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: Claude goes pro – memory, documents, and enterprise AI
This week, Claude by Anthropic made a major leap toward becoming more than just a chatbot — it’s shaping up to be a serious enterprise assistant. Three key updates signal this shift: persistent memory, built-in document editing, and a potential integration with Microsoft Office.
First, Claude now includes automatic memory for Team and Enterprise users. This means it can remember your preferences, project context, and even how your team works — reducing repetition and making collaboration smoother. It’s not just answering questions anymore, it’s actually following along.
Second, Claude can now create and edit files directly in chat — from Excel spreadsheets and Word docs to PowerPoints and PDFs. It even runs code and performs data analysis in a private computing environment, turning it into a full-blown productivity companion.
Third, reports say Microsoft may integrate Claude into Office 365, alongside OpenAI’s tools. Insiders suggest that Claude’s Sonnet 4 model actually performs better than GPT-5 for tasks like Excel automation and presentation building — marking a shift toward more pragmatic, multi-model AI strategies.
Taken together, these upgrades show that Claude is entering a new phase — one where AI tools don’t just assist your work, they do it. From memory to multi-document workflows, this might be the clearest sign yet that the AI assistant race is accelerating — and it’s not a one-player game anymore.
LLMs & AI Models
- Claude gets smarter and more useful, with memory, document tools, and a possible Office 365 integration — see Focus Topic for full breakdown.
- OpenAI published new research on why AI models hallucinate, arguing that current training rewards confident guesses more than honest uncertainty. They propose a new system that penalizes false confidence more heavily than “I don’t know” answers. This shift could pave the way for more trustworthy AI systems. I will be diving deeper into this topic in a separate post this week.
New Tools
- AI solved a complex math theorem in just 3 weeks using 25,000 lines of verified code — a task that had stalled top mathematicians for over 18 months. Check out the Gaussmodel.
- Alibaba released Qwen3-Next, a hyper-efficient 80B parameter hybrid model that tops the performance of Qwen3 at a fraction of the training cost.
- Adobe introduced the AI Agent Orchestrator program with six specialized agents, including Audience Agent and Journey Agent, to automate CX and marketing.
- ElevenLabs introduced Voice Remixing, a new feature allowing users to change aspects of both designed and user voices for more customizable outputs.
- Stability AI launched Stable Audio 2.5, a new audio model for enterprise-grade outputs that can create three-minute tracks in seconds.
- Google introduced learning tools to NotebookLM, including customizable flashcards and quizzes, redesigned reports, and interactive notebooks for academic textbooks.
Other Quick Picks
- Albania appointed the world’s first AI-powered minister, a virtual avatar managing public procurement to improve transparency and meet EU standards.
- Ethan Mollick warns AI is becoming a black box, with powerful systems like GPT-5 Pro offering results users can’t fully understand or control.
- OpenAI backs ‘Critterz’, an AI-generated animated film, aiming to reach Cannes with a full-length movie made using GPT-5 and AI image tools.
- Amazon is developing smart glasses for consumers and delivery workers, aiming to compete with Meta and launch logistics-focused AR in 2026.
- Reddit, Yahoo, and Medium launched a new licensing standard, giving publishers control over how AI companies can use their content
🇪🇪 AI News from Estonia
- R8 Technologies’ AI ‘Jenny’ cuts energy use by 20% in office buildings by optimizing systems like ventilation and refrigeration in real time. You can listen about it in AI faktor podcast.
- Creem raised €1.8M to automate compliance for AI startups, helping 343 clients scale finance operations without platform fees.
- Estonia created just 29 startups in 6 months, marking a record low; experts cite lack of early-stage capital and fewer new founders. Read more.
- 53% of Estonian employees don’t feel expected to use AI, and 63% haven’t received guidance — a gap Helmes says could risk regional competitiveness. Read more.
- Pactum AI raised $54M and ranks 6th among Estonia’s top tech firms, preparing to bring autonomous contract agents to local companies.
- AI Literacy Day at TalTech will bring together 300 Estonian leaders on Sept 29 to boost AI skills, strategy, and ethics with a 90-day roadmap.
🎙️ AIPowerment Podcast dropped a new end-of-month episode featuring Annika Helendi, a marketing automation expert discussing AI tools for content and B2B growth.
🎧 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, 21.05.2025-28.11.2025 – information
- Training “AI võimalused projektijuhtimises” @ Äripäeva Akadeemia, 17.10.2025-28.11.2025 – information
- Speaking at Äripäev Best Internet Conference, 19.09.2025 – information
- Speaking at The Institute of Internal Auditors Estonia annual conference, 01.10.2025 – information
