Hi

Even though we should be used to it by now, it is still difficult to comprehend how quickly the pace of change is accelerating. As I look back over the year 2025, I am...well, frankly... staggered by some of the things that are now possible, that even a few years ago, we wouldn't have believed would happen in our lifetime.

Two examples have really stuck with me.

First, glasses you can wear that effectively subtitle the world. Real-time captions, right in your field of view, so you can “read” a conversation as it happens. And the same set-up can translate live, across well over a hundred languages and dialects. That’s not science fiction any more. It’s a product you can purchase. 

For deaf and hard-of-hearing people, that kind of live captioning could make everyday chats, meetings and public announcements far more accessible—often without waiting for an interpreter (though it won’t replace BSL where nuance and context really matter).

Second, the frankly bonkers idea of putting serious computing infrastructure in space. Not just “satellites exist”, but a deliberate attempt to design scalable AI compute in orbit, powered by solar energy and connected by optical links. Google has published this as a research “moonshot” (their words), under Project Suncatcher. 

If you’d described either of those to me in 2020, I’d have assumed you’d been reading too much sci-fi.

And here are four more “how is this real?” developments that I’d put in the same category.

Your normal mobile phone connecting directly to a satellite
Not a specialist handset. Not an emergency beacon. Your everyday phone, using satellite connectivity when there’s no mobile coverage. Virgin Media O2 has announced a UK-first partnership with Starlink “Direct to Cell”, including a planned new service called O2 Satellite. 

That has big implications for rural businesses, lone workers, resilience planning, and basic safety. “No signal” stops being a hard constraint and becomes more like an inconvenience.

Robotaxis coming to London (properly, not as a gimmick)
London has now moved into the “serious trials” conversation. Waymo has publicly said it intends to offer rides in London in 2026. 

And in the last few days, multiple outlets have reported that Uber and Lyft are partnering with Baidu for UK robotaxi trials in 2026, in the context of a UK government pilot programme. Will it be seamless on day one? Of course not. But the point is: the debate has moved from “can it be done?” to “how quickly can it scale, and under what rules?”

Video created from a sentence (and it’s not just silent clips)
This is another threshold moment. OpenAI released Sora 2 on 30 September 2025, positioning it as a video (and audio) generation model, with a clear emphasis on responsible deployment. 

The practical impact isn’t “Hollywood is finished”. It’s more mundane and more disruptive than that: training videos, product demos, onboarding explainers, internal updates… content that used to require kit, people, and time can now start as a first draft in minutes.

Controlling a computer with your thoughts
This one still makes me pause. Reuters reported that Neuralink implanted its first human patient in January 2024. Reuters also reported the patient was able to control a computer mouse using their thoughts.

 
I’m not making a medical point here. I’m making a “trajectory” point: the interface between humans and computers is changing shape. We went from keyboards, to touch, to voice, to wearables… and now we have credible proof that direct neural input is possible.

When you line these up, a pattern starts to emerge.

• Language barriers are being attacked head-on (not perfectly, but materially). 
• Connectivity is being redesigned so it’s not dependent on terrestrial infrastructure alone. 
• Transport is moving from “human-operated by default” to “software-operated by default”, at least in bounded use cases. 
• Content production is being compressed, dramatically. 
• Compute itself is being reimagined, even to the point of relocating it off-planet. 
• And the human–computer boundary is getting thinner. 

Now, if you run a business, there’s an obvious temptation here: to treat all of this as interesting trivia. “Nice to know, but not relevant to me.”

I’d challenge that.

The relevance isn’t that you need robotaxis or space compute next week. The relevance is that the time between idea and usable reality has shortened.

So the better question, as we head into a new year, is: is there a better way to run your business that you’ve been ignoring because it used to be too hard, too expensive, or too complicated?

A few prompts I’d put on the table:

• Where are you still relying on manual re-keying, copying, pasting, and spreadsheet patch-ups because “that’s just how it’s done”?
• Where are you carrying a known weakness (slow month-end, patchy data, unclear margins, inconsistent follow-up) because fixing it felt like a massive project?
• If you could compress one cycle time by 50% in 2026 — quoting, invoicing, reporting, cash collection, onboarding — which one would change the game for you?

Looking forward, here’s my best guess at what we’ll call “impossible” today and “obvious” tomorrow.

• Always-on business intelligence, where your finance system doesn’t just report what happened, but flags what’s drifting as it happens and suggests fixes you can approve.
• Meetings that translate and summarise themselves, producing actions and follow-ups without anyone “doing the notes”.
• Near-zero friction compliance, where good underlying data means the outputs (returns, filings, management packs) are continuously prepared, not painfully assembled.
• Agent-like software that can move across your tools (banking, accounting, CRM, project management) and execute a workflow end-to-end, with audit trails and controls.

We won’t get all of that in one jump. But if 2025 has shown anything, it’s that “that’ll never happen” isn’t a strategy. It’s often just a failure of imagination.

So the best question to end the year with isn’t “what’s coming?”. It’s what “impossible” thing are you currently ignoring that, in three years, your competitors will treat as standard?

Best wishes for a peaceful and prosperous 2026.

Noel Guilford