Try Google NotebookLM.

 Due to receiving news of a new feature update, I visited the website to test NotebookLM. Developed by the Google team, this tool allows AI to analyze the content and summary of articles. Currently, the beta version is only available to users in the United States. This tool can obtain content for analysis through inputting text and files. Therefore, logging into a Google account to allow access to Drive data is a convenient method.

Currently, it may still be in the early stages of development, and the interface design is not very user-friendly. Here, I am documenting my operating process so that those interested in trying it out can also give it a try.

• Step 1: Create Project

Upon entering the project, the first thing you'll see is the Note page. You can directly type and add notes here. However, during my testing, manual input did not have the line break functionality.

• Step 2: Add source

Switch to the Sources page and add a document. It can read files on Drive, accept PDF files, or have text pasted into the input field. The content I tested was the Christmas edition of Edge magazine, with a preface and the first article discussing the controversy over Unity's charges.

• Step 3: Summary and prompt

The screen that opens after uploading is a bit difficult to understand. It actually has three layers.

The first layer is the list of article squares from before. The file names on the white background constitute the second layer, and the movement direction is likely from bottom to top. The blue-purple area, which contains the summary, is the third layer, and the action is from top to bottom.

If you click the upward arrow in the upper right corner of the blue box, the blue-purple area will shrink upward, revealing the original text from the first layer. Key topics are items that have been listed for you. The preface is very short and can be summarized into five items.

Step 4: ping notes

Clicking on the suggested sentence will display the content analyzed by AI. This is the fourth-layer screen. The downward arrow next to the options for suggested phrases can close this layer, allowing you to select another line of suggested phrases (perhaps you can select the line with suggested phrases directly?). If you click "pin," that item will stay on the note page. (Currently, our tests are all on the source page.)

If you click on a suggested phrase five times, it will run five times, and there is no way to delete the record in the analysis on the fourth-layer page. When I confirm the screen, switch to a new topic by clicking on it, close and confirm, then open it again with the arrow on the prompt line, it appears at the first line. You have to scroll down to the very end to see the new one.

Here is the screen pinned to notes.

Other thoughts

In this preface, the first three suggestions almost complete the article. The summary is too simplified and lacks almost any content. The last two are a bit of self-expression. The AI deliberately stretches the suggested sentences to five items. The fourth item is gathering questions, and the fifth item is gathering exceptions. After confirming, those listed items don't seem to have much connection...

In the second article about Unity's charges, the four-page study has a more normal summary. All five items are summaries. I'm not sure if any crucial points were skipped. Texts that are too short might be challenging for AI to guess. Jokes or proverbs are based on rich life experiences.

I always feel that the risk of AI summaries lies in the possibility of skipping the key points that readers need. If the document is crucial and every word is important, it's better to read it yourself. AI summaries might be suitable for non-literary or detective genres. Texts in the realm of literary arts and poetry should probably be read personally.

I really hope to read more of the magazines I want to see each month. The number of magazines I want to read exceeds eight every month... so far, I haven't achieved that yet. If I scan through the summaries, at least I can observe the trends in those fields.