Open New Learning Lab Resources

Lebensstile: Eine neue Sicht auf Kunden und ihre Bedürfnisse. Die Unterteilung eines Großteils der deutschen Gesellschaft in 18 Lebensstile ermöglicht es, ein gut ausdifferenziertes Bild der eigenen und potenziellen Kunden zu erhalten. Ein gekürzter Artikel aus der Trendstudie und dem Workbook “Lebensstile”.

Explore interactive visualizations of a large scientific research article dataset built in collaboration between Inference.net and LAION

Choose your track and dive in.

👇 The study analyzed AI agents across 50+ occupations, from software engineering to marketing, HR, and design, and compared how they completed human workflows end to end. What they found is both exciting and humbling: • Agents “code everything.” Even in creative or administrative tasks, AI agents defaulted to treating work as a coding problem. Instead of drafting slides or writing strategies, they generated and ran code to produce results, automating processes that humans usually approach through reasoning and iteration. • They’re faster and cheaper, but not better. Agents completed tasks 4 – 8× faster and at a fraction of the cost, yet their outputs showed lower quality, weak tool use, and frequent factual errors or hallucinations. • Human–AI teaming consistently outperformed solo AI.🔥 When humans guided or reviewed the agent’s process, acting more like a “manager” or “co-pilot”, the results improved dramatically. 🧠 My take: The race toward “fully autonomous AI” is missing the real opportunity, co-intelligence. Right now, the biggest ROI in enterprises isn’t from replacing humans. It’s from augmenting them. ✅ Use AI to translate intent into action, not replace decision-making. ✅ Build copilots before colleagues, co-workers who understand your workflow, not just your prompt. ✅ Redesign processes for hybrid intelligence, where AI handles execution and humans handle ambiguity. The future of work isn’t humans or AI. (for the next 5 years IMO) It’s humans with AI, working in a shared cognitive space where each amplifies the other’s strengths. Because autonomy without alignment isn’t intelligence, it’s chaos. Autonomous AI isn’t replacing human work, it’s redistributing it. Humans shifted from doing to directing, while agents handled repetitive, programmable layers. Maybe we are just too fast to shift from "uncool" Copilot to sth more exciting called "Fully Autonomous AI", WDYT? | 36 comments on LinkedIn

🚨 Just dropped! McKinsey report on AI in 2025: the hype is loud, the impact is.... All the CEO must read this: almost everyone is “using AI,” but only a small slice is wiring it deep enough to move the needle. 88% of companies use AI somewhere, yet ~⅔ are still stuck in experiments/pilots, not scale. Agents are real but early: 62% are experimenting; only 23% are scaling in at least one function (and typically just one or two). Only 39% report any impact from AI at the enterprise level. The rest have scattered wins, not system change. High performers (≈6%) think bigger: they aim for transformation, not just cost cuts, and are ~3× more likely to redesign workflows around AI. Leadership matters: where the CEO and senior team own AI, adoption scales and budgets follow (many leaders spend 20% of digital on AI). Value shows up fastest in software eng, IT, mfg (cost ↓) and in marketing/sales, strategy/finance, product (revenue ↑). Risk is real and showing up: inaccuracy and explainability issues top the list, mature orgs pair ambition with stronger guardrails and human-in-the-loop. My take: Most firms bought tools; the few winners rebuilt work. Agent pilots are cool, but without workflow redesign, data plumbing, and clear governance, you’re funding demos, not outcomes. The org that rewires will beat the org that “rolls out.” Leaders should set the bar higher than “efficiency.” Tie AI to growth, new offerings, and customer experience, then go after costs. Redesign 3–5 critical workflows end-to-end (not feature by feature). Ship, measure, harden, repeat. Put ownership at the top. If the CEO isn’t accountable for AI governance and ROI, it will stall. Invest in the platform: data products, evaluation, CI/CD for models/agents, human-in-the-loop checkpoints, risk controls. Skill the workforce for agents: task decomposition, prompt/context ops, verification, and change management, at scale. AI ROI doesn’t come from the model. It comes from the company willing to change its operating system. wdyt? | 76 comments on LinkedIn

Archive ‘How to AI’ (most recent to oldest) Delve, and the many words to ban on ChatGPT. (soon) From Youtube to your own AI. (the last one) Your ChatGPT prompt is too long. Remove em-dashes (and more). How to stop receiving the same ChatGPT answer. The new ChatGPT Atlas is live. Is it any good...

