Today, we're co-hosting the largest AI film festival in the world. $2 million in prizes and investments is on the line. Two dozen finalists of the @AstanaAIFF will screen their films at the cinema near Network School's campus. We received over 8,000 submissions from filmmakers from 125 countries. The theme was "The Future Worth Living In" and the films were really fun to sift through alongside other @ns members. A decade ago, I interned in Hollywood. My job was to read scripts and tell my bosses which ones were good and which ones were bad. Even the good ones usually didn't get made. Studios are loss-averse. They optimize for films that will return a multimillion-dollar bet, like sequels and Marvel movies. Ten years later, we're in a very different moment. AI video models and tools are ushering in a new era of storytelling that makes original films far less expensive to produce. From a budget perspective, all you need to create the next Inception or Parasite is a laptop, some credits, and a good story. Nobody needs to greenlight your film anymore. There's no studio to pitch, no executive to win over, no intern like me writing "pass" at the top of the coverage. But you still need gumption. A human will absolutely still have to will the film into existence. AI alone won't make the film for you. It can render a shot, but it can't decide what the story is about or why anyone should care. That part is still yours, and every finalist tomorrow made hundreds of decisions to make their stories happen. To the skeptics: watch what AI video looked like two years ago. A few seconds of melting faces and extra fingers. Today, filmmakers hold characters consistent across scenes, direct the camera, and cut complete stories with real emotional arcs. That's why the theme matters. The tools are going to get built either way. The question is what stories we tell with them, and whether those stories help us imagine a future worth living in. This weekend, creators from around the world will show us their answer. Here's a trailer with clips from the submissions:

NS Kazakhstan is open! Excited to build a bridge between our international technology community and KZ’s many talented engineers, dedicated researchers, and friendly people.

I ran an AI-powered studio in 2024. Back then, generating a clip took 2–10 minutes and looked broken. Now, the best shots look real and earlier this month, MiniMax generated 5 seconds of video in 2.53 seconds. Cost per minute also fell by an order of magnitude: traditional video is still $1,000–$50,000; AI is $0.50–$30, depending on # of iterations. The tool layer took off too. Higgsfield went from $20M to $700M ARR in a year. Kling is at ~$500M. Runway ~$300M. Consumption is following. A handful of AI-only YouTube channels combined for 63B views by late 2025. And in Q1, the AI microdramas category hit 718M monthly viewers at 129 min/day. Hollywood-scale for AI video is coming, but the story still has to be human. Excited to support the next Christopher Nolans and Greta Gerwigs with millions in grants and investment.

Here it is from the outside

People > place

network school boyfriend, edge city girlfriend

College students living near stairwells make the most friends! Shared interests, values, personality? All downstream of foot traffic Friendship formation can be sorted into 3 conditions 1) proximity 2) repeated unplanned interactions 3) settings where people lower their guard

Insanity - especially contrasted with Singapore, where you just scan your passport in a machine upon ingress and egress. Takes 10 seconds. American customs is already bad but Europe's EES rules are going to take the crown for worst travel bureaucracy now.

After playing with @GoogleDeepMind 's newest video model Omni, the best prompts define a video's CLASS - Composition - Lighting - Action - Style - Setting The model really excels at video-to-video, especially when including these five elements in your prompt.

📈

At @ns, we're doing the work to build that confidence. Each month, a new cohort of 100+ people come from around the world to our island campus. Like clockwork, a mix of NS Core and longtermers help folks from every culture move in, find a rhythm, and level up (physically, professionally, or otherwise) together. This is a giant effort by many people. And while there is still much to improve (and I am as self-critical as they come), I do look around with increasing confidence at the beginning of each month, knowing that we continue to iterate and get better with each rep.

Tonight was one of the best examples I’ve seen of the @ns community coming together. This event was run by the community - they invited the guests, created the agenda, and transformed the café.

The new white collar vs. blue collar: Are you telling the AI what to do, or is the AI telling you what to do?

One of my favorite @DwyaneWade quotes: "Real confidence comes from doing the work." After Game 4 of the 2006 Finals, Wade went to the gym at 3 AM just to shoot free throws. He told his trainer the title would come down to those shots. He was right. He hit two free throws with 1.9 seconds left in Game 5 to take a one point lead in the game and a 3-2 lead in the series. Later, he explained why he wasn't nervous: "You've done it over and over. You've seen yourself do it." He watched the ball go in the net so many times that he knew he wouldn't miss. Most people try to "visualize" success as a form of wishful thinking. For DWade, visualization was a memory. True confidence is the calm that comes from having a mental library of successful reps to fall back on.