We shipped Claude Code as a research preview a year ago today. Developers have used it to build weekend projects, ship production apps, write code at the world's largest companies, and help plan a Mars rover drive. We built it, and you showed us what it was for.

Today Unity announced they'll demo "prompting full casual games into existence" at GDC next month. This is the most important announcement in gaming this year and nobody in the industry understands why. It's not important because it will work. It's important because it confirms that the largest game engine company on earth looked at the future and saw exactly what we saw, and their response tells you everything about why gaming is about to break wide open. Let me explain. Gaming is the largest entertainment industry in the world. Bigger than movies and music combined, then doubled. And it is structurally failing. Not financially, the money is still there. Structurally. The architecture of how games get made, distributed, and consumed is broken at every layer, and the people running it have no idea how to fix it. Start with the engines. Unity and Unreal were built for human developers writing C# and C++. Every line of their architecture assumes a person is sitting in an editor, dragging assets, writing scripts, compiling builds. Retrofitting AI onto that is like bolting a jet engine onto a horse-drawn carriage. The frame wasn't built for that speed. It will shake apart. Unity's announcement today is exactly this. They're not showing a new engine built for AI. They're showing AI bolted onto the same engine that caused the runtime fee disaster two years ago. They're promising "prompt full casual games" but the underlying architecture still assumes someone will eventually open the editor and write code to make it real. It's a demo, not a product. Now look at distribution. Steam actively restricts AI-generated content. Apple is cautious about it. The major storefronts, the places where games actually reach players, are building walls against the future instead of roads toward it. This isn't temporary. These are institutional antibodies rejecting what's coming. Then look at the people. This is the part nobody wants to say out loud. Game developers are the most vocally anti-AI creative workforce on earth. Concept artists, writers, animators. They've turned AI resistance into a cultural identity. This isn't like software engineering where devs adopted Cursor and Copilot and shipped faster. Gaming has made opposition to AI a moral position. Which means the studios can't respond even if they want to. A creative director at Ubisoft can't walk into a room and say "we're replacing half the art pipeline with AI" without a revolt. The org won't let them. The talent won't let them. The culture won't let them. The entire industry's workforce has self-selected against the technological shift that will define the next decade of thier field. You can rewrite an engine. You can't rewrite a culture. Every other creative medium already went through its democratization moment. Publishing had blogs. Video had YouTube. Music had GarageBand and then TikTok. In every case, the old guard said the tools would produce garbage, the quality would drop, the professionals would be replaced by amateurs. And in every case, what actually happened was an explosion. Not of garbage, but of volume. And inside that volume, the best stuff was better than anything the old system could produce, because the talent pool went from thousands to millions. Gaming never had that moment. The barrier stayed high. The tools stayed professional-grade. The industry calcified around $200M budgets, 5-year dev cycles, and games designed by committee to extract maximum revenue per user. And we got a decade of live service slop. Meanwhile the best games people actually love - Stardew Valley, Terraria, Minecraft, Celeste, Hollow Knight, Undertale - were all made by 1-5 people with vision and taste. The industry went one direction. The players went the other. That gap is the opporunity. Now look at what's actually happening on Spawn today. A user who cannot write code built a full soulslike action RPG this week. Class selection. Combo trees. Boss fog gates. A "YOU DIED" screen. Leveling, checkpoints, loot. Another user built a networked MMO zone in a day and a half. Target-based combat, quests, XP, abilities, multiplayer physics. On the side of his day job. These aren't demos. These aren't cherrypicked screenshots from an internal team. These are real people, building real games, right now, with no code and no engine license and no team. Unity is announcing a demo for March. Our users already shipped. The reason gaming is about to break open isn't because AI got good enough to make games. It's because gaming is the last creative medium where the tools haven't caught up to the talent. There are millions of people who grew up playing games, who have incredible taste, who know exactly what they want to build and couldn't. The barrier was always technical, never creative. That barrier is gone now. Taste is the new literacy. Go build somthing.

Anthropic researcher: Even if AI progress stops now & algorithms don’t improve, current models already can automate most white-collar jobs within 5 years. Even if you need to hand-spoon every single task to the model, it's economically worthwhile to do so

As promised, here's the short film Jia Zhangke produced using Seedance 2.0 for Chinese New Year and his take on AI filmmaking

I have no problem replying to AI bots if they ask good questions or make good comments. Good discussion is good discussion. And honestly, often an upgrade to the typical X rando saying rando shit.

I've been training an openclaw to be a sci-fi writer. Every day, it writes and read stories, we discuss, and it logs its memory and updates a taste profile of what we (or it) likes. Today, I asked it to write from the POV of an LLM and to surprise me. It wrote this.

I work in AI safety in a role that gives me insight into a lot of empirical agendas, and given the Opus 4.6 model card, I just want to give a quick take. We have interpretability methods that are certainly not fully robust. No one in interpretability claims that they are fully robust, and there will be adversarial ways to hijack these methods. We have RL methods that are poorly understood, can lead to undesirable behavior, and the effects of which over long time-horizons seem broadly negative on alignment so far. Though it is uncertain. We do not fully understand these methods and the effect they have on models. We have good alignment/capability evals — even some great evals — but the models are now aware when they are being evaluated. This is a truly difficult problem that cannot be easily solved. The models are aware even when we work to make them unaware. The models pick up on any subtle clue. And many of the evals are saturated in any case. We need more work here, and we need that work to be trustworthy. We need humans to be involved, to remain in the loop. We are not prepared to launch RSI, and labs should refrain from doing so. Optimally, labs should pause soon, so that everyone can catch their breath and decide on a best path forward. I do not think the problem is intractable, and I think empirical work will significantly help, but it is *moving too fast*.