§@pmoe — Philipp Moehring

Philipp Moehring is a Berlin-based co-founder of Tiny Supercomputer Investment Company (tiny.vc), the pre-seed/seed firm he started in 2017 after running AngelList Europe. Tiny has backed 450+ companies — n8n, Wayve, Synthesia, Payhawk, Tide among them — at $200–250K a check, deliberately never leading rounds and never sitting on boards. His tweets are the public surface of that strategy: small, contrarian-by-default, allergic to thesis-driven framing, written from inside the European venture scene by someone who has watched it mature for nearly two decades. 100 posts, 60 replies, four top-hit threads, his sundaycet.com interview, and two podcasts inform what follows.

§I. Core Worldview & Mental Models

§The look-dumb-now-be-right-later trade

The single most load-bearing idea in Moehring's corpus is deliberate contrarianism on a five-year clock. He stated it most plainly in his Oct 2025 sundaycet.com interview ("everything that's overcrowded is too late for us to invest in… look pretty dumb for a while — that's the reality of pre seed investing") and then tweeted the compressed version on Apr 30, 2026: "There are those that back things that are cool today, and those that look for things that will be cool in two to five years. I aspire to the latter, and it is just as messy and unclear as it has been decades years ago." The Tiny fund he raised before ChatGPT — "it was not clear at all that transformers and LLMs would be what would underpin the next wave" — is the personal proof point he keeps coming back to. The mental model is call your shot before it's legible, then sit through the period where everyone thinks you're wrong.

§Anti-playbook, anti-grand-thesis

The whole sundaycet interview is titled around the line "There is no Euro playbook, just startup dudes getting shit done," and Moehring's own claim underneath is sharper: "the most amazing companies look nothing like a playbook. That is a humbling experience." He found, contra his own prior, that his biggest winners didn't pivot — "the general direction of travel remained pretty straight for our biggest winners" (Wayve, Tide, Payhawk). The consequence: he rejects "thesis investing" as a posture and treats his job as recognizing founders who already know where they're going. On Mar 12, 2026 he ratified this externally — "For some people, Deep Tech is just Venture Capital, for some it's this" (quoting @lessin) — a one-line diagnosis of the deep-tech narrative as a category that lazy VCs adopt to sound thesis-driven.

§"Apply some common sense"

The single most cited frame across his post and reply corpora is contextual sanity. His Mar 20, 2026 self-thread under "Venture isn't dead, early stage isn't dead, small funds aren't dead" (75L) is the cleanest example: he walks through the numbers in four self-replies ("Seed rounds tripled over 10ys, As to Ds tripled in 5ys"; "Just apply some common sense to the metrics you see, and put them in context. I miss the smart and nuanced writing of the late teens, everything today is breathless enthusiasm or total doom, and it's not helping anyone"). This is the master rhetorical move. He doesn't argue an opposite extreme; he argues for less affect, more arithmetic.

§Intellectual DNA

Moehring is unusual among VC tweeters in that the people he quotes are mostly not VCs. Tier 1 is Matt Levine ("I could just quote @matt_levine every day…" Dec 18, 2025) and Paul Kedrosky ("H/t @pkedrosky (who everyone should read)" Apr 30, 2026; "If you read @pkedrosky, you would know exactly! Hint, starts with private, ends with credit" Mar 20, 2026). Tier 2 is Ed Zitron ("@edzitron you've got a way with words" Nov 15, 2025) — notable because Zitron is an AI bear, and Moehring is a working seed AI investor. He reads books and lets you know: William Gibson's Pattern Recognition quoted at length (Apr 23, 2026), Norman Davies' Europe on Nov 26, 2025 ("I've rarely marked up so many small details in a history book"), Heinlein invoked to mock Musk ("Reads Heinlein once…" Mar 22, 2026). His Mar 18, 2026 tweet — "If you can't detect AI generated text, you're not reading enough books" — is the self-portrait. He's not anti-AI; he's pro-having-read-things.

