The recent NASA images genuinely blew me away. I know humans have made the trip a few times now, but it still fascinates me and literally blows my mind.
The thought that also surfaces when this topic comes up is when my primary school English teacher told me to ´Get your head out of space and back into the books,” when I told her I wanted to be an astronaut. My head has always been in the clouds I guess. It turns out it fits pretty well over in the Google Cloud, where I have been learning some new skills recently.
All this NASA news and the incoming EU AI Act got me thinking about the same metaphor. We often try to fly to the moon before exploring the world right around us. The hidden gems are always closer than they look.
The same is true for technology, especially now, with a new AI-powered something launching literally every hour. If you’ve been in the tech space this past year, it felt like the wild west. A lot of uncertainty. But that feels like it’s finally changing, and I know many nonprofits will be sleeping much better for it. We now have real-world case studies and frameworks for how to implement AI tools in ways that actually serve your mission.
Claude for Nonprofits is here… and it’s genuinely impressive
I’ve tried and tested most of the major AI tools and found value in all of them, honestly. But Claude has been consistently strong: reliable output quality, a clear ethical stance, and real depth in research and writing.
I’ve been using Claude for writing, research, and building and testing apps for over a year now, back before any of us knew how big this would get. I started when GitHub was known more as a forum than a platform, and terminals lived on Windows desktops. I’m currently building an app almost entirely inside Claude!
Claude for Nonprofits brings that same capability to mission-driven organisations at accessible pricing. For nonprofits juggling limited staff, limited budget, and unlimited complexity, that matters.
Big innovations from tiny little models
Another shift making serious waves is coming from much smaller models like local LLMs that run on your own machine. Large language models are extraordinary, but local models are where something different happens: models that learn how you work, that you can tune and improve, that don’t send your data anywhere.
With the EU AI Act coming into force, this is also a much safer and more ethical way to manage the sensitive data nonprofits handle every day.
Using local LLMs isn’t just an ethical choice. It’s a modular one. Instead of one general-purpose model doing everything, you can run separate purpose-built models for specific functions: one for marketing and comms, one for grant writing, one for donor reporting. Pair those with lightweight automations, and you have a genuinely impressive workflow without getting overwhelmed by the hype around autonomous agents or full AI transformation programmes.
Get the A to B sorted before you book the spaceship.
Space travel, clouds, and going local
Vibe coding isn’t for everyone, but it’s a remarkably accessible way for small nonprofits to build real tools. I’ve built several using Replit and Lovable, even on free tiers, you can produce fully functional apps. They may not be beautifully branded out of the gate, but you’ll uncover capabilities and workflows you didn’t know you could access at such low or zero cost. Internal comms is one use case I haven’t seen many talk about but would work really well and possibly save some budget.
I won’t leave it so long next time. If you’re a nonprofit wanting to learn more about the tools, terminals, or just have a chat about your strategy. Get in touch!
