OpenAI推GPT-5.5/5.6，政府强限预览
OpenAI发布Sol（GPT-5.5同级性能同价）和Terra（5.5水平半价），但应美国政府要求改为有限预览而非原计划的开放发布。Sam Altman称这符合迭代部署策略，但承认并非完全自愿。
Sam Altman主动把政府干预包装成"迭代部署"，但措辞里"at the request"和"this isn't quite"暴露了张力。这是首次有大模型发布被政府直接踩刹车，不是建议而是要求。
定价策略也很微妙：Sol同价对标、Terra半价下沉，明显是在抢企业客户预算。但有限预览让实际竞争力打折扣——企业做技术选型时，"可用性"权重正在上升。
做AI infra和合规工具的创业者注意：模型发布的不确定性正在从纯技术风险变成政策风险，帮企业做多模型备案和切换方案的服务需求会涨。
RT @PeteButtigieg: This week someone targeted my family for harm with a false report. We’re physically OK, but that doesn’t mean we weren’t…
查看原文 →Good new first: Sol is a smart, efficient, and a significant step forward. It is the same price as GPT-5.5. Also launching in the GPT-5.6 family is Terra, with 5.5-level performance at half the price. Bad news: at the request of the US government, it is launching today in limited preview instead of the open access launch we were planning on. We are working with the government to get to general availability as fast as we can. I think it is quite reasonable to roll out models--especially as they reach significant new levels of capability--in this way. It fits with our long-held strategy of iterative deployment. But this isn't quite the process that we think is optimal. Now we will with the government to attempt to get to a transparent, reliable process for early access, and to ensure that as long as our safeguards work as intended we can release widely. We want to be a reliable, dependable partner that works with all stakeholders, and we also want to live by our mission of benefiting all of humanity. I believe the government shares most of our goals, and that they are overall doing a good job in a very difficult situation. We will work as quickly as we can to get this model in your hands and we hope you will love it.
查看原文 →RT @paulg: Step 1: Invent a new extra-bad category of crime, like "hate crime" or "terrorism", with extra-harsh punishments. Step 2: Exp…
查看原文 →in other news, we updated the 5.5 instant model used in chatgpt this week. i like its vibes.
查看原文 →RT @lionel_trolling: Yeah it's a real mystery why Vance would identify with an unlikable guy whose entire life revolved around resenting an…
查看原文 →RT @oziadias: My friend the brilliant @kevin_volpp dropped dead He survived (read abt it: https://t.co/UdybYJbUBr) But every year 300-400…
查看原文 →RT @FoundationOAI: The OpenAI Foundation is joining Intercept as a founding partner through our AI resilience program. As AI accelerates a…
查看原文 →RT @KenRoth: The decades-long prison sentences imposed on a group of Texas activists who were protesting against ICE, using the guise of "t…
查看原文 →RT @photoroom_ML: 🚀 Meet PRX Pixel. Our new open-source 7B text-to-image model that generates images directly in pixel space. After months…
查看原文 →RT @KenRoth: The Trump administration claims that mass deportations will free millions of jobs for Americans by reducing foreigner competit…
查看原文 →RT @Dan_Jeffries1: It doesn't take much skill to game this out and see where we're going if we keep this up. Kneecap our own companies. Ga…
查看原文 →This is all open source, and it's probably the direction we'll be automating a ton of tasks in the future: 1. Record yourself using a web browser doing something you want to automate. 2. The system takes the recording and cleans it by removing retries, dead ends, and anything that's specific to one page and not reusable. 3. Then, it turns the clean process into a skill by only keeping the logic to accomplish the task, not the specific clicks. 4. Finally, it organizes the skill into a "skill graph", so an agent can retrieve it to work on a new, related task. And this is now open source!
查看原文 →RT @ClementDelangue: It's quite rational to regulate frontier API models, especially to get more transparency for the government, without r…
查看原文 →We are in skill hell. We already need skills to discover other skills. 2020: Yet Another JavaScript Library. 2026: Yet Another Skill.
查看原文 →