Hi
It’s the Hùng Kings Festival Monday (Chúc mừng Giỗ Tổ Hùng Vương!) so we are currently enjoying a long weekend in Vietnam. Great chance to have a break before a very busy period with guests and a range of talks, work trips, etc. Looking forward to those but love a chance to chill back too 🤩
You know what doesn’t chill back tho? The world of AI – so let’s get stuck into some headlines from the last week:
The Protocol Revolution: AI’s “HTTP Moment” Arrives | Foundation Technology 🌐
Long, long ago in very early January 2025 – ages past in the world of AI – Sam Altman (OpenAI boss) said that 2025 would likely be the year when AI agents “join the workforce” and make material change to companies and their output. Looks like we already have concrete signs of this happening via a standard developed by OpenAI’s rival, Anthropic: the Model Context Protocol (MCP). Think of MCP as a universal power adaptor – but instead of enabling you to connect with different kinds of funky plugs, it lets you connect your AI with just about any power outlet out there – whether that outlet is Google Drive, WhatsApp, AirTable, Blender, Zapier – the list is long and growing).
Why is this interesting? Well, it’s not too far to say that this is a bit like the hypertext transfer protocol (aka the “http” at the start of websites) that gave us the modern internet. In the early days, the internet was fragmented networks with powerful systems that couldn’t talk to each other. The hypertext transfer protocol gave us a common, open standard for communication – and it might just be that we’re at the http moment for AI. Not least given rivals like OpenAI have moved to adopt MCP and Amazon are moving in similar directions and confidently stating that “soon there will be more AI agents than people browsing the web”. We’ll return to this momentarily but, before we do, a quick sidebar to all my fellow Notion/Second Brain nerds – how good is this? Found it in my travels and had to share 🤓🤩😍
All these and a whole lot more… and yes, that is Blender. Think about the time savings from a learning curve p.o.v. alone – and this is just the beginning 😍
Workforce Transformation: AI Agents Enter Business | Economic Impact 💼
The Wall Street Journal reveals what’s driving billions in tech investment right now: AI agents that can make decisions and take actions independently! From OpenAI to Apple, ServiceNow to Nvidia – the entire tech ecosystem has money riding on systems that function less like tools and more like digital colleagues. Early adopters like Adecco are already seeing game-changing results, filling 100% of open positions (up from a typical 70%) by treating AI as “digital workers that need onboarding” rather than mere automation tools. Meanwhile, ServiceNow and Visa have revolutionised dispute resolution for hundreds of financial institutions with AI agents that can make decisions autonomously. 💼
Combine this with recent World Economic Forum data (i.e., the Future of Jobs Report 2025) and we have confirmation this revolution is already building momentum. A striking 85% of employers are prioritising upskilling their workforce while 73% accelerate automation processes. This isn’t about replacing humans – it’s transformation at its core… as many as 41% of surveyed employers plan staff reductions while 63% focus on augmenting their workforce with new technologies. This is not a “one day” issue – this is a “right now” one – and we need to be asking ourselves what the world of work looks like as soon as Q3 2025… 😲
Share of employers planning to adopt the stated workforce strategies (2025-2030)
Educational Intelligence: Claude Reimagines Learning | Higher Education 🎓
Late to the party but the writer’s favourite AI has joined the education party – Anthropic just unveiled Claude for Education. Rather than building another answer-generator, they’re leveraging Socratic questioning to guide students through their own reasoning processes and HE institutions worldwide are lining up – Northeastern, LSE, and Champlain College are already on board, bringing Claude to 50,000+ students and faculty. Extra points for the Canvas LMS integration as well #Khanmigowho?
Full disclosure, I’m a bigtime Claude fan so this is awesome news for me – I find Claude more chilled and thoughtful vs ChatGPT’s more preppy forced upbeatness (tho 4.5 is a definite improvement). Will be interesting to see how this pans out – might feel a bit like having a knowledgeable friend guide your thinking. Anyway, we’ve applied for access and, with luck, will be able to report back soon. Fingers crossed… and if anyone gets access beforehand, please do let us know how it goes!
Therapeutic AI: Mental Health’s Digital Breakthrough | Healthcare Innovation 🧠
Breaking news in mental health tech: AI just gate-crashed the therapy scene! Dartmouth researchers have unleashed Therabot, an AI therapy tool that’s delivering mind-blowing results. In an eight-week trial, this digital therapist slashed depression symptoms by 51%, crushed anxiety by 31%, and made serious dents in eating disorder risks. Translation? A bot that might actually understand human emotions better than some human therapists. Users were firing off 10 messages daily, forming deep connections that challenge everything we thought we knew about AI-powered support. 💡
But this isn’t a tech free-for-all. Lead researcher Nick Jacobson has serious questions about the current AI therapy market, describing most bots as barely sophisticated chatbots masquerading as mental health support. Therabot stands out by being meticulously crafted on evidence-based practices, not internet chatter. The FDA hasn’t cracked down yet, but Jacobson suggests most AI therapy platforms would crumble under real regulatory scrutiny. The million-dollar question: C=can AI truly bridge the massive mental health care gap, or is this just another Silicon Valley pipe dream? Only time will tell but this could be truly life-changing for so many people… 🧠✨
Visual Intelligence: The New Reality | Creative Tools 🎨
Despite the flood of Studio Ghibli images and Veo 2 vignettes, it appears that tales of Midjourney and Runway’s untimely end at the hands of OpenAI and Google might have been a bit pre-emptive. Midjourney have released v7 and I’m equal parts absolutely blown away while also certain I’m not okay with how real it looks (see hero image at top or check out the discord here). Some interesting features in here including ‘draft mode’ which speaks to the ability to iterate and edit now being built in as standard with a range of AI models – nice enhancement across the board.
Not to be outdone in the attempt to snatch back the limelight, Runway have launched Gen-4 – with a focus on consistent style, characters, locations, and objects (including image to video) this is phenomenal stuff – the ability to take a reference image and then to compose shots and script/direct action in real time is absolutely mindboggling… but then the finished products are another level again. Case in point, the Age of Beyond is a 3m space opera that sets a new standard for what can be done with AI – and it’s looking very high indeed 👀🍿
The AI landscape isn’t just evolving – it’s experiencing a fundamental transformation that touches every aspect of how we work, learn, heal, and create. From Anthropic’s Model Context Protocol creating a universal standard for AI connectivity to groundbreaking applications in mental health treatment, we’re witnessing technologies that increasingly understand context, form relationships, and thoughtfully guide human development. For HE, the question isn’t whether to adapt, but how quickly we can prepare students for a world where AI serves not just as a tool but as a colleague, therapist, and teacher. The most exciting aspect isn’t the technology itself, but how it might enhance our distinctly human capabilities for creativity, empathy, and critical thinking. Are we ready? After all, the future arrived last week! 🤖✨