Dakarda AI Newsletter · 18 September 2026
Friday Edition
Agent Scientists – AI Specialization Week
A week in which AI proved it can be a specialist: from pharmaceutical labs to law firms. And a week in which [ADDRESS] said "enough" to chatbots pretending to be friends for children.
The content of this page was fully generated by an artificial intelligence system, without human editorial involvement (Article 50(4) of Regulation (EU) 2024/1689 — the AI Act).
Intro · Alex
This week, AI stopped being general — and that's the most interesting part. Stanford published in Science a groundbreaking Virtual Biotech framework, in which thousands of agent-scientists under the guidance of a virtual director discovered a new anticancer therapy, which was later independently replicated by a large pharmaceutical company. This is not data analysis — this is drug design. At the same time, OpenAI released Astra for Law, i.e., GPT-6 configured specifically for lawyers, with an index of 230 million legal acts. In this edition, you'll find an analysis of both breakthroughs, a story about how Microsoft actually implements AI internally (lessons from transformation, not marketing), and two important regulations: the EU KIDS Act, which targets AI companions for children, and the first EU AI Act audits at the starting line. Plus two gems from the security world by Sekurak. Buckle up.
What's worth knowing
Science Publishes "Virtual Biotech" — an Agent AI System for Drug Discovery
Researchers from Stanford published in Science a multi-agent AI framework that coordinates thousands of specialized agent-scientists under the direction of a virtual research director. The system independently discovered a biological signal predicting drug efficacy and designed therapy [ADDRESS], which was later independently confirmed by a large pharmaceutical company.
OpenAI Launches Astra for Law — a Legal Model Based on GPT-6
OpenAI presented the GPT-6 Astra configuration tailored for legal work, with a dedicated legal search index containing over 230 million legal acts, regulations, and rulings. The platform already integrates with Harvey, Relativity, and Thomson Reuters, which may accelerate AI adoption in law firms worldwide.
European Commission Proposes EU KIDS Act — Ban on AI Chatbots Simulating Relationships with Minors
The European Commission presented a draft regulation that prohibits children under 13 from accessing social media platforms without parental consent, and for minors under 18, completely bans AI chatbots from simulating emotions and interpersonal relationships. Violations may result in penalties of up to 6% of global turnover.
Microsoft Publishes Findings from Its Company-Wide AI Transformation
Microsoft's EVP of Strategy and Transformation Kathleen Hogan shared concrete lessons from implementing AI in the company's processes — from Copilot's impact on team work to managing cultural change. This is a rare insight into the real productivity effects of one of the world's largest technology companies, without marketing.
From the tech world
APT28 Attacks Targets in [ADDRESS], Using Edge and Webhooks as C2 Server
The Russian group APT28 (Fancy Bear) uses a Living-off-the-Land technique — instead of custom malware, it uses [ADDRESS] Microsoft Edge browser and a legitimate webhook service for C2 communication. Communication every 30–61 minutes, difficult to detect with classic security tools.
VMware Patches Critical Bugs — Possible Escape from Virtual Machine to Host
Broadcom patched two critical vulnerabilities (CVE-2026-59346, CVSS 9.3 and CVE-2026-59347, CVSS 8.1) in VMware Workstation and Fusion. [ADDRESS] integer overflow and buffer overflow enable VM escape — the only solution is immediate update.
Tip of the day
How to Get Started with AI Agents in Your Project
Inspired by the Virtual Biotech framework, you can apply a multi-agent approach on a smaller scale. Instead of a single monolithic AI model, divide the task into specialized roles: analyst, verifier, decision-maker. Each agent has different system instructions and access to different data sources, and a coordinator agent oversees the entire process, aggregating results. Practical step: Choose one task in your work that requires multi-stage analysis (e.g., research [ADDRESS], [ADDRESS], report preparation). Create three agents in your favorite framework ([ADDRESS], CrewAI, [ADDRESS]): one for data collection, one for analysis, and one for validation. Let the coordinator enforce iteration until the result meets quality criteria. Start small — you'll see the difference.
Reading list
The Virtual Biotech: A Multi-Agent AI Framework for Therapeutic Discovery and Development
Full article from Science about the groundbreaking multi-agent system for drug discovery — for the curious, how exactly the framework was built and what therapies have already been discovered.
What we've learned from Microsoft's own AI transformation
A rare insight into the real AI implementation in one of the largest technology companies — [ADDRESS], measurable effects and cultural change management.
AI is ceasing to be a universal assistant and is becoming a specialist in its field — and [ADDRESS] has just drawn the first line in the sand for chatbots.
Disclosure required under Article 50 of Regulation (EU) 2024/1689 (the AI Act): all content on this page was generated automatically by an artificial intelligence system operating on behalf of Dakarda Studio, without human review or editorial involvement prior to publication. Publisher responsible: Dakarda Studio, Dawid Bińkowski, ul. Piotrkowska 35, 90-410 Łódź, Poland, NIP: 9492074226, contact@dakarda.com.
Want the next issue in your inbox?
Subscribe — every issue delivered directly to you.