Raphael Poss

Time to Ground Ourselves Again

· Raphael Poss · reading notes, ai commentary, economics, entrepreneurship, policy

I first started reading Halvar's Guide to Entrepreneurship by Thomas Dullien as a distraction from another task, and then caught myself reading it with attention. This is an experience report on organizational design; it confirmed my own experiences in some areas (product design, management), so I became curious about what the author had to say in other areas. I'd recommend it! It is not as comprehensive as the High Growth Handbook that I read two years ago, but certainly more compact and approachable.

❦❦❦

In the Netherlands, the months of July and August are called “cucumber times”. That is because real news is thin in summer and so all there is to talk about is cucumbers. Keep this in mind. We'll come back to it in a second.

Meanwhile, usually, historical story arcs are covered through calendar years. The following arc is best viewed as what happened between summer 2025 and summer 2026:

  1. business owners are induced into executive psychosis by their “smart” chatbots and start making dubious decisions.

    Example: Bosses Are Becoming Obsessed With AI, Using It to Make Every Decision, Barraging Their Employees With Nonsensical ChatGPT Directives, and Even Asking It Who to Fire, Maggie Harrison Dupré (Futurism).

    (This one would have made me laugh if it was satire—it's so grotesque! Sadly, it's all real.)

  2. ... some time passes ...

  3. It did not work as expected! Oh no! Who would have known?

    Ford hired AI and sacked humans. It backfired badly, Anthony Cuthbertson.

    Meta's Zuckerberg says AI agent tech progressing slower than expected, Katie Paul and Courtney Rozen.

  4. experts in the industry start to pile on with “I told you so” retrospectives. This is going to be the cucumbers of summer 2026.

Maybe that summary sounds glib? I don't know. In any case, I found these following two analyses quite good. If you want to take only one piece of LLM-related news away from this month, take these two.

In hindsight, it looks like I chose the right time in history to start my own bootstrapped project away from corporate madness. 🤷

❦❦❦

Maybe the right move at this stage is to start grounding ourselves again.

One piece that touched my heart is this plea by Jay Acunzo to celebrate lived experiences: The best response to AI slop, infinite advice, and online noise is from Robin Williams. Choice quote:

Right now, the mere idea of "knowing" is winning out, when it's hilariously insufficient to do anything effective or meaningful or good. The internet and many voices using it would have us convinced that because there's so much to know and because someone else knows more and because a software tool can tell you things it "knows," then all of that should cause you to sit down, shut up, stop feeling confident in what you've done or seen or felt. Because what's your life in the face of infinite knowledge?

Turns out, everything.

In a nutshell, the author reminds us of what makes us interested (and interesting) to each other and asks us to do more of that, with or (preferably) without LLM assistance.

❦❦❦

Switching gears and away from LLM news.

One of my teenage years interests was computer vision, and today is as good a day as any to recall that our eyes see more than a screen (or a printer) can display.

In Where to Find the Colors Your Screen Can't Show You (Ryan Moulton), we learn the math of why the light diffused under the leaves when we walk in a forest is a shade of green that can only be found outdoors. Iridescence and colors on birds and butterflies are unique experiences as well. And, perhaps surprisingly, the green LEDs of traffic lights. Per the author:

I can’t show you these colors, but by telling you about them I can help you notice them. When you notice, you may be astonished to find that they were there all along, and that your screens are duller than you thought they were. When you drive home today and see a green traffic light, notice it. Try to see it as bright and as beautiful as it really is.

Incidentally, What every coder should know about gamma by John Novak is also relevant here.

❦❦❦

Looking at economics, a lot of big and important things are happening in the world in 2026.

Somehow, I feel that a lot of the the big item news can be looked at through the lens of textbooks from the 20th century. However, two things happened in 2026 that really feel new, dangerous and complicated, for which we don't have any handbook yet: the liberalization of prediction markets and the first steps towards legal integration of cryptocurrencies into the financial system.

The best piece I've read on these topics is Crypto in 2026: Oh, This is the Bad Place by Stephen Diehl. The title borrows from The Good Place, where the protagonist realizes the “paradise” she's living in is actually an engineered hell. Diehl's conceit is that hell isn't a fire pit but a world so absurdly awful that the laughter it provokes is what eventually tips you off that you're in it. He argues 2026's crypto-financial landscape has reached exactly that level of normalized absurdity, and organizes his case around a few fronts.

The core thesis here is that new high-risk instruments are being sold to unsophisticated retail customers, a business the author calls “sucker farming,” a practice that used to be prohibited but is becoming legally sanctioned. He identifies two specific harms:

  • The “casino pipeline”, the modern economics variant of gateway drugs. Poor people get onboarded through memecoins, graduate to leveraged options, then Kalshi-style event betting, then late-night sportsbook parlays, all while believing they are “investing.” Meanwhile, the corresponding regulatory institutions (like the CFTC) are being neutered.
  • Stablecoins. The GENIUS Act arguably formally integrated dollar stablecoins into the global monetary system: the author contends the OCC has been pulling crypto issuers inside the federal bank-chartering apparatus — and thus inside an implicit federal backstop — on what he considers a bogus legal theory that issuing redeemable par-value claims is a “trust” activity. This leads to a major new financial stability risk: the largest crypto issuers hold Treasury piles that rival sovereign holders, so a run could force a fire sale into a fragile Treasury market.

The piece ends with a number of policy proposals; even though their phrasing is US-centric, most can be translated to any other locale.

❦❦❦

References:

← Prev: Life Games: The Conducts We Choose Next: Things That Happened Quietly →
← Back to blog