Pasi Tuomainen

👋 We’re thrilled to welcome back Heikki Pitkänen to Health Design 2025! Last year, Heikki demoed e-tools for regulatory pathways — this year we want to learn more about the new Lean Entries e-tools that simplify the AI Act, Data Act, and EHDS Regulation. These tools, developed in a joint initiative led by Sitra with partners in Finland and Denmark, provide instant, personalised guidance on complex EU regulations. Join us in exploring how Lean Entries turns regulatory complexity into clear, actionable knowledge: https://lnkd.in/eFRFzuwQ Laura Halenius, Meeri Toivanen LL.M. Ville Peltola, Päivi Ahlgren, Signe Ulrik Holm, Ilona Raitakari, Teemu Moilanen, Youssef Zad, Outi Tuovila, Aurevia, Meeri Säily, Timo Nore, Sandra Liede, Saara Hassinen #HealthDesign2025 #AIAct #DataAct #EHDS #DigitalHealth #Regulation

26

2 kommenttia

🎬 AI is rewriting the script in indie filmmaking! In a recent talk at Amman, expert Aleksi Hyvärinen shared powerful insights on how artificial intelligence is changing the game for independent filmmakers. From generating scripts to streamlining post-production, AI tools are opening doors for creators with limited resources to compete at a whole new level. This is just one example of how AI is revolutionizing industries worldwide—and it's only the beginning. 💡 If you're a business curious about how AI can transform your work, we're offering a FREE AI Automation Assessment. Let us explore what’s possible for you. 📩 DM me or comment “Interested” and we’ll connect! #AIinFilm #ArtificialIntelligence #Automation #DigitalTransformation #BusinessGrowth #IndieFilm #AIforBusiness #FreeAssessment

2

2 kommenttia

Interesting... Although artificial intelligence has rapidly become commonplace, its use is still experimental in many companies. Read more 🔗 https://ppe.link/7v #AI | Työterveyslaitos - Finnish Institute of Occupational Health

1

Recently, Nikita Sobolev highlighted an interesting case around Bun: a large rewrite experiment done with AI agents. A significant part of the runtime was reportedly rewritten from Zig to Rust in a very short time using AI agents. The speed is impressive. But the more important engineering question is not how much code AI can generate. It is what happens after that code exists. Who can understand it? Who can review it properly? Who can own it long term? Who can safely evolve it when the system changes? This is where AI-assisted development becomes an architecture topic, not just a productivity topic. In complex systems, generated code is not automatically valuable code. It still needs clear boundaries, deterministic builds, maintainable structure, and strong technical ownership. AI can help teams move faster. But in serious backend and infrastructure systems, speed without ownership can quickly become a liability. At Kanzu, this is exactly the kind of question we care about: where engineering speed creates leverage, and where it starts creating hidden cost.

2

I recently joined Namya Joshi for a conversation on AI, XR, learning, and the future of work. Many people still ask the same question: Will AI replace experts? I’m more interested in a different angle: Can AI make expert work more measurable, visible and therefore more leadable? In the podcast, I shared how I use AI as a thinking partner. Not a boss, and definitely not a replacement. For example, I set up a 5-agent review board on Discord to challenge my own ideas, expose weak points, and make my reasoning process visible. When we use AI and immersive environments this way, they can: • shorten the distance from a raw idea to a testable experiment • expose where decisions actually take shape • help us safely test our own assumptions The real value for leadership isn’t in the technology itself. It’s in gaining visibility into how thinking evolves and how decisions are formed. If you’d like to hear more about this, future skills, and why I sometimes talk to an AI assistant while driving so I can focus on cooking at home. The full conversation is here: https://lnkd.in/dNAfrbBv

6

2 kommenttia

This short tutorial shows how to create a crypto monitoring dashboard with Dashtera. It's not that hard, is it? 🤗 Thanks Jarkko Tirkkonen for the data, and Elmeri Keinänen for your help with the InfluxDB data feed 🙏 #dataviz #crypto #trading #dashboard #data_analytics #realtime #fintech #charting #tutorial

25

1 kommentti

Finland's innovation fund Sitra has proposed a national health data infrastructure to accelerate AI development in healthcare. The Finnish Health Data Space model would consolidate fragmented data reserves, unify the research permit process, and connect to LUMI supercomputer, Europe's most powerful supercomputer. Pilot schemes are planned for 2026-27, and the proposal aligns with the EU's Health Data Space regulation. Some industry voices have raised concerns about access for smaller companies and whether a purely national approach goes far enough — but the ambition is clear. Fragmented data is one of the biggest bottlenecks for AI in healthcare. Countries that solve it first will have a lasting advantage. Or what do you think? Link to the article is in the comments. Olli Kallioniemi Kimmo Porkka University of Helsinki Valtioneuvoston kanslia

