I was lucky enough to receive the beta of YouMind 0.2 as soon as possible! After a quick experience, there are still many interesting changes, so I wrote a short report. There are many places that I didn’t have time to talk about, so I will save them for future use~ 1. Initial interface optimization It is obviously much more beautiful, although I can vaguely smell a hint of “Ai design”, but it looks good! On the right is a dynamic example diagram, and on the left is an interactive creation bar. In my opinion, “Name” and “Description” are equivalent to the title and body of OB. In this way, the creation process is simplified than before. It is worth mentioning that the small icon in the upper left corner will change. I refreshed it five times and there was no repetition. I don’t know how many kinds there are. Is it a little easter egg~? 2.Boards section optimization The overall style remains consistent, and “New snip”, “New Thought” and “Types” have separate interactive areas. The first two can be created by clicking, and the latter will list the display columns downwards by clicking, similar to the “filter” in Excel. There’s a new Open Mind Studio icon in the upper right corner. We’ll talk about it in a bit. It’s very powerful! 3. Open Mind Studio

Huge changes, countless new features, and a very high ceiling.

“Share” button in the upper right corner (link form)

Reader, chat, note three basic functions, plus (potentially) unlimited custom functions

Chat Update:

1. Added Claude 3.7 Sonnet. When I saw it for the first time, I was stunned. Can I use it directly? I didn’t expect that YouMind would provide me with an experience that I couldn’t find anywhere else. Thank you! The only drawback is that the generation speed of 3.7 is very slow at present. I am not sure whether it is a model or API problem. It is recommended to use R1 for now, which is fast and stable.

2. Support using @ to select multiple materials at the same time to ask questions: Obviously, this is to serve the creators. As long as multiple similar materials are organized into a board, Ai can summarize the common points among them. What I can think of at the moment is to empower the personal knowledge base.

3. Support one-click model change to regenerate answers, and one-click save answers to Thoughts: a step towards convenience Reader update: The previous summary function is basically unchanged. It would be great if it could support free changes to the model and regenerate answers like Chat does. It may still be the first version. Switching back from Chat’s high degree of freedom feels like “taking care of the head but not the butt” (it seems a bit harsh, I hope the team members understand what I mean) Note update: I have a say in this! I still remember one night in early February, I suddenly had an idea and sent an email to the YouMind team, asking whether the “class annotation” function could be added. It was around 11 o’clock at that time, and I received a reply half an hour later. I felt a little flattered and was thinking that the team members were too hardworking. I didn’t expect the function to be launched so quickly and with high usability. First, you can choose to add notes to the current material or board. The former is an “annotation”, while the latter can be your thoughts on the entire board. This is a very important function for me. If before it was a “bookshelf”, where one could only know the type of book but not the specific content, now it is equivalent to annotating each book. At a glance, one can know what the author was thinking while reading. This is an extremely significant benefit for creators. Once again, I would like to commend YouMind for its emphasis on user feedback and in-depth thinking on features.

Assistant update: ceiling-level functions. It supports creating a single-aspect “assistant” with the help of model capabilities, and adds it to the top bar for one-click calling. The picture shows the translation assistant I created simply, the R1 base. It only requires a few lines of commands. Just tell it what you want it to do and it will be generated perfectly. After experiencing this function, the only word that comes to my mind is “invincible”. I no longer have to be restricted by the official websites of various models. All general large models can be trained into independent assistants with just a few simple words, and there is no upper limit. The strength of the assistant depends on the sophistication of the command. So cool. From this feature, I can vaguely sense the ambition of YouMind. After evolving into a true “creator community” in the future, users will share their “assistants” with each other, just like the complicated plug-ins of VSCode today, which are flourishing and unparalleled.

At this point, I suddenly thought of one thing: the above can be used not only on web pages. In my opinion, e-books are the only thing that can truly meet “hard needs” and is useful to every user. Summarize, ask questions, annotate, translate, no matter which one, they are all perfectly adapted to the reading scene. So I tested the PDF scene without stopping, as shown below:

There is generally no problem, but there are still two flaws:

1. The layout of the translation assistant is not beautiful: Maybe it’s because my prompts are too simple. The text can be extracted and translated, but the feeling… is really hard to describe. At present, I often use the translation of Tongyi, which is natural and beautiful, and has many practical functions. For example, clicking on the translation will automatically correspond to the original text, which can be used as a reference when I question the accuracy of the translation. Of course, I know this is not YouMind’s fault, but it has discouraged me from creating a local library, which is a pity. I would like to make a wish that the official will release a translation assistant to optimize this problem. After all, this may become one of the most frequently used features.

Tongyi’s translation interface, for comparison

2. Annotations cannot be directed to a specific line or paragraph: This is not a big problem in non-reading scenarios, but it is a big problem when reading. I still hope that the Note function can be optimized. Being able to launch it so quickly shows that the team is sincere enough. I hope it will be better in the next step!

Of course, these are all for reading, which may be a bit nitpicking. After all, there are only a few people like me who treat it as a library. I’m not even sure if the official is aware of this purpose. Regardless of the above, or the fact that Epub import is still not supported, there is no consideration of “reading” anywhere. Anyway, it is also a demand. I won’t send an email this time. I hope that when it is released in 0.3, I can exclaim “Wuhu”.

The above is what I have experienced so far. Overall, it is a big surprise. Whether it’s the generous 3.7 Sonnet, the Assistant that gives full play to the user’s freedom, or the “advisable” Notes. What I see is YouMind’s unlimited potential. I sincerely believe that this can become a community for content creators. The team’s technical strength, understanding of functions, and seriousness towards users are all YouMind’s “golden signboards.” Thanks to the team for bringing such an excellent product to the public. We look forward to updates again and again, and we will move forward together!