Vor einigen Jahren entwickelte ich zusammen mit Niclas Schaper und Andreas Seifert nicht nur das Oktagon-Modell, sondern auch eine dazu passende Skala zur Messung des informellen Lernens am Arbeitsplatz. Später folgte, in Zusammenarbeit mit Prof. Dr. Michael Knappstein, auch noch die Publikation einer 8-Item-Kurzversion. Die Skalen lagen zwar bereits auf Deutsch vor – allerdings nur schwer zugänglich im Anhang der englischsprachigen Publikationen, und Evaluationen sowie Handhabungshinweise waren ausschließlich in englischer „Wissenschaftssprache“ verfasst.   💡 Aus der HR/PE-Praxis kam daher immer wieder die Bitte: Gibt es ein deutschsprachiges, anwendungsorientiertes Manual? Diesem Wunsch sind wir nun nachgekommen. Die „Zusammenstellung sozialwissenschaftlicher Items und Skalen“ (#ZIS) des GESIS - Leibniz-Institut für Sozialwissenschaften bietet genau die Plattform, um diesen Transfer transparent und praxistauglich umzusetzen. In der ZIS-Datenbank stehen daher ab sofort bereit: 📝  Zwei Versionen der Skala für informelles Lernen am Arbeitsplatz: Kurz (8 Items) für schnelle Erhebungen & Monitoring, Lang (24 Items, 8 Facetten) für differenzierte Analysen. 📝 Praxisnah dokumentiert, alles auf Deutsch: klare Anwendung, Auswertung und Interpretation. 📝 Offen zugänglich (kostenfrei) über ZIS, inklusive Materialien. 📝 Transparenzzertifikat des Diagnostik- und Testkuratoriums (DTK) von BDP und DGPs: bestätigt die vollständige, standardkonforme Dokumentation – als verlässliche Grundlage für informierte Anwendung und Bewertung. Direkt zu den Inhalten: 🔗 ZIS – Langskala: https://lnkd.in/eiMUWhBz 🔗 ZIS – Kurzskala: https://lnkd.in/e_rVMdqs 📄 Originalpublikation der Langversion auf Englisch (2019): https://lnkd.in/gbGpxKY 📄 Originalpublikation der Kurzversion auf Englisch (2023): https://lnkd.in/e5WsWvcK   Mein Dank geht an #GESIS, insbesondere an den betreuenden Reviewer Julian Urban, für die Möglichkeit, die Skalen offen und nachhaltig zu publizieren – das hilft, Brücken zwischen Forschung und betrieblicher Praxis zu schlagen. 🙌 Rückmeldungen aus der Anwendung sind sehr gern willkommen, z.B. zu Einsatzfeldern, Betrachtungen zu Aufwand & Ertrag und Einbindung in organisationale Lernevaluationen. Viel Erfolg beim Nutzen der Skalen! 🙂 #InformellesLernen #NewLearning #HR #Personalentwicklung #Praxistransfer| 19 Kommentare auf LinkedIn

Last time I asked what would happen if L&D went away. This week, we’re asking how it can stay relevant as everything around it changes. We created the L&D Strategy Framework to account for all the things L&D should consider in its strategy. We’re in the process of figuring out what it looks like with an AI lens. What areas do ya'll think are changing, and how? P.S. - We're discussing this in this week’s L&D Huddle on Wed 1PM EDT. RedThread members can RSVP on the platform. I also have a few guest spots. DM me if you want in. P.P.S. - This L&D Strategy Framework is a part of a larger infographic. You can grab the full infographic through our Starter tier (read: free). Link in the first comment. | 33 comments on LinkedIn

