Amp(在新窗口中打开)正在构建一款前沿 AI 智能体，帮助开发者将雄心勃勃的想法转化为可运行的软件。对于 Amp 首席执行官 Quinn Slack 而言，模型的每一次改进都是扩展智能体能力的机会，从帮助开发者决定构建什么，到编写代码和设计页面。
Amp 根据对用户最有效的表现来选择并推荐模型，而 OpenAI 模型占 Amp 中使用的 token 的大多数。Quinn 分享道：“OpenAI 非常透明，当有保障措施被应用时，用户知道并且理解。这让使用 OpenAI 模型变得非常容易。你可以在 Amp 中用 OpenAI 模型完成所有工作，并且非常满意。”
“我们的用户非常欣赏 OpenAI 模型变得越来越好的一致性。”
——Quinn Slack，Amp 首席执行官兼技术员工成员
通过 Sign in with ChatGPT 降低构建成本
随着构建者给智能体更多工作，token 成本可能成为限制。Quinn 描述了一些 Amp 用户每周消耗 100 亿个 token。他认为，如果人们关注每一次使用量的增加，可能会阻碍他们发现智能体的能力。
为了帮助缓解这一点，Amp 推出了 Sign in with ChatGPT 集成。Quinn 说，这使 Amp 用户能够访问更多 token 和 OpenAI 模型，包括 GPT‑6 Astra，这改变了他们让智能体承担更大任务的意愿。
“Sign in with ChatGPT 意味着人们可以拥有那种不必计算每一个 token 的感觉，”他说。“而一旦你真正让智能体放手去做，给它更长的牵引绳，你就会意识到它能做多少事情。”Quinn 说，他看到人们使用智能体来规划、决定构建什么，并在发布后继续工作。
使用 GPT‑6 Astra 进行写作、设计和创作
Amp 希望智能体能够帮助完成构建和发展产品所涉及的所有工作。它自己的 20 人团队就是这样使用它们的，包括在营销方面。“我们没有专门负责这件事的人，但我们有一个专门负责这件事的智能体，”Quinn 说。团队依靠智能体来改善新用户的体验，并帮助更多人通过搜索和其他渠道发现 Amp。
对于这项工作，拥有高质量的写作和设计至关重要。GPT‑6 Astra 帮助团队制作符合 Amp 严格标准的文档和设计页面。
“我们有一种非常独特的设计风格和写作风格。我们对此非常挑剔，但我们对 GPT-6 Astra 的表现非常满意。”
——Quinn Slack，Amp 首席执行官兼技术员工成员
Quinn 对 GPT‑6 Astra 的实验还在家里继续。他把自己房子的平面图交给它，并用它在 Blender 中创建了一个 3D 模型。然后他和孩子们把这个模型变成游戏：在房子里加一条漂流河，或者围着它追大蒜，灵感来自他们的祖母不喜欢大蒜。“而这在几个月前是不可能的，”他说。
通过 GPT‑Live‑1 与你的智能体交谈
Amp 最近宣布了 Puck，这是一个内置于 Amp 的对话式 AI 助手，帮助用户启动和管理其他智能体。用户可以要求 Puck 启动一个智能体来调查 bug、协调多个智能体，或查找之前关于技术问题的对话。
通过由 GPT‑Live‑1 提供支持的 Calling Puck，用户可以与 Puck 实时交谈，包括从手表上。Quinn 设想人们用它来让智能体在他们离开电脑时继续工作，甚至在外出跑步时也是如此。
使用 Ultrafast 加速智能体工作流
Amp 一直在用 GPT‑6 Astra 测试 Ultrafast 模式，Quinn 说更高的速度改变了人们使用智能体的方式。如今，人们常常在等待每个智能体任务完成时在多个任务之间切换。更快的响应可以让人们同时处理更少的智能体，或者给智能体更多时间在返回结果前检查自己的工作。
“人们还没有见过一个这么好、这么快的模型。”
——Quinn Slack，Amp 首席执行官兼技术员工成员
对于 Amp 来说，下一步是将更多编码前后的工作带入智能体，这样构建者就可以把一个想法推进得更远，而不必自己处理每一步。“人类将创造出什么，这是一个未知数，而这就是我们都能看到的。能参与其中很酷，”Quinn 说。
Amp(opens in a new window) is building a frontier AI agent that helps developers turn ambitious ideas into working software. For Quinn Slack, Amp’s CEO, each improvement in the models is an opportunity to expand what the agent can do, from helping developers decide what to build to writing code and designing pages.
