Aggregators pull headlines from many publishers and put them in one place. They are very good at breadth: if you want to know what is being published right now, and then read the original article, that is what they are for.

What they hand you is still a list of headlines. Reconstructing one story means opening several of them and holding the pieces together yourself.

Use an aggregator when you want coverage — many outlets, many stories, the original articles.

Use gg2 when you have already seen the headline and want the story: the background, the sequence, the disagreement, and what is not yet established.

Ground News is built around how a story is covered: which outlets are reporting it, where they sit on the political spectrum, and what one side is not covering at all. If media bias is the thing you want to see, it is purpose-built for that and gg2 is not a substitute.

gg2 is aimed at the story rather than the coverage of it. A card does note where sides disagree, but its centre of gravity is the event: what happened and why it matters.

Use Ground News when your question is “who is reporting this, and how differently?”

Use gg2 when your question is “what actually happened, and why does it matter?”

A human-edited daily digest — 1440, Morning Brew, Axios and the rest — arrives once, has an editor deciding what matters, and has a voice. Those are real advantages, and gg2 has none of them: it is written by AI.

The trade-offs run the other way too. A newsletter is a fixed edition: it lands in the morning and does not update when a story moves, it covers what fits the format, and it rarely shows you the reporting underneath. gg2 refreshes through the day, marks the newest development on each timeline, and links its sources on every card.

Use a newsletter when you want one considered edition a day, chosen and written by people.

Use gg2 when you want to check in more than once, go story by story, and see where each claim came from.

This is the fairest comparison, and worth being straight about: if you know what you want to ask, a chatbot is excellent. It will go as deep as you like, answer follow-ups, and adapt to what you already know. gg2 does none of that.

The catch is that you have to know what to ask. A chatbot will not tell you that something happened this morning that you should care about — and when a model answers from memory rather than from today’s reporting, a confident, plausible, wrong answer is a real risk on a story that is hours old.

gg2 inverts that. It looks at what is trending, searches reporting published that day for each story, and writes the card from that material. You get the stories you did not know to ask about, in a fixed shape, with the sources attached — and anything not confirmed by the material is labelled rather than blended into the summary.

Ask a chatbot when you have a specific question and want to go deep.

Use gg2 when you want to find out what is worth knowing today without writing a single prompt.

聚合器把众多媒体的标题汇到一处，强在广度：想知道此刻都在发什么、然后点进原文，它们就是干这个的。

但它给你的仍然是一串标题。要还原一件事，得自己点开好几篇、再把碎片拼起来。

用聚合器：你要的是覆盖面——更多媒体、更多条目、原文入口。

用 gg2：标题你已经看到了，你要的是这件事本身：背景、顺序、分歧，以及哪些还没定论。

Ground News 围绕「一件事是怎么被报道的」：哪些媒体在报、它们在光谱上的位置、哪一边干脆没报。 如果你要看的是媒体立场差异，它就是为此而生的，gg2 替代不了。

gg2 对准的是事件本身而非对事件的报道。卡片里确实会标出各方分歧，但重心在「发生了什么、为什么重要」。

用 Ground News：你的问题是「谁在报这件事，报得有多不一样？」

用 gg2：你的问题是「这件事到底怎么回事，为什么重要？」

人工编辑的每日简报——1440、Morning Brew、Axios 之类——一天到一次，有编辑判断什么重要，有自己的语气。 这些都是真优势，而 gg2 一样也没有：它是 AI 写的。

反过来的代价也真实存在。邮件是一份定版：早上送达，事情后续变化它不会更新；能装进版面的才会写；也很少把底下的报道摊给你看。 gg2 全天刷新，每条时间线上标出最新进展，每张卡都附来源链接。

用邮件简报：你想要一天一份、由人挑选并撰写的定版。

用 gg2：你想一天看几次、一件事一件事地读、并且想知道每句话的出处。

这是最该正面回答的一条：如果你知道自己要问什么，聊天机器人非常好用——想挖多深挖多深，能追问，能迁就你已知的部分。 这些 gg2 都做不到。

问题在于你得先知道该问什么。机器人不会主动告诉你「今早发生了一件你该关心的事」； 而当模型凭记忆而非当天报道作答时，对一件几小时前的新事，一个自信、通顺、但错的答案是实打实的风险。

gg2 把这件事反过来：先看什么在热起来，再针对每条事件检索当天发布的报道，然后才动笔。 你拿到的是那些你本来不知道该问的事，形状固定、附着来源——材料证实不了的说法会被单独标出，而不是混进正文。

问机器人：你有具体问题，想往深里挖。

用 gg2：你想知道今天有什么值得知道的，且一个字的提示词都不想写。