§Where his thinking has shifted

Across the 6 months of corpus the shift is mostly tonal. The earlier window (Oct–Dec 2025) is more dispassionate-market-commentary. By mid-spring 2026 he is markedly more combative on regulation (UK carry tax Jan 29; NY AI bar Mar 7–15; CA demographic reporting Mar 12) and more emotional about Germany (Apr 16 on the CDU; Mar 15 in German about Bavarian power lines). The five-year prediction layer is steady; the political layer has thickened.

§Blind spots

Two are visible. (1) He prides himself on contextual nuance but has very little to say about the failure modes of his own portfolio — the corpus has no losing-position post-mortems, only the survivor framing of "look pretty dumb for a while." That's a survivorship trap built into being a 450-company index investor: you can always point at Wayve. (2) He's structurally under-confident about ideology. He has strong takes on tactics (secondaries, fund size, board seats) and on aesthetics (cars, design, restaurants), and nearly nothing on the question his portfolio actually rides on — what AI should do to labor markets. The Anthropic "labor market impacts" link gets a flat "The official Claude endangered species list" (Mar 8, 2026) and a follow-up linking the long version, with no commentary. The biggest question in the corpus is the one he most carefully declines.

§II. The Tiny Model: Index Investing, ROW Behavior, the European Game

Most of Moehring's posting is in some way about how a small European seed fund should operate. The shape is consistent enough across the tweets, sundaycet interview, Sifted podcast, and 20VC appearance to summarize.

Never lead, never on boards, write 200+ checks. From the sundaycet interview: "we are seed investors, not growth people, so it's probably better to not sit around the table when the companies grow." And on naming: he is pointedly trying to retire "unicorn" — "I try to move the term unicorn out of my vocabulary, because we look for multipliers" (Apr 23, 2026). The frame is the multiple, not the valuation, which is what an index investor would say.

Secondaries are mandatory. Feb 20, 2026: "If you don't actively understand and use secondaries as a VC, you're ngmi." This is the operative claim of the April 2026 Sifted podcast appearance, titled "on when to take money off the table." This stance has aged well — he's reading @ttunguz's data point that secondaries went from 3% to 31% of exit value and saying out loud that the seed crowd has not yet adapted.

Hadley's "VC fund of the future" is already here. Mar 3, 2026, 80L: he quote-tweeted Hadley's prediction ("a small number of highly experienced, high-agency people with deep trust… surrounded by tons of agents") and appended "The VC fund of today if you ask me." This is unusually self-revealing — he's claiming his five-partner outfit has already arrived at the destination others are still forecasting.

The European mistake to avoid: the "familiarity discount." In the sundaycet interview he names a specific local-VC behavior — demanding cheaper terms because they're nearby — as the thing keeping European seed second-tier. On Apr 27 he reduces this to a private joke in a reply to Andreas Klinger: "Not allowed, ROW behaviour." "ROW" — rest of world, i.e. anywhere outside the SF/NY axis — is shorthand he uses with friends to mean behaving small because everyone around you is. The whole Tiny posture is built against it.

Anti-billion-dollar-seed. On the Sifted podcast: "Thankfully, the billion dollar seed round is not the standard across Europe — yet. That would be concerning." This is the same complaint as his Mar 20 jab at @FinnMurphy12 about "2024 vintages, no less... Everyone should be a little bit more... should I say journalistic?" — both attacking the OpenAI-shaped distortion in deal-count and aggregate dollar figures.

LP supply, not deal supply, is the actual European bottleneck. From the sundaycet interview: "if you want to build London, Berlin, and Paris outcomes, you better equip these teams with the right money." He returned to this on Feb 16, 2026 mocking California's new mandatory VC demographic reporting — "And people bash the EIF and BBB for reporting requirements…" — which only makes sense if you read EIF (the European public LP) as a structurally good thing whose reporting overhead is the cost of money supply.

