What Tobi is about to do today is one of the most insane things I’ve seen in athletic performance. He will be racing for 2.5 hours straight. Before immersing myself in this sport for the past few days, I would have thought “oh that’s cool, he’ll be driving a fast car around a track.” That is not what this is. And why I have so much respect for him. Tobi will be driving a car over 180 mph, experiencing up to 4.2 g of force. That will throw his body forward with 715 lbs of force and his head by 67 lbs. Braking action requires up to 200 lbs of leg force. The cockpit will average 104-122°F (40 to 50°C). His multi-layer suit barely allows for evaporation. This will cause his core body temperature to reach 102°F. I achieved this same body temp after 200°F of dry sauna for 34 min. It’s overwhelming. His heart rate will beat between 140-170 bpm. He will sweat up to 170 fl oz (5 kg), losing 11 lbs of body weight. Some classes run driver cooling systems. His LMP2 doesn't. Under these conditions, he will need to maintain perfect concentration for the entire 2.5 hours. He’s fighting for tenths of a second on every turn and every stretch. One small slip up and trouble awaits. Today it’s even more intense… it’s raining. The track is wet, scrambling everything. He has to maintain spatial memory of every other car on the track, the complicated rules to avoid penalty, and constantly strategize overtaking other drivers, communicate with his team on the radio, watch for debris, be aware of slightest nuances of his car, and anticipate other drivers. He did a practice run on Thursday and then a qualifying race yesterday, when he emerged from the car, he was a different human. Unrecognizable from the person who entered the car. Beaming with something like an omnipresence of reality. We celebrate those who are especially good at kicking balls, shooting balls into baskets, and running with balls. What Tobi will do today is alien to those. I struggle to think of a sport equally as demanding. Tobi and I have a lot in common. We’ve both built companies. We both spend a lot of time at the keyboard trying to organize the world in ways that we find beautiful. We both love data, method, and analysis. We both have a love for continuous improvement and will endure any pain required to achieve that next goal. What I’ve not experienced is the unmatched flow state Tobi tells me he drops into when under these racing conditions. I deeply understand what it means to achieve a 102°F core body temperature after 34 min in a 200°F dry sauna. I also know what 3 g’s feels like from being a pilot. These are extremely intense psychological states. I know what it means to apply 200 lbs of pressure with my legs. What I’ve never experienced is all these things, at the same time, for 2.5 hours straight, while maintaining perfect focus. He is new to car racing. Late by racing standards. DHH got him into it a few years ago. During this first year of racing, he endured to overcome the extreme motion sickness that accompanies the physical and mental violence of the sport. The race starts at noon eastern. Tobi will be driving for the first 2.5 hours and then @dhh will drop in, followed by Mathias Beche. The race will conclude at 10 pm. I’ll be in the pit watching as all the data streams in from his car and body and imagining what it’s like to be in that car.

I have yet to see any sort of happy ending write ups on a post AGI future. Like what, best case is that centaur period ends, we have our aligned AIs, and we all sit around all day doing what? We can’t all open up cute little French bakeries off UBI until we die at like 150

We did it, the first Halo CE match ever played with more than 16 players happened today at 2:07 CST

new preprint: language models are “insecure” reporters. researchers and engineers have become increasingly reliant on LLM-written reports---second-hand accounts of code and experiments---to understand agent-generated work. in this study, we ask: “how transparent are these LLM-generated reports?” 🔍