It’s finally happening. 🎉 We’re ditching LMS and SCORM and building learning resources right where our learners already spend their time: in the CRM and sales enablement tools they use every day. It’s literally learning in the flow of work. With AI, we’ve created agents that deliver learning exactly when it’s needed, put together role-specific learning paths, and (soon) will really push personalized lessons based on our team’s performance and platform usage. When AI tools first arrived, the fear was real—and honestly, it still is. As an #LXD, I see every day how AI is “taking over” parts of my job: - I no longer have to set up or record audio and video; AI does that for a fraction of the cost. - I don’t need to code interactions anymore; AI handles that with clear instructions. - I don’t have to manually gather and review progress reports or analyze and present data; AI does it faster and more accurately than I could. Since AI is taking care of those routine and time-consuming tasks, I can focus on designing learning experiences. I actually have time to sit down with my team, brainstorm strategy, and think long term. We get to be more creative and experiment with new ideas (whether they work or not) without burning through tons of resources. I never thought I’d see the day, but yep, LMS and SCORM are dead (or about to be). AI image created in Canva. | 134 comments on LinkedIn

The Headline: 👉 Global learning is at a crossroads: learner outcomes have dropped sharply worldwide, and UNESCO projects a shortage of 44 million teachers by 2030. 👉 AI is positioned as *the* tool to save us from an impending education crisis BUT... 👉 The real "secret weapon" for improving education isn't the tech: it's the learning science we build into it. According to Google, the four biggest opportunities offered by AI in education are: 🔥 Learning Science at Scale – Embed evidence-based methods (retrieval practice, spaced repetition, active feedback) directly into everyday tools. 🔥 Making Anything Learnable – Adjust explanations, examples and complexity to meet each learner where they are. 🔥 Universal Access – Break down language, literacy and disability barriers through AI-powered translation and transformation. 🔥 Empowering Educators – Free up teacher time through AI-assisted lesson planning, resource creation and differentiation. Overall, Google's latest white paper signals an evolving ed-tech culture which centres on a more substantive partnership between ed & tech: 👉 Co-Creation: Google commits to investing in evidence-based approaches to learning design and development and to rigorous evaluation, pilot studies and educator-led research to test and demo impact. 👉 Collaborative Development: Google commits to working with schools, NGOs, researchers and learning scientists to co-design tools for learning. You can read the white paper in full using the link in comments. Happy innovating! Phil 👋 | 26 comments on LinkedIn

In this 2025 edition of the annual McKinsey Global Survey on AI, we look at the current trends that are driving real value from artificial intelligence.

Resilience Science Must-Knows: Nine Things every Decision-Maker Should Know about Resilience! I. Resilience Resilience has become a central consideration across practice, policy, and business. It is increasingly integrated into public health strategies, private-sector risk management, corporate planning, development, and financial investment II. Knowledge Decision-makers across regions and sectors urgently need clear, science-based, and actionable knowledge to maintain the resilience of people and the planet and to ensure societies have the capacity to cope, adapt, and transform in order to thrive amid uncertainty. III. Nine Must-Knows are: 1. Navigate accelerating risk: Resilience offers pathways toward more just and sustainable futures for people and the planet. 2. Cope, adapt, and transform: Resilience is more than just bouncing back from shocks. 3. Invest today – benefit tomorrow: Resilience protects and strengthens the foundations that support long-term human well-being and prosperity. 4. Cultivate continuous learning and innovation: Resilience is a cycle that demands continuous experimentation, learning, and innovation. 5. Foster diversity in all its forms: Diversity is both a source of persistence, providing multiple options, and a source of adaptation, and transformation. 6. Nurture relationships: Resilience grows through relationships and these connections strengthen the flow of resources, knowledge, trust, and care. 7. Govern and negotiate trade-offs: Addressing trade-offs is vital to avoid unintended harms, prevent conflict, and build just, lasting resilience. 8. Empower agency: Supporting and developing agency means enabling people and institutions to take intentional and grounded action. 9. Address power imbalances: Failing to address social inequalities, power imbalances, and historical injustices risk reinforcing the very systems that cause vulnerability. IV. From science to impact This report is not an endpoint. In the next phase of the efforts—the Resilience Road to Action—the initiative will work closely with decision- makers across a range of sectors, including food and agriculture, urban development, health, and finance, to translate the Resilience Science Must-Knows into actionable guidelines. Make sure to check out the important report by Stockholm Resilience Centre , Global Resilience Partnership, Future Earth here: https://lnkd.in/dmaARyPU ______ Stay Ahead of Transformative Innovation Follow The Futuring Alliance for regular insights, foresight, and practical tools to help your organization thrive in times of change. We support leaders across industries in turning future-focused ideas into real-world impact—through collaboration, innovation, and bold action. Let’s shape what’s next—together. #reslience #sustaunability #innovation #foresight #system #systemschange #strategy #venturing #impact

