Time for some updates, isn't it?

So for the last few months I've taken mostly off of work for personal reasons (no worries, I'm doing well!). I'm ramping back up now. During that time I've not maintained full radio silence, but I've not done much more work other than basic maintenance, some wrapping up of projects, and a little post here or there — in other words, nothing major, and it's been good. I'm very happy I had the opportunity to do that, I needed a bit of a mental reset.

So, here I am, back and ready to roll. Which reminded me that this newsletter list exists 👋

--x--

As you might know, I have a someone complicated relationship with AI: While I most certainly am not on the hype train, I also don't fall squarely into the critics camp. But given the speed and momentum of what's been going on, I figured it's time to share how I see this all going right now. In no particular order:

1. The AI story is, right now, the story of everything. - It's almost the whole story of the economy (together with, but quite possibly more important than, even the turmoil around oil and energy more broadly.) The US economy is basically an AI story at the moment, and since it's the world's biggest economy, that turns the global economy into an AI story. One easily accessible source I find helpful is Kyla Scanlon, an economist and author with a strong Insta presence who picks up on this thread well I think. - It's a key story in geopolitics. There's a race to set the standards (i.e. to dominate) AI globally, even though the people behind the big AI labs all claim that whoever wins the race basically is a risk to society. (But at least it's our risk! Or: At least it's an American risk!* Etc. etc. But any race to the wrong place just gets you to the wrong destination more quickly: Ugh!) - As part of geopolitics, it's also a story of completely reordering global power dynamics. Sovereignty around tech, military, security, future industries: Right now, that's all AI. At least that's the dominant narrative.
2. The economics of AI don't make sense to me. I see generative AI and LLMs as useful in some areas. I also see them as wrong too often to be useful in many, many others. For something to be this huge (see my points above) in terms of investment and power and politics and resources, a technology would have to be massively more useful and reliable than this. It's not, though! At least not as far as I can tell. The labs also seem to be mostly losing money the more people use their services, which isn't really a business model. Alternatively, to not lose money they'd need to raise prices, but then suddenly the users would stop benefitting from their investor-subsidized prices and would have to pay the real prices for their use, which would lead (I think) to very different cost-use tradeoffs. Would you really have AI write that LinkedIn post or that Powerpoint if it cost $5 per pop? Or 25? So, the economics to me seem idiotic. Yet, we're gambling whole economies on this to pay off. If it did pay off, the value capture would be highly centralized in the hand of a few companies, though: It's not like slightly higher productivity would automatically lead to more buying power. So even the underlying premise doesn't quite seem to align for me. I could be wrong on this, and frankly on everything here, but I don't think I am. Feel free to convince me otherwise, I'm more than willing to learn.
3. The laser focus on one subset of technologies and approaches (LLMs, gen AI) within a small subset of technologies (AI more broadly) at this scale is a major issue no matter what else might be true. Nothing good comes from such a focus. The opportunity cost is simply too high, plus it introduces potentially catastrophic break points. Every dollar spent on AI data centers is a dollar not spent, for example, on prevention of climate change. It's also a dollar spent on a bet on basically a handful of companies, literally all of which have problematic leadership and governance structures. They're simply not good places on which to put our bets!
4. I don't think of AI as a bubble. Not because I don't think it won't blow up in our faces, but because the term bubble doesn't seem to be the right term for the scale and intensity of what we're seeing in terms of flow of money and political efforts. A bubble bursts, and it stops being shiny. It's mostly air. This right here has more tectonic scale. If and when this blows up, it's going to be a lot more hurtful than any bubble we've seen before.
5. If you'll indulge me, let's discount the extreme utopian vision of AI: humanity automatically will flourish, or rather an automated humanity will flourish, and everything is perfect and everyone is rich and humanity will healthily settle among the stars, through magic something something question mark. Even without going to the extreme dystopian end, there's no plausible happy end here, no good, desirable end state worth aiming for. The labs talk of existential risk, which seems bad. Mildly better outcomes still come at significant external costs in the form of climate impacts (via extreme energy use), as well as as-of-yet-unknown impacts on our cognitive capacities over the longer run if we automate more stuff rather than challenging ourselves. The AI promise is based on the premise that it might offer shortcuts. Maybe it does, maybe it doesn't. We don't know, and we don't even know what we don't know. When it comes to learning, I'm somewhat skeptical of shortcuts.

Which brings me to this shortest of summary: I'm not sure if there's anything great at the end of this particular rainbow worth chasing. But I am sure that this concentration of effort, resources, money and political power should only be spent on areas where we know it to pay off. Currently, and probably for the next 100 years, that's climate, and nothing else. AI would be a nice enough side bet, but it shouldn't be the top priority. That simply doesn't make sense.

--x--

Just by way of re-upping two pieces of writing in case you missed them, early in 2025 and 2026 I wrote two little pieces. Booklets might be the right descriptor? I wrote them with the intention to be really quick, skimmable reads:

• The Little Book of Strategy is a quick, actionable guide to get you unstuck on your strategy and leadership journey, i.e. some distilled little nuggets from about 20 years of strategy work. This version is an updated and slightly expanded version from when it first came out last year.
• The Little Book of Public Interest Technoliogy is what I think of as a practitioner’s field guide to power, infrastructure, and tech governance.

Both are very much written from my personal experience in those areas. They're not academic, not heavily sourced, they're basically a glimpse into my head when it comes to these specific areas.

--x--

I just realized I should probably also post this on my blog. But hey, you heard it here first :)

If you're mulling questions or exploring projects, hit me up and let's chat. But more importantly, I hope this finds you well and you're having a great start to this week.

All the best from Berlin,

Peter