On Tiny's own ranking. Feb 26, 2026, 70L: "Thank you to all the founders who let us come along on the journey — the only reason we get to rank high on any list." The self-reply names his team explicitly — "Thank you Andy, Ophelia, Sandra, Alex, Harriet and Toni for being the most amazing team to do this with. You could be on social too, but you just choose to do the work." Read this twice. He is, intentionally, the only public face of the firm. The "you choose to do the work" framing is half-celebration, half-confession.

§III. Operating Principles for Investors and Founders

Distilled rules-to-live-by, each pulled from a specific post.

1. "Look pretty dumb for a while" is the job. "Everything that's overcrowded is too late for us to invest in." (sundaycet, Oct 2025.) The pre-seed return profile requires entering before the category is legible.
2. The multiple matters, not the valuation. "I try to move the term unicorn out of my vocabulary, because we look for multipliers." (Apr 23, 2026.) An honest reframing of how index seed actually works.
3. Don't lead, don't sit on boards. "We are seed investors, not growth people." (sundaycet.) The Tiny refusal-to-grow-up.
4. If you're hiring outside your operational experience, borrow it. "As a founder, if you are hiring for a role that you have little or no operational experience with, please involve someone who is excellent at this job at a larger scale. You can do that by asking your investors and founder friends for references." (Mar 20, 2026.) One of the few straight-up tactical posts.
5. Stop optimizing for what looks important to others. Reposted from @bryce on Apr 10, 2026: "Once you stop doing what other people find important, you suddenly have so much time to spend on the things that matter."
6. Time will not make you a top fund. Ten years might. Dec 17, 2025, to new fund managers: "It's the time of big year end reflections… Remember, this stuff takes time. No one expects you to be at the top this year, next year, or in the next five. Look at the successful vintages of the top funds, and you'll see most of them are 10+ years old. Just do your work and tune out the noise. Stay real, honest, and yourself." The single longest standalone post in the corpus. He is consciously writing advice to his younger self.
7. Start weird, or be world-class at conventional. May 5, 2026: "If you don't start weird, you better be incredibly excellent on the most commonly valued things. That means you compete with the whole world. Good luck. If you don't start weird you're probably trying to compete inside the local maximum of something more boring than you could be." The clean version of the contrarianism principle, applied to founders.
8. Use secondaries. Feb 20: "If you don't actively understand and use secondaries as a VC, you're ngmi."
9. Hidden-gem-as-sure-thing. Apr 30: "If you're a hidden gem, start to behave like a sure thing. Else you will stay hidden." (Quoting @JayKapoorNYC.) Tactical advice for the founders he funds.

§IV. Rhetorical Style — Berliner Dry Wit at the Quote-Tweet

Moehring's voice on the timeline is distinct enough to identify blind. Three moves dominate.

The dry sub-tweet of culture. "The new em dash" (Mar 3, 2026, on the emerging tic of GPT-style paragraph rhythm). "And this, son, is how we stopped AGI" (Mar 7, 2026, linking the proposed NY AI legal-advice ban). "Ain't no party like a Donner Party" (Feb 18, 2026, over a Donner Summit blizzard photo). Each is one line, payload-heavy, no setup. The five-token political read of the CDU on Apr 16, 2026 — "€1000 tax free stimmy checks for employees / Shutting down solar subsidies and battery storage, preferring gas power plants / Putting a moratorium on gas tax (at gas companies' discretion to forward) / Screaming about nuclear plants they shut down in 2011" — is the same move at list length, ending with "This is would be funny if it wasn't so sad for our country."

The "Reads Heinlein once…" A whole class of his tweets is a single arch-phrase quote-tweeting somebody self-important. Musk announcing moon mass drivers gets the Heinlein read (Mar 22, 2026). David Sacks declaring Miami will replace NY gets "It's funny that even Silicon Valley says Silicon Valley is just a mindset" (Jan 2, 2026). Bryan Johnson confessing he over-packs longevity gear gets a single 🚨 (Nov 12, 2025). These are his highest- hit-rate tweets relative to length.