Last week, a group of L&D peers gathered to discuss budgeting. The initiative started from a Slack thread and turned into a space where we discussed all our questions about L&D budgets and the budgeting process. There were SO many brilliant questions asked, and I wanted to share my favs: 👉 What percentage of your L&D budget is directly linked to measurable business goals or KPIs? 👉 When the budget gets tight, how do you decide which initiatives to keep and which to pause or cut? 👉 How do you plan or forecast your L&D budget for the year when future needs aren’t yet fully clear? 👉 How do you measure the impact or return on your L&D investments, beyond just participation rates or satisfaction scores? 👉 How do you factor in external shifts, like economic changes, new technologies, or workforce trends, when planning your L&D budget? 👉 Do you ever include the cost of employees’ time spent learning in your budgeting decisions? Could that actually be your biggest investment? 👉 How can L&D teams move from a “repeat last year’s budget” mindset to a more intentional, impact-driven budgeting approach? What other questions would you add to the list? 💡 Big thanks to peers such as Adam House, Chris McLaughlin FLPI, Jen Collins, Lisa-Noreen Kröger, Mair Horscroft Assoc CIPD, Janelle M., Tereza Kenova, Lauren Vecchio, Marine Petitpas, among others, for joining the conversation with so much openness ☂️💜 #learninganddevelopment #learningbudgeting

What happens when learners meet AI? Think of skill development as a road from beginner to expert. You normally start with basic practice, work through tough problems, reflect on what's working, and eventually reach the point where you can handle anything that comes up. Now AI has entered this picture. Depending on how we use it, we end up on completely different roads. Use AI too early and you risk never-skilling. You skip the fundamentals and never develop real capability. Hand over too much and you risk de-skilling. Abilities you once had start to fade. Copy AI outputs without thinking and you risk mis-skilling. You learn the wrong lessons and build on faulty foundations. But there's another path. Use AI while staying critical. Question its outputs. Think through the logic. Verify the answers. This is AI-enhanced adaptive practice. AI becomes a sparring partner that helps you learn faster without replacing your own reasoning. The difference comes down to one thing: who's in control. The people who'll succeed with AI aren't avoiding it or surrendering to it completely. They're the ones who keep thinking while using AI to compress learning cycles and test ideas faster. AI shouldn't replace your thinking. It should make your thinking better. The question isn't whether to use AI when learning. It's whether you're driving or just sitting in the passenger seat. How are you seeing this play out in your work? ✍ Raja-Elie Abdulnour, Brian Gin, Christy Boscardin. Educational Strategies for Clinical Supervision of Artificial Intelligence Use. N Engl J Med. 2025;393(8):786-797. DOI: 10.1056/NEJMra2503232 | 10 comments on LinkedIn