On this topic, I just stumbled upon an extraordinary number released by China's National Radio and Television Administration (NRTA): 430,000 micro-dramas (short TV series) went online in China in the first 8 months of 2026, and over 90% of them were AI-made (news.ifeng.com/c/8wUTuZWFeeU) 🤯 And micro-dramas aren't a micro-phenomenon in China: still according to the NRTA, they're watched by 800 million people - a good majority of the population - and the market passed 100 billion yuan last year, about twice China's cinema box office (chinafilm.gov.cn/xwzx/gzdt/2…). Which means that, that's it, we're there: the majority of new TV-style series in the world's largest market are made by AI. I knew anecdotally that AI entertainment content was big in China (hence my casual take last night 👇) but I'm genuinely awed by the scale and speed of it: AI dramas were virtually non-existent at the beginning of last year and, fast forward a year, it's now human-made dramas that are the exception! To be clear, this is about what gets made: many of these AI dramas are - like YouTube videos - made by everyday folks and don't that much viewership. But that's how every creator economy works: a long tail few people watch, and a handful of hits that reach hundreds of millions. This might possibly become China's answer to Hollywood. We're talking immense production cost advantages: the cost per minute of footage for AI micro-dramas is roughly 1,000 yuan ($140, yicai.com/news/103084978.htm…) while your typical Netflix drama costs $5M to $10M per hour-long episode, so $85,000 to $170,000 a minute. In other words, it's roughly a thousand times cheaper to produce. And there are other advantages too: with AI, it's ridiculously easy to re-render the footage with local faces, languages and other local cultural characteristics for every single international market at near-zero cost. Which would be something Hollywood never had: a culture-neutral fiction factory. All the more because we're also starting to see full-length AI dramas appear. In fact, just 3 weeks ago, Hunan TV, one of China's biggest channels, began airing in prime time "Hou Xiyou Ji" ("The Later Journey to the West"): a 30-episode series of 40 minutes each, entirely AI-generated, without a single human actor. You can watch the first episode here in full: youtube.com/watch?v=T2Xzd7KH… and see for yourself that it looks remarkably well made. It went from approval to broadcast in just under 4 months and, according to the series' director Li Dongshen, costs about 900,000 yuan per episode to make ($135,000), or 50 times less than your typical Netflix series (bjnews.com.cn/detail/1788401…). And it did very well: it topped its time slot among provincial satellite channels five nights running, passed 150 million plays on Mango TV in its first five days, and sent Mango Excellent Media's stock up 44% in two days, adding over 11 billion yuan to its market value (m.21jingji.com/article/20260…). Bear in mind that at the beginning of 2025 the technology could not do this at all (back then, no-one could produce more than a few seconds of coherent AI video, let alone a 40-minute episode with consistent characters) so it's an understatement to say that things are moving fast. Imagine where this will be in another 18 months: today is the worst AI dramas will ever look. Interestingly, this technology is overwhelmingly Chinese. Where China is roughly on par vs the West for frontier LLMs, it is clearly ahead in AI video: on the main blind-test leaderboards (artificialanalysis.ai/video/…), out of the 10 top spots, 9 are held by Chinese models (ByteDance's Seedance, Alibaba's Wan, Sand.ai's MAGI-2, Minimax), with Google the only Western lab still in the race. And they've clearly unlocked some sort of flywheel which gives them a considerable advantage: the companies building the models are the same ones that own the platforms where the content gets watched. ByteDance makes Seedance and owns Douyin and Hongguo, China's biggest micro-drama app. Alibaba makes Wan and owns Youku. Kuaishou makes Kling and owns, well, Kuaishou. So hundreds of thousands of AI dramas get made with their models, then watched by 800 million people on their app. Better models make better dramas, better dramas bring more viewers and more creators, and round it goes. No Western lab has anything like it: OpenAI had a model - Sora, being discontinued this very week (help.openai.com/en/articles/…) - but no audience, and Netflix has an audience but no model. Only Google has Youtube but, so far, it strangely doesn't seem to have connected the pieces: it has the best video model in the West and the biggest video audience in the world, but no serialized drama format built on top. If anyone in the West can replicate the Chinese flywheel, it's them. But for now, they haven't. In any case, long story short, what this all means is that the entertainment industry, in Hollywood and elsewhere, is on the cusp of being seriously disrupted, and its epicenter is China. Humans won't disappear from the process - even in China, the regulator's line is that "AI is a tool, humans are the creators" (china.org.cn/2026-09/20/cont…) - but it matters a great deal which humans, and whose tools, and it looks increasingly like the next generation of storytellers will either be Chinese or using Chinese tools.

This actually left out the most unique thing I did. (Each item on Andreas's list was done by at least one other person besides myself.) Starting in 2004, I concluded that Friendly AI would be too hard/risky to attempt, and repeatedly tried to talk Eliezer/MIRI out of pursuing this path to the Singularity (before they changed their minds largely by themselves in the early 2020s). AFAIK I was the only person to publicly make this argument during that time, which seems striking and notable, e.g., as evidence for how rational or strategically competent the rationalists and humans in general were/are. Not that the argument was clearly correct back in the 00s and 10s, but rather in retrospect it seems like a natural argument for people in the rationalist circles to make, and surprising that nobody else did.

People high in ADHD traits have an underappreciated advantage. Curiosity.