Dispatches 001: Shifting Futures, Quad Espressos & Meridian 2.0

Long time no see, dear reader.

I confess that I have been a little busy.

When I originally started Meridian, I wanted to understand how innovation ecosystems work. What programs exist, what gets funded? How does the innovation picture look like at a national+ level?

As it happens, as I started digging into all this, we’ve seen perhaps one of the fastest reshuffles in the tech space for a long time. I’m talking about the knock-on effects of AI, of course, but it’s not just chatbots and Claude. The picture seems to be shifting a lot towards deep technology as software loses its moat. Biotech, novel compute, ever-grinding progress in robotics – it’s all looking interesting indeed.

The question I have begun to ask myself is less just how does innovation tick and more what gets built?

Naturally, my work has expanded into that direction.

AI has advanced massively over the past 6 months, especially AI-powered research. I finally have some pretty serious foresight setups that let me do the horizon scanning work at a scale I’ve always wanted. As a consequence, I can shift much more of my attention from simply gathering strands of thought to outputting more thoughtful, in-depth pieces on topics I care about.

I’ll talk a bit about that in a moment.

As for the Dispatches, this is my version of a miscellaneous section where I gather my thoughts on interesting shifts in tech, speculate on long-term progress and share tidbits of content I’ve been consuming. Considering my tendency to keep endless winding journals, this should have been a thing from the get-go.

Drawing up Meridian V2

I have always been a writer. I enjoy it. It clarifies the mind in a way that few other activities can, letting you unbundle interesting strings of thought while also forcing clarity. Deeply exploratory writing is a slow, deliberate process.

Sometimes a little too slow.

One of my initial mistakes was going after a far too massive chunk of information. Hence, as the astute among you will have noticed, I never actually finished the Compendium, although I did build a database of emerging innovation programs and initiatives across the world.

The research itself was fast enough, but actually covering it all at the depth I would have liked would have left me with about a year of work. All well and good, but a non-starter in the current fast-paced tech landscape.

On top of that, it was never the goal to just be a full-time writer. Analysis to me is a tool. Same with foresight. I do it because I deeply believe that knowing what to work on, and what the most powerful at-scale instruments are, is key to driving real progress. By the same token, bridging domains and disciplines — how education links with innovation, how policy can shape science, and so on — is my other favorite hobby.

So I went back to the drawing board.

I envisaged the original Meridian as a sort of proto-think tank. The goal was to do some interesting work around innovation toolkits and long-term trends. The logic was to approach progress from two sides: which acceleration or orchestration mechanisms we currently have, and where should they be pointed.

What I was truly interested in is intersections and long-horizon foresight directly linked with new seed projects. Case in point: FROs and novel lab-style orgs.

One thesis I’ve held for a while is that AI is driving a shift towards non-obvious problems; intersections between domains, long-term research, or downright futuristic stuff like how to govern autonomous digital organisations. These surfaces of contact between fields are not generally something most spaces create. Scientists exist in labs, tech people in startups, policy folks in governance, so on. With one exception: deliberately intersectional organisations.

These come in a couple flavors. There are your Labs and Studios: open-ended orgs that work on interesting problems instead of a specific product or vertical. Think Dark Matter Labs and their intersectional work. The second are Connectors: event series, fellowships, and organisations that bring together people who normally wouldn’t meet. ARIA in the UK is doing a lot of such convening work across science and tech; The Entrepreneurs Network, whose events I often frequent, does the same in bridging startups with policy.

Meridian was intended to achieve something of the same. It didn’t quite work on first pass, so I began developing a successor.

V2 was born out of three questions:

- What work is currently not being done that we are uniquely positioned to do, and that would be highly valuable for long-term progress? I like AI Agents as much as the next guy, but there are roughly seven trillion AI agent startups. And while I’m deeply interested in deep tech, I am alas no scientist. And so I landed at inflection-oriented foresight bridged with innovation ecosystem projects.

- Can we concentrate our work directly on crucial inflection points? That means understanding what those are, what they enable, and what sort of new solutions they demand. Take AI and Future of Work. This is a massive reshuffle that raises questions about new education models, policy, the job market, all alongside how technical capabilities of current systems will develop. AI itself has been a major breakthrough here. I’ve ported many of my foresight frameworks to Opus 4.7, which now augments a lot of my ‘cascade analysis‘ specifically looking for intersections and feedback loops. This volume of work just wouldn’t be feasible, but now I have a large-scale initial mesh to pick up these dynamics.

- How can the analysis, foresight and research work be linked to actual output? Analysis is fun. It also lives in a vacuum unless deployed or applied, never to be witnessed by human eyes again outside of a cursory read. My response to this was across two areas: 1) co-releasing all major essays alongside an artefact such as a prototype, recommendations report or a project blueprint, and 2) seeding ecosystem projects like fellowships, events and communities where we find gaps.

Think of it as a direct foresight-to-artefact loop. We run 5-10 year horizon foresight around long-term macrotrends, identify where the waves are shifting, which gaps exist, and directly seed projects as a response.

I want to be transparent here: this is work for a major organisation with the resources of a caliber I simply do not possess. A kind of ultimate form of this is a Foresight-Venture studio that directly spins out projects through collaboration with external teams, like a fellowship or venture builder model.

Meridian V2 is in some sense a dry run for exactly such an org. There will have to be a Meridian Labs V3.

Alas, I simply will not be able to achieve the kind of impact I would have liked with this. What we can accomplish, however, is extend how models like this can work and create the groundwork for further expansion. That is the intent.

Needless to say, I will need a lot of coffee. Blessed be the baristas serving me my favorite Colombian filter.

Probes, Sonars & How Will This Work?

There are something like a dozen ways I could do this.

I already run more classic industry analysis over on LVCN, for whom I publish DeepTech VC research covering emerging technologies, fundraising movements, and companies to watch. You can read it on LVCN Publications.

What I wanted for Meridian was more long-term. This sort of work is usually called strategic foresight, a horrendously fancy name that I am unfortunately very fond of using because it’s both awesome and very fitting.

My (not overly sincere) apologies.

For those interested in how this works, I invite you to look up the Copenhagen Institute of Futures Studies, CIFS for short. Somewhere between them and government foresight taskforces is where the most interesting futures work is carried out.

Since I wanted something bounded but with enough room to explore long-term implications, I designed our own probe cycle.

Much of this is powered by Sonar, a collection of Claude Skills and overlaid frameworks that let me do large-scale horizon scans. To give an example of what this looks like, I’ve been demoing an early version of Sonar for an upcoming issue on AI Economies:

Beyond just research, I’ve been playing around with a framework that models interaction pairs between trends & effects.

It’s all just raw information for now, of course. I’m looking to get to a point where we’re designing artefacts directly responding to several overlapping trends. Eventually, this will be done in collaboration with external teams.

Meridian Fellows, perhaps?

We’ll publish the first couple of Sonars over the next two months. For now, I’m digging into AI x Future of Work as above, plus blueprints for novel education formats – something I believe will be urgently needed as the entire entry-level role landscape and job progression gets reshaped in the span of just under a decade.

That’s all for now. Perhaps I’ll draft up a proper announcement at some point. We do need a new logo for one, and I’m upgrading the branding, since Meridian is now closer to a foresight lab than just a publication.

I will say, though, this project is one of the most interesting things I have ever worked on. It has taught me an immense amount – and I’ve barely even launched the damn thing!

Do follow along, because this is going to be fun.

‘Till next time.

Dispatches 001, Meridian Labs