Leopold Aschenbrenner is the most famous man on fintwit.
Plus: a new Polymarket-backed media company, Andrew Carmellini's new tavern, and more.
Good afternoon, everyone.
Has anyone had the steak frites at Bar Oliver? I got a drink there last night with Jasmine Sun and I was distracted by the two guys next to us happily eating steak at the bar, which according to their menu is served with piquillo jus. Next time. On the way out of the bar, I saw Brendon Holder. A very small (China)town night.
In today’s newsletter: A new Polymarket-backed media company, Andrew Carmellini’s new restaurant is a tavern, an L.A. artist embarked on a six-week bender in New York’s downtown art scene, Teddy Kim on streamers going vertical, Leopold Aschenbrenner is the most famous person on the internet today, and a fall fast fashion campaign shot at Fanelli.
Once again we’re reminded that Quibi was ahead of its time. By Teddy Kim.
Stay Tuned is a Feed Me guest column about film and entertainment, written by Teddy Kim.
Rewatching The Social Network last year, I was struck by how outdated the name “social media” is now. We used to go to the apps and see what our friends were up to. Now, for the most part, what we see on Instagram or X is videos made by people we don’t know. Oftentimes the most social part of it all is just sending Reels to people in the DMs.
Somewhere along the way, there was a business and product decision made to transform “social media” into “algorithmic media” because doing so increased the amount of time people would spend in-app and maximized the number of ads that could be shown, juicing revenue.
I was reminded of this when I saw the news that HBO Max is launching “HBO Max Shorts,” which will feature clips from HBO content in a vertical short-form feed so people can scroll endlessly through scenes from the streamer’s entire catalog. Once again we’re reminded that Quibi was ahead of its time.
As a business decision, it makes complete sense. Netflix already launched a “Fast Laughs” feature in 2021 and then a “Clips” feed earlier this year. Half the clips I see on X videos and YouTube Shorts are TV and movie scenes (with annoying music and AI narration over them). So why wouldn’t the streamers move to do it themselves if their AVOD (advertising-based video on demand) competitors are just going to boost their own engagement stats with clips of scripted content they don’t even own?
“Ad revenue is still a small minority of total streaming revenue, but it all makes me wonder what the end game is here. Do the financial incentives of the ad-based business model put us on an inevitable trajectory of slop carcinization?”
This outcome didn’t always feel so inevitable. Not long ago, streamers like Netflix were ideologically opposed to ads, proudly brandishing instead their subscriber numbers and their recurring revenue business models as they invested billions of dollars in cash flow into original content. Most investors were more than willing to fund that cash burn. Media competitors wanted to get into the streaming game and get treated, and valued, like sexy tech companies instead of boring old media companies.
That all changed in April 2022 when Netflix reported its first net subscriber loss in over a decade and its stock dropped 35% in a day. Investors realized ZIRP (zero interest rate policy) was ending, capital was not free, and even companies like Netflix were now expected to turn a profit, especially if growth was now leveling off.
Later that same year, Netflix launched an ad-supported tier to the service, allowing subscribers to pay a lower monthly subscription in exchange for watching ads while streaming. Disney+ followed the next month, in December 2022. HBO Max had launched one the year prior in an effort to juice subscriber numbers and catch up in scale with its competitors.
Looking back, that’s when the Faustian bargain was made. Streamers started to realize that ARPUs (average revenue per user) for subscribers on these ad-tiers could actually be higher than the non-ad ARPUs and started pushing their mix of revenue towards ads to keep growing their overall revenue even as subscriber growth stalled off. Ad-supported subscriptions now account for nearly half of US streaming users.
Netflix has always been notorious for its expansive view of its role in our lives, introducing the concept of binge-watching to the world and, in 2017, famously describing sleep as its top competitor. But back then, the traditional Netflix model didn’t literally rely on maximizing every second you spent on the platform. You just had to value your subscription enough to never churn, something that could be achieved by watching a few movies and a TV show every month. That allowed, too, for a potentially healthier alignment between the interests of the companies and the lives of its users.