𝗘𝗺𝗯𝗿𝗮𝗰𝗶𝗻𝗴 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗮 𝗗𝗶𝘀𝗿𝘂𝗽𝘁𝗲𝗱 𝗪𝗼𝗿𝗹𝗱 𝗪𝗮𝗿𝘂𝗺 𝗱𝗲𝗿 𝗠𝗲𝗻𝘀𝗰𝗵𝗻 𝗱𝗲𝗿 𝘄𝗶𝗰𝗵𝘁𝗶𝗴𝘀𝘁𝗲 𝗙𝗮𝗸𝘁𝗼𝗿 𝗶𝗻 𝗱𝗲𝗿 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗶𝘀𝘁 In einer Welt voller Unsicherheit, technologischem Wandel und geopolitischer Spannungen ist Transformation längst kein Projekt mehr – sie ist Dauerzustand. Die aktuelle Studie von Arthur D. Little – „Embracing Transformation in a Disrupted World“ (2025) zeigt eindrucksvoll, wie tiefgreifend der Wandel bereits in Unternehmen verankert ist – und wo er scheitert. ➡️ 65 % der Unternehmen befinden sich aktuell in umfassenden Transformationsprozessen. ➡️ 95 % der Führungskräfte glauben an ihren Erfolg. ➡️ Doch nur 7 % schaffen es, eine wirklich kontinuierliche Transformation zu leben. ➡️ Der größte Stolperstein? Nicht Technologie – sondern Menschen. Die Studie belegt: Ohne echte Einbindung der Mitarbeitenden bleiben Strategien, Strukturen und Systeme leere Hüllen. Denn: Transformation gelingt nicht um Menschen herum – sie gelingt nur mit ihnen. 𝗪𝗮𝘀 𝗯𝗲𝗱𝗲𝘂𝘁𝗲𝘁 𝗱𝗮𝘀 𝗳ü𝗿 𝗛𝗥 𝘂𝗻𝗱 𝗟𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 ➡️ Upskilling & Re-Skilling sind kein „Nice-to-have“ mehr, sondern Voraussetzung für Zukunftsfähigkeit. ➡️ Kulturarbeit muss Teil der Transformationsarchitektur sein – nicht ein Begleitprogramm. ➡️ Führungskräfte sind die entscheidenden Übersetzer zwischen Strategie und Emotion. Sie schaffen Sinn, Vertrauen und Energie. ➡️ Iterative Transformation statt Großprojektdenken: Wandel wird dann nachhaltig, wenn Organisationen lernen, sich selbst permanent weiterzuentwickeln. 📊 Besonders spannend: Nur 5 % der Unternehmen bewerten ihre Lernkultur als „sehr effektiv“. Das zeigt, wie groß der Handlungsbedarf ist – gerade im HR. Hier entscheidet sich, ob Transformation getragen oder gebremst wird. 🎯 𝗠𝗲𝗶𝗻 𝗙𝗮𝘇𝗶𝘁 Technologie mag der Katalysator sein – aber Menschen sind der Motor jeder erfolgreichen Transformation. Wer ihre Potenziale entfesselt, gestaltet nicht nur Wandel, sondern Zukunft. Quelle: Arthur D. Little (2025): “Embracing Transformation in a Disrupted World” – Autoren: Francesco Marsella, Wilhelm Lerner, Ben van der Schaaf, Marten Zieris, Alexander Buirski, Francesco Cotrone, Alexis Ost Duchateau.

Generate use case ideas and identify the most promising ones to implement.

Ich habe mir die Megatrendstudie zu Wissenskultur angeschaut und mit unseren Thesen zu "New Learning" und warum es eine Transformation benötigt, verglichen. Kurz: Die Schnittmenge ist sehr hoch.

Learn and earn with Google Skills, a platform that provides free training and certifications for Google Cloud partners and beginners. Explore now.

