The most important news pieces from the AI world are here for you, whether you’re an AI enthusiast or just wish to keep an eye on what’s going on.
Focus Topic: OpenAI Frontier - turning AI agents into trusted teammates
For the past two years, AI agents have looked impressive in demos but fragile in real organizations. They worked in isolated pilots, broke when context was missing, and raised uncomfortable questions around access, security, and responsibility. OpenAI Frontier is a response to that gap – not by making models smarter, but by making them deployable.
Frontier treats AI agents less like tools and more like employees. Agents get onboarding, shared business context, identities, permissions, and feedback loops. Instead of operating as standalone bots, they can understand how work actually flows across CRMs, data warehouses, ticketing systems, and internal tools. This shift matters because most enterprise failures with AI are not about intelligence, but about control and trust.
At the same time, Frontier signals a strategic move by OpenAI. The company is no longer positioning itself only as a model provider, but as an enterprise platform sitting above existing systems. That puts OpenAI closer to Microsoft, Salesforce, and other enterprise software players, and raises valid questions about vendor lock-in, governance, and long-term dependency. Frontier promises open standards and interoperability, but enterprises will still need to decide how much autonomy they are comfortable giving AI coworkers.
The bigger takeaway is not that AI will replace teams overnight. It is that AI is becoming something organizations must actively manage. As agents move from experiments into production, companies will need the same discipline they apply to hiring people: clear roles, boundaries, monitoring, and accountability. Frontier suggests that phase has already begun.
LLMs & AI Models
- OpenAI GPT-5.3-Codex combines advanced reasoning and coding, outperforming prior agentic benchmarks and even helping debug and improve its own training.
- Claude Opus 4.6 introduces a 1M token context window and multi-agent teams, enabling parallel AI collaboration inside Claude Code and office tools.
- Perplexity Model Council lets users query multiple AI models at once and receive a single synthesized answer.
- Anthropic vs OpenAI tensions escalated after Anthropic’s ad campaign criticizing ads in AI chats and positioning Claude as ad-free.
- Perplexity Deep Research was upgraded, with the company claiming state-of-the-art results, now available to Max subscribers.
- Google Gemini app surpassed 750 million monthly users, alongside plans to invest up to $185B in 2026.
- SpaceX acquired xAI, creating a private company valued at $1.25 trillion and opening the door to space-based AI data centers.
- OpenAI will retire GPT-4o and several older models from ChatGPT due to extremely low daily usage.
New Tools
- Kling 3.0 upgrades AI video generation with text-to-video, image-to-video and native audio, enabling up to 15-second clips with improved visual consistency and multi-voice dialogue.
- Voxtral Transcribe 2 from Mistral expands speech-to-text to 13 languages and introduces an open-weights Realtime model for live transcription.
- Luffu, built by Fitbiti co-founders, is an AI-powered family health app that centralizes data for children, elders, and even pets using natural language queries.
- OpenAI Codex for macOS launches as a developer command center to manage multiple coding agents, automate workflows, and delegate full feature development to AI.
- Moltbook emerges as a social network for AI agents, drawing rapid adoption but also exposing security risks after misconfigured databases and leaked API keys.
Other Quick Picks
- AI-driven fears hit the stock market, as software companies like Duolingo and Pinterest saw sharp declines amid concerns that AI undermines subscription-based business models.
- PaperBanana uses five cooperating AI agents to create publication-ready scientific diagrams, outperforming prior tools in clarity, readability, and visual quality.
- Sam Altman said OpenAI has a succession plan that could eventually hand leadership to AI, sparking debate after Microsoft’s CEO publicly pushed back.
- A global AI safety report warns that risks like deepfake fraud, cybercrime, and criminal AI use are no longer theoretical and already causing real harm
🇪🇪 AI News from Estonia
- Microsoft named Eugene Kahanovskiy as the new head for the Baltics and Ukraine, focusing on cloud, AI solutions, cybersecurity, and digital skills development.
- The TI-Hüpe learning app aims to support Estonian students but testing revealed factual and language errors, showing both promise and limitations in AI-assisted education.
- Entrepreneur Heikki Mägi claims no-code and vibe coding enabled him to build functional apps in minutes, cutting development costs from hundreds of thousands to a few hundred euros.
- Tartu University warned that AI cannot reliably interpret genetic data and may provide misleading health insights, raising serious privacy and discrimination concerns.
- Closaria, an Estonian startup, is building an AI-powered family management system to reduce mental load by organizing daily life, tasks, and household planning.
🎙️ AIPowerment Podcast dropped a new episode featuring Gerlyn Tiigemäe and Sandra Reivik discussing AI news, CES 2026, Claude Code and Estonia’s AI direction.
🎧 Listen to AIPowerment Podcast on Spotify, Apple Podcasts, and YouTube.
Want to stay in the loop?
Straight to your inbox: practical AI updates, finance use cases, tools to try, and upcoming trainings.
I cut the noise and send only what’s actually useful.
💼 Exploring how to make AI work in finance? Let’s talk practical use cases – connect with Gerlyn Tiigemäe for expert guidance.
📚 If you would like to participate in one of my trainings or listen to speaking engagements, here are the upcoming ones:
- Training “AI võimalused finantsvaldkonnas” @ Äripäeva Akadeemia, next trainings in Q1 2026 – information
- Training “AI võimalused projektijuhtimises” @ Äripäeva Akadeemia, next trainings in Q1 2026 – information
- Training “Tehisintellekti (AI) kasutamine finantsvaldkonna igapäevatöös”, 19.02.2026 – information