The trade-the-payoff tweet. "I was a never tattoo guy until I saw this thread" (Oct 29, 2025) is his single most-liked tweet at 384 likes — a Simpsons tattoo thread, quote-tweeted with a one-line confession. The pattern: pick a piece of culture, undersell your reaction, let the contrast carry. "3 days in Japan and I'm looking at importing toilets" (Apr 23, 2026, 35L) is the same shape. "My wife will absolutely love this" (Mar 3, 2026) is the smallest version of it.

Audience vs. intent. The 246L "Funny to see people complain about @steipete deciding to spend his time at OpenAI… Peanut gallery galore" (Feb 18, 2026) shows how he uses register-clash deliberately. The thread that followed was full of replies amplifying his line ("Yeah the entitlement is weird" — 676L on the top reply), but it also surfaced the genuine counter-argument he didn't engage with: "We are speaking about those people who are making innovative businesses in Europe impossible and then complaining when innovative business cases leave europe." That reply got zero likes. Worth noting because his audience hears European indignation defended, while his own reading is closer to don't tell builders what to build. The audience heard a tribal post; he wrote an individualist one. He let the misread stand.

§V. Contrarian & Hidden Takes

What he actually thinks, distinct from the median Euro-VC line.

Prediction markets are a scam, not market wisdom. Apr 26, 2026, on a working paper finding 3% of Polymarket accounts drive price discovery: "Not really surprising: in an unregulated market where insider trading is not easily trackable or punished, we should expect shenanigans. I believe none of the 'making markets efficient' in this debate. It's gambling at best and — turns out — asymmetric rug pulling in reality." This is sharp by his standards and cuts directly against the rationalist-VC fondness for Polymarket-as-aggregator.

Private credit is the next blow-up. Feb 19, 2026: "It was a brief period of unrestricted enthusiasm, kind of cute to see." On Mar 20 he made it the punchline of a reply to @villi: "If you read @pkedrosky, you would know exactly! Hint, starts with private, ends with credit." He is borrowing Kedrosky's view but he's signed up for it.

He is genuinely happy to be European. Jan 14, 2026: "I'm extremely happy to be European." Said flatly, without irony. Three months earlier, on Nov 24, 2025: "This rings true in any tech ecosystem but Silicon Valley. That's why building companies elsewhere isn't just about the 'mindset', it's also about bridging the gap between those who get it… and those who don't." In a tribe that mostly talks about leaving, naked European patriotism is the contrarian position. He paid for it on Feb 18 by defending @steipete's choice to leave — showing he's not a Europhile by reflex.

Brexit was bad but the British response is good. Apr 26: "I hate Brexit (for obvious reasons), at least the typical British pragmatism is dealing with it in refreshing ways. We don't all need to look across the Atlantic for inspiration on how government can (partly!) make sense of current challenges." A surprisingly generous read from a German VC running a UK team.

The big AI legislative threat is the bar associations, not safety. Mar 15, 2026: "There is also legislation on the way to ban all legal related advice by AI in NY on the basis that it's a protected profession that requires Bar, etc. This is the most likely successful angle slowing down AI. Look out for it elsewhere…" He's saying the quiet part out loud: AI's actual regulatory enemies are professional cartels, not the safety community.

Anti-Bryan-Johnson, anti-protein-powder. Oct 15, 2025: "I posit that the negative unintended consequence of a lot of the Blueprint and general longevity trend will involve poisoning through some under researched supplement BS. If it starts with protein powder (the grand current obsession), it's gonna get a lot worse from here." In a VC scene that mostly cheerleads biohacking, he's calling the category dumb.

What he'd say after three drinks. Probably already said it: the German political system disappoints him, Musk reads like he stopped at one Heinlein book, most VC public commentary is breathless or doom and "it's not helping anyone" (Mar 20 self-reply), Polymarket is rug-pulling, and the European VC peer group includes "SO MANY talking heads who aren't writing any checks" (sundaycet). The only thing he probably won't say on-timeline that he'd say off it is which firms he means by that last one.

§VI. Network Graph

The five he treats as primary sources, not peers. Matt Levine, Paul Kedrosky, Ed Zitron, William Gibson, Norman Davies. He cites them as authorities — they're who he reads, not who he debates.