Amp chooses and recommends models based on what works best for its users, and OpenAI models account for the majority of tokens used in Amp. Quinn shares, “OpenAI is very transparent, and when there are safeguards applied, users know and they understand. That makes it really easy to use OpenAI models. You can do all of your work in Amp with OpenAI models and be really happy.”
“Our users really appreciate the consistency with which OpenAI models have been getting better and better.”
—Quinn Slack, CEO and member of technical staff, Amp
Lowering the cost of building with Sign in with ChatGPT
As builders give agents more work, token costs can become a constraint. Quinn describes some Amp users consuming 10 billion tokens a week. He believes that if people watch every increment of usage, it might hold them back from finding out what an agent is capable of.
To help alleviate that, Amp launched its Sign in with ChatGPT integration. Quinn says it gives Amp users access to more tokens and OpenAI models, including GPT‑6 Astra, which changes how willing they are to let an agent take on a larger task.
“Sign in with ChatGPT means that people can have that feeling where they’re not counting every single token,” he says. “And once you actually let the agent go and have a longer leash, then you realize just how much it can do.” Quinn says he sees people using the agent to plan, decide what to build, and continue working after they ship.
Writing, designing, and creating with GPT‑6 Astra
Amp wants agents to help with all the work that goes into building and growing a product. Its own 20-person team uses them that way, including in marketing. “We don’t have someone devoted to it, but we have an agent devoted to it,” Quinn says. The team relies on the agent to improve the experience for new users and help more people discover Amp through search and other channels.
For that work, having high-quality writing and design is critical. GPT‑6 Astra helps the team produce documentation and design pages that meet Amp’s exacting standards.
“We have a very unique design style and writing style. We’re very picky about this, but we’ve been really happy with how GPT-6 Astra does.”
—Quinn Slack, CEO and member of technical staff, Amp
Quinn’s experiments with GPT‑6 Astra continue at home. He gave it the plans for his house and used it to create a 3D model in Blender. He and his children then turn that model into games: adding a lazy river through the house or chasing garlic around it, inspired by their grandmother’s dislike of garlic. “And that was not possible just a couple months ago,” he says.
Talking to your agent through GPT‑Live‑1
Amp recently announced Puck, a conversational AI assistant built into Amp that helps users launch and manage other agents. Users can ask Puck to start an agent to investigate a bug, coordinate several agents, or find a previous conversation about a technical problem.
With Calling Puck, powered by GPT‑Live‑1, users can talk to Puck in real time, including from a watch. Quinn imagines people using it to keep their agents working while they’re away from their computers, even while out for a run.
Accelerating agent workflows with Ultrafast
Amp has been testing Ultrafast mode with GPT‑6 Astra, and Quinn says the greater speed changes how people use agents. Today, people often switch between several agent tasks while waiting for each to finish. Faster responses could let people juggle fewer agents, or give agents more time to check their own work before returning a result.
“People have not seen a model that’s this good, that’s this fast.”
—Quinn Slack, CEO and member of technical staff, Amp
For Amp, the next step is to bring more of the work before and after coding into the agent, so builders can carry an idea further without handling every step themselves. “There’s this unknown of what humans will create, and that’s what we all get to see. It’s cool to be playing a part in that,” Quinn says.
本文内容采集自官方网站，排版和翻译可能与原页面存在差异。
阅读官方全文