But now that more time spent watching directly translates to more revenue, I’m not sure what stops the streamers from continually feeling the pressure to evolve their brands and businesses in order to feed the new financial reality. Forget sleep, their competition is companies like YouTube, TikTok, and Instagram, platforms where people spend hours watching highly engaging content that other users upload for free. In the attention-maximizing world, does it still make sense financially to spend billions on TV shows and movies? Once you accept you’re in the “content” business, doesn’t any loyalty to those traditional media seem like romantic attachments?
Last April, I wrote about the TikTokification of The White Lotus and how disappointed I was by its rapid cutting and editing, as if it were engineered to keep people from looking away. There’s been a lot written about second-screen content and how streamers are dumbing down shows so that people can have them on in the background. In the case of The White Lotus, the show seemed to have done the opposite, adopting the nature of short-form video to compete with it head-on.
But these stylistic changes, however unpleasant, can only be so effective. That’s why we see Netflix moving decisively to bring in all sorts of content that would have once made little sense on the platform because it risked diluting Netflix’s brand. Live sports, video games, video podcasts, and most recently the kind of short videos, albeit well-produced, that you’d normally find on YouTube.
More ad revenue might not necessarily be bad. After all, some version of this happened before in the past. Network television was a robust business that financially supported the creation and distribution of some great shows, like Breaking Bad, Seinfeld, and Lost. The reintroduction of some of that ad money could go a long way to supporting the artists and artisans that make great stories (that is, assuming we don’t see more industry concentration that enables the corporations to hold onto most of it themselves). But network television also never had to compete with a free product driven by an increasingly powerful programming algorithm that knows just what to play next. When every impression is individually targeted, you no longer need a mass simultaneous audience to command a premium, so the reach advantage of expensive scripted content collapses.
Ad revenue is still a small minority of total streaming revenue, but it all makes me wonder what the end game is here. Do the financial incentives of the ad-based business model put us on an inevitable trajectory of slop carcinization? Will all content providers ultimately look very similar as they converge on short-form video as the optimally engaging/addictive delivery mechanism to maximize ad revenue? Because if the end goal for the streamers is to compete with business models with cheap user-generated content, I think they will have to look and think more and more like those businesses and produce the kind of content they put out. Everything that brainrots must converge.
New releases this weekend: The Samurai and the Prisoner, Kiyoshi Kurosawa’s first crack at the samurai genre. Kurosawa is best known for his slow psychological thrillers like Cure and Pulse, which were major genre inspirations for Obsession’s director, Curry Barker. I’m excited to see how that sensibility translates to feudal Japan. Kurosawa’s Q&As this weekend at IFC are sold out but he’s also doing one for Throne of Blood on Monday.
Director Gregg Araki returns with I Want Your Sex, his first feature in several years, starring Olivia Wilde, Cooper Hoffman, and Charli XCX and co-written by Karley Sciortino. Araki has always occupied a funny and provocative corner of American independent cinema. I got the chance to see this last night and it’s a punchy send-up of the art world and Gen Z sexual anxieties, The Substance crossed with Babygirl (with some references to Sunset Boulevard).
Finally, we have Spider-Man: Brand New Day. People were quick to call an end to uninspired Marvel and Star Wars IP films and usher in the era of Backrooms and Obsession. With this movie tracking for a massive opening weekend and a potential global box office of $2+ billion, that all seems premature. Side character superhero movies may no longer be guaranteed box office hits, but Spider-Man shows that the A-tier characters can still bring people out to the theaters. A huge year for Tom Holland—and Zendaya, who’ll be in Dune 3 as well.
Here are some other screenings I’d encourage you to check out in the coming week:
💎 Friday 7/31 & Saturday 8/1 @ IFC — Thief (1981) — The summer of Michael Mann continues with this screening of his first theatrical feature, which premiered 25 years ago. Thief contained all the hallmarks that would define his career: lonely men, serious professionals, neon-soaked city nights, and a fantastic score. James Caan delivers a career-best performance here, enhanced by his having to actually learn how to crack safes in preparation for the role. I love almost every movie Mann has directed but this is my personal favorite, and one of the few films I own on Blu-ray.