26

3 kommenttia

The State of AI Video Generation. AI video models are moving beyond experimentation. Evaluation now centers on motion realism, temporal consistency, and camera stability. A recent benchmark compares four leading systems under the same high-impact scenario: Kling 3.0 Seedance 2.0 Veo 3.1 Sora 2 When scenes become chaotic — collapsing structures, dynamic movement, rapid camera shifts — differences between models become visible. For agencies and production teams integrating AI into commercial workflows, model selection directly impacts revision cycles, rendering stability, and asset usability. Links in the first comment. #AIVideo #VideoProduction #GenerativeAI #AIVideoGenerator

4

1 kommentti

Swedish AI company Lovable believes so. Founded in 2023, the company has quickly emerged as a rising force in the fast-growing field of “vibe coding” – where users can build apps and websites with the help of AI. 🚀 In just a short time: $100M+ in annual revenue Valued at $1.8B after its first major funding round 2.3M active users, including 180,000 paying customers Unlike OpenAI and Anthropic, which are tied to their own models, Lovable combines several technologies – including GPT-5 and Claude – to give users more flexibility. Recently, they even launched an AI agent that can debug, read files, generate images, and search the web. Founder Anton Osika argues that this puts Lovable in a stronger position than its much larger rivals. Still, critics warn that AI-generated code remains fragile and better suited for prototyping than production. With competition heating up – and players like Figma going public at a $19B valuation – the question is: can Lovable truly carve out its place among the global AI giants? Lovable Anton Osika

2

1 kommentti

Mesh setup… it’s one of those things we all spend way too much time on. Copying settings from one case to another, trying to keep everything consistent, or waiting on a fresh mesh when all you changed was a small parameter. It adds up fast. I’ve been looking at how bramble tackles this, and a few things jumped out: 1. Re-using meshes - For sweeps (speed, yaw, whatever), you don’t need to re-mesh every case. You can clone from a baseline RANS mesh and carry it straight into new runs, even into DES. 👉 https://lnkd.in/e5QyBteX 2. Template mesh settings - Instead of manually repeating refinements across similar cases, you set them once in a template and bramble applies them everywhere. Keeps things consistent and saves a lot of repetitive clicks. 👉 https://lnkd.in/epnr8pdg 3. Geometry morphing - Not strictly meshing but so helpful! For those last-minute geometry tweaks, you can morph the geometry directly rather than bouncing back to CAD. Handy when you’re iterating fast. 👉https://lnkd.in/eTSHvus6 None of this is flashy for the sake of it, it just cuts down on the grunt work and lets you get to the part you actually care about: engineering. If meshing has ever felt like it’s eating half your project time (been there…), you might find the above links interesting, or you can see more of bramble cfd feature's below: https://lnkd.in/e8yesmPZ #ComputationalFluidDynamics #MeshGeneration #AerodynamicDesign #CFD #Simulation

103

2 kommenttia

🎙️IT Talks 266. AI as a tool for programmers Guest: Kim Wessel Bjørneset Language: Norwegian In this episode of IT Talks, Kim, a system developer at Redpill Linpro Oslo, discusses how AI can be used as a tool for developers, including code generation, documentation, automated comments, and configuration files. AI can help improve efficiency, but it’s also important to understand the risks, the purpose of its use, and how it may affect developers’ learning. https://lnkd.in/d4DH36EY #AI #ittalks

9

🚀 Last night Stockholm MLOps hosted the launch of Arkus AI and it turned into one of the most interesting healthcare AI discussions we’ve had so far. A packed room with: clinicians, healthcare founders, AI engineers, researchers, infrastructure builders and operators from across the ecosystem. What stood out was how grounded the conversations were. This was not: ���AI will change healthcare.” This was: How does AI actually work inside real healthcare systems? Some quotes: “Healthcare workflows still operate like 1991.” “The difficult part is not the LLM. It’s the routing, integrations, security, and compliance.” “Patient empowerment will force healthcare transformation.” “Operationalizing compliant AI at scale is the real challenge.” “AI agents are becoming operational healthcare infrastructure.” The strongest signal from the evening: Healthcare AI is moving from pilots and demos into operational systems. The focus is shifting toward: interoperability, governance, workflow integration, trust, runtime evaluation and how AI coexists inside real healthcare ecosystems Huge thanks to the Arkus AI team, all speakers, panelists, and everyone who joined the discussion. Real systems. Real stories. Real people. https://lnkd.in/dx6HDrbx #StockholmMLOps #HealthcareAI #AIAgents #AIInfrastructure #MLOps #AIInProduction #ArkusAI

