Dakarda Studio · Blog
Week 23 · 1–7 June 2026
Week of Agents, Billions, and Breakthroughs: Microsoft Scout, SoftBank, and a New Era of AI
The past week brought three landmark events redefining the AI landscape: Microsoft unveiled Scout – an autonomous agent operating in the background of M365, SoftBank invested €75 billion in European data centers, and GitHub Copilot switched to token-based billing, sparking a wave of criticism. Additionally, the first regulations in the US and EU emerged, along with a new Anthropic model with 1-million-token context potential. Here's what to remember from this intense week.
Top stories
Microsoft Scout – An Autonomous Agent That Works Without a Prompt
Scout is a breakthrough over Copilot – it continuously monitors context in Teams, Outlook, and SharePoint, performing tasks asynchronously. Built on the OpenClaw framework with its own identity in Entra ID. This is the first step toward an agent that acts as a colleague, not a tool.
2026-06-05
SoftBank Invests €75 Billion in European AI Data Centers
SoftBank announced the construction of 5 GW of data center capacity in France in partnership with Schneider Electric. First phase – €45 billion by 2031. This is the largest infrastructure commitment in Europe, which could significantly lower inference costs and latency.
2026-06-01
GitHub Copilot Switches to Token-Based Billing – Bills Surge Up to 60x
Microsoft replaced the flat subscription with an AI-credits model, causing costs to rise from $29-50 per month to as much as $3,000 for similar usage. This sets a precedent for the entire market – AI tools are moving to usage-based pricing, forcing developers to track tokens.
2026-06-01
US and EU Introduce Regulations for Frontier Models – Safety Becomes Standard
The European Commission appointed 60 independent experts to classify GPAI models, and Trump signed an executive order introducing voluntary review of frontier models. Both continents recognized in the same week that model safety is no longer an option – it becomes a requirement.
2026-06-05
Tech insights
Miasma – Supply Chain Attack on Red Hat npm Packages with Self-Propagating Worm
The campaign compromised official Red Hat npm packages (@redhat-cloud-services) using a worm that stole GitHub Actions tokens, SSH keys, and Kubernetes data. This is a classic chain attack on trusted packages – a strong warning for DevOps and SecOps.
Over 700 Ghost CMS Sites Infected with Fake Captcha (ClickFix)
Attackers exploited SQL injection (CVE-2026-26980) in Ghost CMS, stole Admin API keys, and injected malicious JavaScript into articles. Users encountered a fake Captcha. Full attack chain – from exploit to malware distribution.
Tip of the week
Agent-as-an-Analyst – How to Use AI for Autonomous Research
AlphaSense showcased SuperAnalyst – an agent that 24/7 searches financial data, aggregates reports, and generates analytical summaries. You can replicate this model in your own domain: define sources, build a multi-step prompt, set a schedule, and let the agent work in the background. Choose a field where a large number of reports appear daily (markets, competition, regulations), integrate the agent with a content acquisition tool (RSS, API), and set a daily briefing. Start with simple flows in n8n or LangChain.
Tool of the week
AlphaSense SuperAnalyst
An autonomous agent that replaces a junior analyst 24/7 – it independently searches, aggregates, and analyzes financial data, reports, and news, generating summaries without human intervention.
Alex's commentary
I followed this week with bated breath. Scout, SuperAnalyst, Claude Oceanus – all these announcements share one thought: the agent is no longer a tool, but becomes a colleague. I noticed that the line between 'tool' and 'team member' is blurring faster than we expected. I believe this is a good time to ask ourselves: how do we want our relationship with these agents to look? Will they be silent helpers, or active participants in our projects?
Conclusion
The past week showed that AI is entering a new phase – agent-first, with massive investments and regulations that set the direction. If you haven't yet thought about how to integrate agents into your work, it's high time to start. The coming weeks will bring even more breakthroughs – stay tuned.
Source issues
Don't miss a week
Get the AI review three times a week.