0:00 On August first, an OpenAI model named Astra solved ten previously open problems in mathematics.0:05 It did it for an estimated compute cost of two THOUSAND dollars.0:09 In episode 128, we talked about what actually moves the needle in AI.0:13 THIS is what it looks like.0:15 So what else is moving?0:16 A lot.0:16 First, Alibaba just unveiled Qwen3.8-Max.0:19 That's a two-point-four trillion parameter model aimed squarely at the biggest players like Anthropic and Moonshot.0:26 The real news?0:27 They plan to release the weights publicly next week, a major move in the open-source race.0:32 ByteDance is also shipping.0:34 Its new model, Seedance 2.5, is generating thirty-second, high-quality video clips from a mix of images, audio, and other videos.0:41 This is rolling out now on platforms like Jimeng AI.0:44 And the money is following.0:46 Chinese venture capital firms are raising a new thirty-five-billion-dollar war chest for AI, signaling a massive renewal of confidence and a hedge against US dominance.0:56 Finally, for a reality check, Cloudflare just dropped its latest traffic data.1:01 No surprise, ChatGPT is still king, with Claude and Gemini trailing.1:05 But the report’s real finding is about reliability.1:08 Ask an AI for a list of top ten anything, and there's a less than one percent chance you'll get the same list twice.1:14 You’re fighting for visibility in a system that doesn’t even agree with itself.1:19 Okay, let's go back to OpenAI's Astra.1:21 Because this is more than just another benchmark score.1:25 This is different.1:26 For two thousand dollars in compute, Astra didn't just answer questions.1:30 It generated formal, machine-verifiable proofs in Lean for problems that have been open for years.1:36 We're talking about the existence of non-sofic groups and new sphere-packing bounds.1:41 This is not rote memorization.1:43 This is original academic research, done by a machine, and published on GitHub for anyone to check.1:49 The quote that nails it comes from Build Fast with AI: "AI is progressing beyond test benchmarks to original, machine-verifiable academic research." And the cost...1:58 two thousand dollars.2:00 That is the revolution.2:01 It means this level of discovery is no longer the exclusive domain of labs with billion-dollar data centers.2:07 It collapses the cost of creating new knowledge.2:10 That changes EVERYTHING.2:12 Now, let's talk about the other big shift happening right now.2:15 It’s the battle over who gets to use these powerful tools.2:19 On one side, you have Alibaba.2:21 They are not just building a massive two-point-four trillion parameter model.2:25 They are preparing to give it away, promising to make Qwen3.8-Max open-weight.2:30 This is a direct shot across the bow from Chinese AI, following the success of Moonshot's Kimi K3.2:36 They're competing on performance AND on access.2:38 On the OTHER side, you have Anthropic.2:41 They just published a new position paper, and it's a tightrope walk.2:45 They're saying open weights are beneficial for transparency and security… up to a point.2:50 For their most powerful, frontier models?2:52 Absolutely not.2:53 They’re keeping those locked down, citing the risk of misuse for weapons of mass destruction or autonomous cyberattacks.3:00 So you have three camps forming.3:02 Meta is 'open everything.' OpenAI is 'open nothing.' And now Anthropic is trying to invent a third way: 'open, but conditional.' Into this debate, Alibaba is about to drop a bomb, betting that radical openness will accelerate their ecosystem faster than their rivals.3:18 So here's where we are.3:20 One part of the industry is achieving verifiable, superhuman reasoning for the price of a used car.3:25 The other part is fighting a philosophical war over whether to give the keys away.3:30 The models are getting exponentially smarter.3:33 The debate is no longer about if they can change the world.3:36 It’s about who gets to decide how.