11

Our paper, "From registers to research: Data quality challenges in Finnish EHR – Enabling AI-driven research through data preparation and standardization", has been published in the Finnish Journal of eHealth and eWelfare. Link to the paper: https://lnkd.in/e6caGQGw In this work, we explore the real-world challenges of working with electronic health record (EHR) data in Finland—highlighting issues such as fragmentation, missingness, and inconsistencies, and how they affect the reliability of AI models, emphasizing the need for careful data preparation and standardization to ensure trustworthy results. Sincere thanks to my co-authors Pekka Siirtola, Outi Laatikainen, Satu Tamminen, Gunjan Chandra, Taneli Mattila, Juha Röning, and Jarmo Reponen for their valuable contributions to this work. #AI #Healthcare #EHR #DataQuality #RealWorldData #PrecisionMedicine #MachineLearning

62

3 kommenttia

Entry-level programming roles are declining, while senior, complex roles are increasingly in demand. Especially those focused on problem-solving, architecture, and design. https://lnkd.in/ecxe6agt ELVA11

7

1 kommentti

Finland is no longer in the “build the digital foundation” phase of healthcare IT. It is in the harder phase: turning a strong national infrastructure into better workflows, cleaner interoperability, stronger diagnostics, smarter analytics, and practical AI at scale. Our new Black Book Research report, State of Digital Healthcare IT: Finland 2026, examines why Finland stands out as one of Europe’s most mature and strategically important digital health markets—and where the next competitive opportunities are emerging. The report looks at: • Kanta and MyKanta’s national impact • County-led modernization and provider-side variation • EHR platform dynamics and procurement priorities • Imaging, diagnostics, interoperability, and analytics • Where AI is becoming operational, not theoretical Finland is not a catch-up market. It is a precision market. Now available from Black Book Research. https://lnkd.in/eWT5JZna #HealthcareIT #DigitalHealth #Finland #EHR #Interoperability #HealthTech #AIinHealthcare #HealthcareInnovation #TietoevryCare #Apotti #Epic #CGI #Mediconsult #BCBMedical #Sectra #Philips #SiemensHealthineers #OracleHealth #Kanta

11

2 kommenttia

Emotion #AI is increasingly shaping discussions around how #digitalhealth tools can better support users by responding to their emotions, moods, and motivation levels. In this article, Rasmus Kyyhkynen from the Turun yliopisto - University of Turku examines how #emotionaware technologies are emerging within fitness and wellness applications and why combining emotional intelligence with #gamification may redefine user engagement in digital health. The article explores: ✅ How emotion-aware systems could tailor feedback in #fitness and #wellnessapps ✅ Why emotions are central to sustained motivation and behaviour change ✅ How gamification elements can adapt to users’ emotional states ✅ What empathetic design could mean for the future of digital health technologies Overall, the piece positions this field as emerging, where technical advancement needs to be balanced with thoughtful evaluation of user experience and impact. ℹ️ Read the full article 👇 #EmotionAI #DigitalHealth #Gamification #AffectiveComputing #HealthTech #UXDesign #Wellbeing #EmergingTech

13

𝗜’𝘃𝗲 𝗮𝗹𝘄𝗮𝘆𝘀 𝗯𝗲𝗹𝗶𝗲𝘃𝗲𝗱 𝘁𝗵𝗮𝘁 𝗼𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲 𝗔𝗜 𝗶𝘀 𝘄𝗵𝗲𝗿𝗲 𝗿𝗲𝗮𝗹 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗵𝗮𝗽𝗽𝗲𝗻𝘀. But what I just saw… genuinely surprised me. A new model from Mistral AI is pushing boundaries in a way that feels straight out of the future: 🎙️ It can clone a human voice from just a few seconds of audio 🧠 It generates natural, emotional speech (not robotic at all) 🌍 It works across multiple languages ⚡ And yes — it’s built with an open approach We’re not just talking about text-to-speech anymore. We’re talking about: → AI narrators that sound human → Personalized voice assistants → Content creation at scale (podcasts, audiobooks, videos) → Real-time dubbing that actually feels real And here’s the bigger shift 👇 For the longest time, the best voice AI was locked behind APIs and expensive platforms. Now? Open models are catching up — fast. This changes everything for: • Developers (more control) • Startups (lower cost) • Creators (more freedom) We’re entering a world where your voice can scale without you being present. That’s powerful. And a little scary too. The real question is no longer “Can AI sound human?” It’s 👉 “What will we build with it?” Curious to hear your thoughts — Would you trust an AI voice that sounds indistinguishable from a real person? #AI #OpenSource #VoiceAI #Innovation #MistralAI #FutureTech

9

3 kommenttia