Nothing is future-proof
”We shape our tools and thereafter our tools shape us.” Marshall McLuhan
Previously at the Volcano Base, I'd been thinking about craft in the context of AI video tools.
Since then, it's been a bit of a work desert. Lots of enthusiastic discussion about projects, but nothing kicking off yet. <cry cry>
Mission Briefing
Nothing is future-proof
If you’ve ever bought a phone, smart thermostat, or robot hoover and comforted yourself with the phrase “this should be future-proof,” congratulations, you’ve lied to yourself just like the rest of us do. When it comes to AI models (LLMs), "future-proof" is the punchline before the pratfall.
Models today, relics tomorrow
The pace of LLM development makes planned obsolescence look like a lazy Sunday. By the time your team gets a model integrated, there's a newer, shinier, more capable AI model already knocking on the door, bragging about its ambivalence to em-dashes.
I've seen companies invest weeks integrating one only to find out it’s been outclassed by something built over a long weekend by a rogue research team fuelled by Monster and questionable sleep hygiene.
What's actually worth future-proofing?
The smart money’s not on the model. It’s on the system. Think model-agnostic design, containerisation, loose coupling, and workflow modularity. If your stack lets you swap models in and out like phone cases, you’re doing it right (using services like OpenRouter to dynamically swap to the most appropriate tool). If you hard-coded GPT-4 into your backend logic like it’s your soulmate, or you’re committed to Microsoft Copilot, we need to talk.
A gentle nudge
Maybe the future isn’t something to “proof” against anyway. Maybe it’s something to prepare for, with flexibility, not false confidence. Future-ready.
Action suggestion: Atomise your workflow to see where AI is being used. If a new model dropped tomorrow, how fast could you switch? If it'd take more than an hour, then start considering a modular approach instead.
Classified Intel
Some interesting stuff I discovered on my adventures.
Yes, you will lose your job to AI
Personal Math
Summary: This piece argues that job displacement from AI isn't hypothetical. It’s already happening, especially for knowledge workers, and that denial is the most dangerous response.
Why it matters: It's a sobering but useful read for anyone still clinging to the idea that "AI can't do my job." Sharing this might spark some honest conversations (and hopefully, contingency planning).
Build your AI identity with Gravatar’s new tool
Ronnie Burt
Summary: Gravatar introduces a new feature that lets users create and manage AI-generated identities tied to their profile, offering a sort of persistent "AI face" across platforms.
Why it matters: As AI agents start representing us more publicly, having a consistent, recognisable identity for both humans and machines might become part of the new professional toolkit.
Story Protocol wants to rewrite IP for the AI era
Summary: Story Protocol is building a decentralised infrastructure to help creators register, licence, and trace ownership of creative works in an AI-dominated future. Copyright is struggling these days, and this is a programmatic alternative.
Why it matters: If you're in tech, media, or anywhere near generative AI, understanding how IP might evolve is essential — this could be the foundation for protecting (or monetising) your next great idea.
If you've got a second, and felt like leaving me a testimonial, that would be super-smashing. It really helps when speaking to people that don't know me, and I could use a boost.
Your data is not shared. Unsubscribe with 1 click.
Member discussion