But people have reacted to genrative AI so differently. Some have embraced it with gusto. Many have shrunk away from it. Thre vast majority of AI experimentation and usage still happens outside of work (ChatGPT has 800m weekly mostly-consumer users now). Most firms don’t have a very good idea of where the individuals and teams that make up their workforce are. Well, a 2x2 matrix almost always helps - so simple, so illuminating. It’s my favourite mental model. In this situation, adoption and capability are two pertinent axes to think about this. It gives a sense of where there’s overconfidence, underconfidence and appropriate confidence. And what actions you might take for populations in each of the quadrants. This enables you to better serve your people, and be better served by them. If you’re interested in a 30-question survey which generates the data behind each axis and forms part of and builds on my AI in the Wild use case research, send me a message. ♻️Please REPOST if people you’re connected to may like to be updated on how AI is being used, out in the Wild. #aiinthewild

In 2014 Michael Staton detailed how the university experience is a bundle of many things rolled up together (see image below), and suggested that many components could be disaggregated, or “unbundled” and provided in alternative ways. He drew on this framework in his “The Degree is Doomed” piece in the Harvard Business Review in 2014. https://lnkd.in/gYdkhCb In 2025, Shannon McKeen, writing in Forbes, considers where we are now https://lnkd.in/gYir-XDe: “Most college students now use AI tools for academic work, yet employers consistently report that new graduates lack the critical thinking and decision-making skills needed in an AI-augmented workplace. This disconnect signals the beginning of higher education's great unbundling. For decades, universities have operated on a bundled model: combining information delivery, skill development, credentialing, and social networking into a premium package. AI is now attacking the most profitable part of that bundle—information transfer—while employers increasingly value what machines cannot replicate: human judgment under uncertainty. Higher education represents a massive market built largely on controlling access to specialized knowledge. Students pay premium prices for information that AI now delivers instantly and for free. A business student can ask ChatGPT to explain supply chain optimization or generate market analysis in seconds. The traditional lecture-and-test model faces its Blockbuster moment. This is classic disruption theory in action. The incumbent model optimized for information scarcity while a new technology makes that core offering abundant. Universities that continue competing with AI on content delivery are fighting the wrong battle. The real value is migrating from information transfer to judgment development, from transactional learning to transformational learning. In an AI saturated world, premium skills are distinctly human: verification of sources, contextual decision-making, ethical reasoning under ambiguity, and accountability for real-world outcomes. This shift mirrors what happened to other information-based industries. When Google made basic research free, management consulting pivoted to implementation and change management. When smartphones made maps ubiquitous, GPS companies focused on real-time optimization and personalized routing. Higher education must make the same transition… The great unbundling of higher education is underway. Information delivery is becoming commoditized while judgment development becomes premium. Institutions that recognize this shift early will capture disproportionate value in the new market structure.” H/t Sinclair Davidson

To address rapidly changing business conditions, companies are cultivating a workforce based on skills. Initiatives that incorporate four success factors are most likely to create tangible value.

I learned AI Agents for absolutely free, you can do it too! AND... the best part is I got to learn from industry experts DeepLearning.AI has done a great job in making these courses. 1. Event-Driven Agentic Document Workflows with LlamaIndex - https://lnkd.in/d7vJEH4H 2. Long-Term Agentic Memory with LangGraph (LangChain) - https://lnkd.in/dKJ-B3ks   3. Build Apps with Windsurf's AI Coding Agents (Codeium) - https://lnkd.in/dTqjjt4Q   4. Building AI Applications with Haystack (by deepset) - https://lnkd.in/d7WnTvTr   5. Improving Accuracy of LLM Applications (Lamini) - https://lnkd.in/dcJvY6kg 6. Evaluating AI Agents (Arize AI) - https://lnkd.in/dvTNKSaq _______________ ♻️ Repost it to help others.  _______________ If you like this, and want more AI resources, images, tutorials, and tools, join Superhuman, my daily AI newsletter with 1M+ subs now: https://lnkd.in/dXQ9-B9A | 55 comments on LinkedIn

