HN user
I like this, and it mirrors my experience.
I felt that, by using the "full agentic way" I am implicitly accepting the fact that all the knowledge I have right now is all the knowledge I will ever need or want to have (with the exception of new knowledge on how to ask AI to do things, I guess).
This seems like a nice way to enable yourself with AI, but not replace your brain completely.
I recently started an internship in a field I am very interested in. I began using claude to write a lot of my code, but realized that:
a) It was way too easy to just auto-approve everything. Answering the 5-10 spec questions it asked me made me feel like I was an important part of the loop, but really it was just a way to make me feel important while spraying my slop cannon.
b) I wasn't actually learning anything, defeating the whole purpose of the internship I worked hard to get.
I am now using a workflow where the brainstorming process is the same, but I have claude write an instructional document for me to implement. It has instructions to ask me questions about what I know / want to know, to lay out the plan iteratively with lots of verification steps, and to heavily explain portions of the code that are unfamiliar to me. It's sorta like making my own custom tutorials specifically for the problem I am working on.
It's a little slower, but not too bad since it does still put whole codeblocks in the instructions. I have a much better understanding of what I am doing, I still get to enjoy learning and programming and improving, and I don't feel like a reverse centaur.
Very cool. The satellite views are awesome! I clicked around for 20 minutes and still felt like there was so much more to explore. Thoughts:
---
Unless there is a soviet military installation in Southend, Saskatchewan, something is seriously funky with the "military installation" layer.
It would be nice if clicking didn't lower the zoom level. It's fine to zoom in and center, but having it zoom me out every time I click something was annoying.
I was just talking to a friend of mine who has been making webapps for himself in a similar fashion. Very little to no programming experience. His first app scans his course notes (med school) and creates structured question banks. He's released it so everyone at his school can sign up with their institutional email. The front end is hosted with vercel and the backend with supabase.
He also has one for tracking the stats of the volleyball team he coaches. He can do things like track the direction a player hits the balls during a game and save it for review later. Hosted with Vercel and Firebase I think.
Point being: he has no experience with software development before this (although he did have some data science experience), and in the space of a couple months has produced two high quality webapps that are being widely used in his circles.
I was pretty shocked, but after seeing the apps Claude made for him (or told him how to make). I can believe this story.
I was in Singapore last summer and have fond memories of going for a run and seeing a family of 6 or so otters running on the sidewalk a ways ahead of me at NUS :). They make such lovely little bouncing arcs when they run.
Cool to hear it's the result of government effort.
Could Tomas something have been referring to some of the risks mentioned here [1]? Particularly the overdiagnosis bullet.
I am not a neuroscientist, but I thought the actual physiological cause of addiction was similar in both nicotine and gambling: you crave the predictable release of dopamine.
If that is the (heavily simplified) case, is there a distinction for you between a chemically-induced dopamine release from smoking and, say, and a button you can press that magically releases dopamine in your brain?
Realized I forgot about the golden rule of feedback: start with something positive! I will end with it instead.
I thought the UI was slick. You have identified a very real problem I am having as a PhD student, and this solution genuinely feels like something I would use. Cool stuff.
Serendipitous, as I was just planning a lit review on a topic I am unfamiliar with, and was frustrated by my increasingly messy list of "to read".
Quick thoughts:
- I would like a "not read" stack, for papers that I peeked at but decided against reading.
- The registration link took a few minutes to come in, I thought something was broken.
- I keep trying to click the stack titles to take me into the stacks, but they aren't clickable.
- I couldn't figure out how to use a summary credits
- I couldn't figure out how to search/export my notes and tags
- I would like a way to quickly see my own 1-line summaries of the papers I have read. Maybe under the titles/authors.
---
Long Thoughts:
What does this do that a "to-read" folder on Zotero doesn't? This post made me ask myself why I don't just use the to-read folder I already set up in Zotero.
1) Adding something to zotero is a bit of a process, and feels like I have to be 100% committed to reading it, and thus I hesitate to put the "I'll check it out later" papers into my to-read folder on zotero. Your website is much more focused and low-effort to use.
2) I often forget why I wanted to read something in the first place. I like to store items on my to-read list according to the parent paper I discovered them in, to remember what I was doing when I added them to the list. Zotero notes/relations are not the easiest way to do this, so I stick to a markdown list.
3) My to-read list is so long, with so many papers from 2 years ago that I didn't end up needing, that I don't know where to start when I look at it. It would be nice to have a way to rank papers on my list by priority, or group them by topic tags, etc.