AI Data Centres Set Sail

This morning I was intrigued by a story I came across about Panthalassa – an Oregon-based startup that’s taking an entirely new approach to powering AI. Backed by Peter Thiel, they’re raising £140M to build floating compute structures that ride the ocean waves for energy. Yes, you read that right – AI data centres are headed for the sea! You can read more about it here.

For digital product designers, this development highlights how resource and environmental challenges are driving innovation in infrastructure. Even if you’re not building data centres, it’s a reminder that our design solutions must evolve to support increasingly complex technologies. It’s a fascinating interplay between physical engineering and digital performance.

Rethinking API Metrics for Better UX

Over on You.com, there’s a detailed guide that challenges the conventional focus on API latency. Most teams, it seems, tend to gloss over true performance by only noting demo benchmarks. The guide explains why raw metrics like p50 latency don’t tell the whole story and introduces a “time-to-useful-result” framework. You can check out the guide here.

For UX designers and developers alike, understanding these hidden cost drivers can mean delivering smoother, more reliable digital products. It’s a gentle reminder to dig deeper than surface-level metrics to ensure our user interfaces run flawlessly under real-world conditions.

Self-Building AI and the Future of UI

Anthropic’s co-founder recently forecast a future where AI systems will be capable of training their own successors – a shift that could revolutionise how quickly and efficiently new features are developed. The idea that AI might soon handle a significant portion of its own R&D is both exhilarating and a tad unnerving. Read about these insights from his latest post here.

For those of us in digital product design, this evolution has clear implications. Imagine interfaces that adapt and evolve almost in real time, optimised by systems that learn from their own use patterns. It’s a glimpse into an era where design and technology might merge even more seamlessly, challenging us to think creatively about integration and user experience.

Local AI on Your Mobile

Ever fancied replacing Siri with a customised AI experience? A handy guide on setting up a local AI model on your iPhone shows you how. With solutions like Google’s open-source Gemma model, you can download an app, bind it to your action button, and enjoy an AI that works entirely offline. More details are available here.

This is a brilliant reminder that as we tailor digital experiences, privacy and performance can go hand in hand. For UX designers, it’s both a creative challenge and an opportunity to literally rethink how we interact with our devices in an increasingly AI-powered world.