📜 Friday 7/31 @ Intrepid Museum — National Treasure (2004) — I can’t think of a better way to watch this comfort classic than on the flight deck of the Intrepid, where scenes from the movie were filmed. A subsequent jump into the Hudson is strictly optional. It’s a reminder that blockbuster adventure movies are allowed to be unabashedly fun, and a relic from a time when Disney made these large-budget, non-IP adventures.
🏀 Friday 7/31 @ Roxy — Uncut Gems (2019) (35mm) — The Safdie brothers’ anxiety odyssey through New York. With every rewatch, I’m more convinced this is a modern masterpiece. A lot has been written about the crazy kinetic energy of the plot and the characters, but the movie really shines with the detailed attention paid to the most minor figures and the city itself. All of it adds up to a kind of cocaine realism.
This concludes Stay Tuned by Teddy Kim.
Some breaking niche Long Island news: Giunta’s Meat Farms is buying King Kullen.
There’s a new prediction market-focused media company called Eventual. Sara Fischer reported that the goal is “to give the world’s top prediction market traders a platform for news and analysis.” Eventual has 201 subscribers on Substack, and per Axios, its site is “powered by Substack.” It’s backed by LightShed Ventures, Riverside Ventures, Juniper Ventures and BDG Media founder Bryan Goldberg. Prominent traders like Domer, Caleb Davies, Brian Golden, and Paul Krishnamurty will also contribute to Eventual. Polymarket — Eventual’s launch sponsor and data partner — joins a growing group of tech companies like OpenAI and Notion who are experimenting with their own media strategies (OpenAI acquired TBPN, Notion hired New Yorker contributor Adam Iscoe). Earlier this year, Substack formed a partnership with Polymarket that gave publishers more tools to access and share prediction market data.
Don Lemon signed a three-year hosting and advertising deal with podcast hosting platform Libsyn for his podcast and affiliated shows. That explains why his team just posted three roles on the Feed Me Job Board.
L.A.-based artist Sammy Loren embarked on a six-week partying bender with New York’s downtown art scene for the sake of journalism. Stops on his journey included an art show in an abandoned Brooklyn WeWork, a reading with Dimes Square niche micro-celebrities, Mamdani-fueled positivity, and the back room at Time Again. I don’t think I knew they had a back room…
Jennifer Hyman, co-founder and former CEO of Rent the Runway, has been appointed CEO of Babylist. Hyman, who founded clothing rental company Rent the Runway in 2009, left the company earlier this year. She took the company public in 2021. I wasn’t familiar with Babylist until this week, when I bought three baby registry gifts on there for friends. The experience was very easy.
Caribbean restaurant Doubles (which has locations in Sag Harbor and Amagansett) is opening a third location in Brooklyn Heights. Per a 2025 New York Magazine story, Doubles owners Daniel and Evan Bennett are also partners in Bar Oliver and own the yet-to-open Springs General Store.
Abercrombie’s fall campaign was shot at Fanelli.
Is the Midcoast Villager the only local newspaper that also owns a cafe that offers catering?
The pinot noir days in California are over. The NYT reports that California winemakers produced less wine last year than at any point in the last quarter century: “It’s often cheaper to let grapes rot on the vine than to harvest them because so many unsold bottles clog the shelves of storerooms and wine shops.” Vineyard managers are ripping out pinot noir grapes, burning their vineyards, and selling their land. Nationwide, the Times reports, wine sales are at their lowest level in more than two decades, as people drink less and younger drinkers turn from traditional wines to canned cocktails, seltzers, and nonalcoholic beers and spirits. But pinot noir, in particular, has been hit hardest, now that the boom caused by Sideways (it was wine snob Miles’s favorite) has subsided a generation after the indie film goosed consumer demand for some California varietals and spiked the fortunes (“I am NOT drinking any fucking merlot!”) of others.
Andrew Carmellini (Locanda Verde, Cafe Carmellini, Bar Primi) is opening a new restaurant in The Public Theater called The Astor. It’s billing itself as, “ a downtown tavern for the public, where dining, drinking, and performance come together under one roof.” Serge Becker (who designed Miss Lily’s, The Box, and La Esquina) designed the space.