AI isn’t just transforming Learning & Development. It’s revealing it. For years, we’ve talked about being strategic partners - about impact, performance, and business alignment - but much of L&D has still operated as a content-production function. We’ve equated “learning” with “stuff we make”. Now AI has arrived, and it’s showing us what’s really been going on. - If your value comes from creating courses and content, AI will replace you. - If your value comes from solving real problems for the business, AI will amplify you. That’s the pivot point we’re in. The new report, The Race for Impact written by Egle Vinauskaite and Donald H Taylor, captures this moment perfectly. Within it, they describe the “Implementation Inflexion” - the shift from experimenting with AI to actually using it - and revealing what L&D teams are doing as they lead the way. The “Transformation Triangle” lays out three models that go beyond content: Skills Authority - owning data and insight around workforce capability Enablement Partner - orchestrating systems that help others solve problems Adaptation Engine - continuously learning with the business to stay relevant Each one moves L&D closer to the business and further from being an internal production house. This isn’t about tech. It’s about identity. And the teams that figure that out now will define what L&D means in the age of AI. Hear more about this from Egle in the latest episode of The Learning & Development Podcast and what this all means in practice. A link to this episode is in the comments.

TeachLM: six key findings from a study of the latest AI model fine-tuned for teaching & learning. LLMs are great assistants but ineffective instructional designers and teachers. This week, researchers at Polygence + Stanford University published a paper on a new model — TeachLM — which was built to address exactly this gap. In my latest blog post, I share the key findings from the study, including observations on what it tells us about AI’s instructional design skills. Here’s the TLDR: 🔥 TeachLM outperformed generic LLMs on six key education metrics, including improved question quality & increased personalisation 🔥TeachLM also outperformed “Educational LLMs” - e.g. Anthropic’s Learning Mode, OpenAI’s Study Mode and Google’s Guided Learning - which fail to deliver the productive struggle, open exploration and specialised dialogue required for substantive learning 🔥TeachLM flourished at developing some teaching skills (e.g. being succinct in its comms) but struggled with others (e.g. asking enough of the right sorts of probing questions) 🔥 Training TeachLM on real educational interactions rather than relying on prompts or synthetic data lead to improved model performance 🔥TeachLM was trained primarily for delivery, leaving significant gaps in its ability to “design the right experience”, e.g. by failing to define learners’ start points and goal 🔥Overall, human educators still outperform all LLMs, including TeachLM, on both learning design and delivery Learn more & access the full paper in my latest blog post (link in comments). Phil 👋

Ermitteln Sie den Reifegrad Ihrer KI-Transformation - Kostenlose Selbsteinschätzung basierend auf dem Learning Ecosystem Framework

Ein Impuls für individuelle Lernpfade durch Künstliche Intelligenz im Bereich L&D „Kunden, die diese Bettwäsche kauften, kauften auch …“ Dieses Prinzip kennen Sie sicher von E-Commerce-Plattformen. Ihr Verhalten wird erfasst, verglichen und in Empfehlungen übersetzt. Das Ziel: Sie kaufen bestenfalls mehr als diesen einen Artikel. Übertragen auf Lernplattformen heißt das: Klicks, Quiz-Ergebnisse und Suchanfragen lassen […]

