean for businesses and EU regulations?

Anthropic surprised everyone. Just days after warning about the risks of overly powerful AI, it launched Claude Fable 5. What does this mean for safety and what impact does it have on users in Europe?

A study in the journal Radiology tested 10 large language models as translators of radiology reports. GPT-4 came out on top, but accuracy of medical terminology still lags behind. What does this mean for patients?

French tech giant Atos has announced the largest-ever investment in technology for its workforce — Microsoft 365 Copilot will be rolled out to every one of its 56,000 employees across 54 countries. Security, management of tens of thousands of AI agents, and European sovereignty form the backbone of a strategy that also reaches the company's Czech branch.

Meta surprised with a fundamental shift. The Muse Spark model replaces Llama 4 in smart glasses, betting on native multimodal reasoning and extreme efficiency. What does this mean for users and the Czech market?

No more counting licenses for every employee. Boston-based Creatio has unveiled Unlimited Enterprise — a model where companies pay for unlimited performance instead of per user. The CRM platform now includes unlimited AI agents, workflows, and apps with no extra fees.

Former OpenAI CTO Mira Murati has unveiled her new startup Thinking Machines Lab. In her first interview since leaving OpenAI, she revealed details about a new model focused on efficient reasoning and safety. What does this mean for the market?

Anthropic launches Claude Fable 5, the first Mythos-class model for general availability. At $10/$50 per million tokens, it outperforms every frontier model on the market while pioneering a new safety architecture with three AI classifiers that block misuse — but sometimes catch harmless requests too.

GitHub Copilot is no longer just smart code completion. From agent mode and cloud agents to its own SDK — we'll show you what this platform can do in June 2026, how much it costs, and how to get the most out of it.