Let’s talk about Leopold. Great reporting by the FT last night. Leopold Aschenbrenner burst onto the scene as a wunderkind stock investor over the last two years, betting on just one megatrend: AI and demand for compute.
A quick primer on his rise and (maybe) fall: At just 23, he left OpenAI’s “SuperAlignment” team, raised a fund called Situational Awareness, generated eye-watering returns, and grew assets under management to $20 billion as of June. Then this morning, CNBC reported that Leo’s dramatic recent losses have wiped out all his YTD gains, and the fund has liquidated all of its public equity holdings.
People have a lot of feelings about this guy. I’ll try to identify how Leo’s shooters would describe him as an investor, and how his haters would describe him. Recent events have proven his haters correct, but let’s try to understand what investors admired in this guy, and why he was such a big deal over the last year.
To his fans, Leo stood out as someone who not only understood AI megatrends, but also deftly traded them in markets. The word was that this kid could read the tea leaves of AI, and was beating the Wall Street veterans at their own game. Like any iconic trader, Leo has an unforgettable signature look, like a foppish elven prince in a movie where elves are good at thinking, but not fighting. Aesthetically, he was an anti Sam Bankman-Fried - whereas SBF survived on cargo shorts and ramen, Leo looked like he might have deadly allergies to both. He also has lore - his deep personal connections to the early teams at OpenAI and Anthropic (and FTX!) supposedly give him an edge as an investor. He also loves Diet Dr. Pepper.
The Leo legend probably reached its peak in a June 8 Wall Street Journal profile. The profile was full of reverential, almost religious, quotes and praise for a whiz-kid investor with a short track record. Go on x.com and look at people’s tweets from early June praising him as a once-in-a-generation investor (some are now deleted). Graybeards will often tell you this type of glowing profile is a perfect signal of a top in the market.
Now for the haters. As Situational Awareness racked up dizzyingly good returns for the first half of 2026, skeptics muttered that Leo might be using extreme, unsafe amounts of leverage (borrowed money) to generate those returns. Some people even claimed he was basically a figurehead for a group of industry insiders who wanted to put on this levered AI trade. “It’s not alpha, it’s levered beta,” said the haters, but the shooters pointed out that Leo’s returns and specific stock picks beat those of the semiconductor indexes, and also beat their estimate of what a levered index bet might look like (always a squishy estimate). Just wait, said the haters: If he’s taking too much risk, eventually his returns will go down as quickly as they went up.
It seems that time has come. With the unwind of the “bottleneck trade” over the last 8 weeks, Leo’s core market of chipmaker, data center, and AI-exposed stocks has traded down sharply. Some of the hardest-hit groups have been Korean retail investors, and investors in Situational Awareness.
As news of the fund’s struggles broke last night, haters claimed he was down 30% year-to-date (YTD), while others claimed his returns and realized gains before June were too good for that, and he’s still up YTD. The FT article is very good, but they did not share his full returns info, so ultimately a lot of these people were just guessing. This morning, breaking reporting from CNBC suggests that Leo’s recent losses have indeed wiped out all his YTD gains, and the fund has exited its entire public equities book. (WSJ is now reporting that Citadel bought all his positions in a single block trade.)
Meanwhile, as recently as last week, Leo was out raising fresh funds, and calling this downturn an excellent buying opportunity. Some smart people agree, and say that behind the scenes, the AI labs and demand for compute are doing nothing but accelerating. But are the bulls just talking their (hurting) books? Time will tell.
Something I learned this summer is that king kullen is considered the first true supermarket (I had to do some googling to get a better idea of what that meant)
I just started a rewatch of White Lotus--Season One––and Teddy's right once again. The scenes are not rushed and attention is rewarded. I think Season Two had that same aspect.
Have not yet rewatched Uncut Gems. I loved it, but recalling my first watch, I'm afraid it will raise my BP. That would be a great experiment––monitoring the pulse and BP of an audience watching Uncut Gems.