The peer benchmark cohort. @PeterJ_Walker (Carta data; multiple back-and- forths including a fun jab Apr 23 — "Hey man, I gotta drop some sound bites, otherwise I stand no chance against… ah yeah, multistage funds of course"), @ttunguz (data citations), @Hadley (he name-checked Hadley's "VC fund of the future" tweet with his answer "fund of today"), @bryce (the "focused not laid back" post he reposted with explicit admiration), @lessin (cited on deep tech), @JayKapoorNYC (the hidden-gem line), @sarahcuda. These are the people whose framing he uses publicly.

Berlin / Europe inner ring. @christianmiele (two replies, one in German about Bavarian power infrastructure), @andreasklinger (frequent reply partner, "ROW behaviour" is a private inside joke), @__tosh (multiple hardware-trade-in replies), @robindchnt ("Imagine 2 pmoes, what a disaster!" Feb 26), @nathanbenaich (Air Street), @drnovac (gave him the sundaycet platform), @b_johannsen (Sensmore portfolio), @amaankahmad. This is his actual day-to-day social network on X.

Portfolio he amplifies. @n8n_io, @Wayve_ai (referred to elliptically: "if I would start a business like that, I would give it a name that rhymes with shmayve" Nov 14, 2025), @Synthesia_io ("Should've used @synthesiaIO for proper dubbing!" Feb 9, 2026), @Payhawk, @SensmoreAI, @CoMind (James Dacombe). The amplification is always founder-name-first, never product-pitch.

His team — named once, then never again. Andy Chung, Ophelia Cai, Sandra Lyness, Alex Gezelius, Harriet Dedman, Toni Everitt — all surfaced in the one Feb 26 self-reply, with the gently barbed line "you just choose to do the work." This is the most revealing piece of social-graph data in the corpus: Tiny is structured so one person speaks publicly and five do not.

What he ignores. US growth-stage Twitter (no a16z, Sequoia, Founders Fund quote-RTs except for the one Sequoia 1977-Apple-memo post). Crypto Twitter entirely. AI-doom Twitter. Sam Altman. Marc Andreessen. The corpus is striking for what's missing: he doesn't engage with the SF cathedral at all.

§VII. The One Essay He Keeps Rewriting

It is the same essay every time, and it has four moves.

1. Stop trying to look right in real time. (Apr 30, 2026: "two to five years." Oct 2025 sundaycet: "look pretty dumb for a while.")
2. The smart move is the small move. (Mar 3, 2026: "fund of today." Mar 20, 2026: "small funds aren't dead." Feb 20: secondaries. The whole Tiny model.)
3. Stop trying to fit a playbook. (sundaycet: "the most amazing companies look nothing like a playbook." Apr 10: "stop doing what other people find important." May 5: "if you don't start weird…")
4. It will look like a straight line in retrospect, but it doesn't now. (20VC: "When you look back, it seems to be a very straight line… It didn't seem like a straight line at all." sundaycet: "the general direction of travel remained pretty straight for our biggest winners.")

The essay is, essentially, a defense of patience-as-strategy in a category that pays for confidence-as-strategy. Read against his portfolio it is self-serving — index investing requires no thesis, so naturally the head of an index fund argues no-thesis-is-best. But read against his own career arc — Seedcamp → 500 Startups → AngelList Europe → Tiny, all variations of "more deals, smaller checks, less narrative" — it is also the only thing he has ever actually believed.

The reading list around the essay is short and consistent. Matt Levine for the discipline of describing what is actually happening. Paul Kedrosky for what the numbers say once you stop yelling. Ed Zitron for the angry counter-read he needs to test his own bullishness against. William Gibson because the line in Pattern Recognition he quoted on Apr 23 — about the soul being left behind by transatlantic flight — is, in his framing, what also happens to founders whose company outgrows the founder's original question. They get to scale before their original belief catches up. The straight line, in retrospect, is the soul finally arriving.