@Suuumedhh @S48794Sinha Some additional thoughts, leave yourself some linear reading cues in the comments, because when you come back to it later, it's often like looking at the code for the first time again. But the notes and thoughts you've left for yourself will help you grasp it better.
@poteto I'm really curious what changes have happened to the way you study in this day and age. Do you still take notes? If so, how do you do it? I'd love to hear about it!
@poteto@bot Oh, looks like non-verified users can't DM you. I mainly use ZCode, Kimi Code, and Codex right now. I used Cursor before and I'm looking for a chance to try it again! Honestly, the Agent experience and UX in Cursor really impressed me.
@kepano For me, I'll write a note on a single topic, and use Canva to visualize the structure. Basically, I use Markdown notes mostly for the Agent to read (and for my own linear reading), while I look at Canva more to get the overall logic.
Just realized in Obsidian Canva that we can narrow a Note down to a heading or block, but right now we can only drag the Note onto Canva first and then narrow it down.
@kepano Would it be possible to add some shortcuts to make this process easier?
@brian_lovin The Grok bot feels like ChatGPT Work with an additional virtual machine. The team has abstracted things even more, it's like @ colleagues on Slack.
@mattyp Thanks for the intro article, but it seems like a more general version of the Cursor Cloud Agent for knowledge work. Why did Cursor decide to launch such a standalone app?
@prades_maxime@cursor_ai I subscribed to Cursor for a while before I realized that BYOK requires a subscription to use. I hope Cursor will make BYOK available to free users so more people can experience the Harness benefits.
@tianzhouchen OMG, that's exactly how I feel. It feels like the team added way too many features, and my brain is completely overloaded. Notion isn't as intuitive as it used to be.
@mattpocockuk I've been thinking that in this era, the importance of having a broad range of knowledge seems to have increased. You need to know certain things in order to describe them to AI, because people can't simply articulate concepts they don't understand.
@ericzakariasson Everyone has different use cases, so it's really hard for Benchmarks to reflect personalization.
I think we could save our own test cases for specific scenarios, like including prompt and the background information, and then test how the model performs.