🚨 OpenAI 𝗷𝘂𝘀𝘁 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗱 𝘁𝗵𝗲𝗶𝗿 𝗼𝘄𝗻 𝗔𝗴𝗲𝗻𝘁 𝗕𝘂𝗶𝗹𝗱𝗲𝗿. [And no — it didn’t kill 99% of startups overnight.] It’s called AgentKit — a no-code, full-stack platform to build, deploy, and optimize AI agents. The UI looks surprisingly clean, but let’s be clear: this doesn’t instantly replace Zapier, Make, n8n, or Lindy. AgentKit is impressive, yes — but it’s still early, still developer-focused, and far from being a plug-and-play automation killer. 𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗮𝘁 𝗔𝗴𝗲𝗻𝘁𝗞𝗶𝘁 𝗶𝗻𝗰𝗹𝘂𝗱𝗲𝘀: ⬇️ 1. Agent Builder – a visual interface to design and connect multiple AI agents → You can drag and drop steps, test them instantly, and track versions → Comes with built-in safety checks (guardrails) → It’s in beta — I haven’t tested it yet, but the interface looks quite polished 2. Connector Registry – a control center for all your data connections → Possible to manage integrations with MCP → Adds content and tools to keep it organized, secure, and compliant for enterprise use 3. ChatKit – provides an interface to add chat to your product → Turns agents into a chat interface that looks native → Handles threads, live responses, and context automatically 4. Evals 2.0 – a system to test and improve your agents → Lets you run evaluations using datasets and automated grading → According to OpenAI companies which used it saw up to 30% higher accuracy after using it None of the announced capabilities are truly new, and I doubt that building agents with OpenAI will offer a better experience than platforms like n8n or Zapier. The output still generates code, and the whole setup clearly targets developers (for now) — which explains why it was introduced at DevDay rather than rolled out to the broader user base. And for enterprise-ready AI agents, you still need solid frameworks like LangChain or CrewAI, not another drag-and-drop automation layer. AgentKit is a strong step, but there’s still a way to go before it becomes a production-grade enterprise solution and kills "99% of all other tools". 𝗣.𝗦. 𝗜 𝗿𝗲𝗰𝗲𝗻𝘁𝗹𝘆 𝗹𝗮𝘂𝗻𝗰𝗵𝗲𝗱 𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿  — 𝘄𝗵𝗲𝗿𝗲 𝗜 𝘀𝗵𝗮𝗿𝗲 𝘁𝗵𝗲 𝗯𝗲𝘀𝘁 𝘄𝗲𝗲𝗸𝗹𝘆 𝗱𝗿𝗼𝗽𝘀 𝗼𝗻 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀, 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀, 𝗮𝗻𝗱 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗮𝘆 𝗮𝗵𝗲𝗮𝗱 𝘄𝗵𝗶𝗹𝗲 𝗼𝘁𝗵𝗲𝗿𝘀 𝘄𝗮𝘁𝗰𝗵 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝘀𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀. 𝗜𝘁’𝘀 𝗳𝗿𝗲𝗲 — 𝗮𝗻𝗱 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗿𝗲𝗮𝗱 𝗯𝘆 𝟮𝟬,𝟬𝟬𝟬+ 𝗽𝗲𝗼𝗽𝗹𝗲. https://lnkd.in/dbf74Y9E

PS: once it's live, I'll make a full guide on how-to-ai.guide. It's my newsletter, read by 132,000 people. Here is all of the (trusted) information I gathered: — The no-code AI era starts now — ✓ Drag-and-drop visual canvas for building agents. ✓ Templates for customer support, data, etc. ✓ Native OpenAI model access, including GPT-5. ✓ Full integration with external tools and services. But there was always a wall: coding. Now, anyone can build advanced AI agents. No code. No friction. Here’s how it (seems to) work: You want to automate customer support: 1. Pick a template for a support bot. But you need it to pull info from your database. 2. Drag in an MCP connector. Link your data. You want human approval for refunds. 3. Add a user approval step. Set the rules. You want to check documents for fraud. 4. Drop in a file search and comparison node. Test it. Preview it. Deploy it. All in one place. OpenAI is more than just an API company. It is building the backbone for the no-code AI economy. Now, anyone can create agents that work across systems, talk to users, and make decisions. The age of visual AI automation is here.

It frontloads skills like learning about learning and adaptability, which seem very hard to arrive at with Bloom’s taxonomy. In Fink’s model, there are 6 dimensions (which are interconnected). 1️⃣ Learning about learning 2️⃣ Foundational knowledge 3️⃣ Application 4️⃣ Integration 5️⃣ Human Dimension 6️⃣ Caring My personal opinion is that Fink’s model is going to be much more useful than Bloom’s, when it comes to understanding how AI is changing learning. I made this case when taking to Tina Austin and Michelle Kassorla, Ph.D., when we talked about Bloom’s taxonomy in The Age of AI. (More on that soon!) ——— Image: a screenshot from Fink’s book “Creating Significant Learning Experiences” (2003). | 147 comments on LinkedIn

How does AI change the role of managers? Research shows a new breed of leader is now needed, the Supermanager.