I agree. But I believe I need to do more baby sitting on a greenfield project than an existing project.
"garbage in, garbage out" applies at every step.
As soon as I realize that it was AI generated, I stop reading. If I am on social media, I also unfollow the account that posted it. There is also a great blog post on this. [1]
There is also one other point. To me, everything that comes from an AI needs verification. It goes to a grey area in my mind. It is neither true nor false which makes it very hard for me to learn from AI as well.
Cursor $20 subscription is also very good. auto is usually decent. I use composer 2 for heavy tasks and it does good job. If I need some UI polish, I switch to gemini 3.1 pro. The trick is to stay at auto/composer 2. But if you try to use other premium models too much, then you’ll hit the limit as well.
I felt that too, when I first used cursor/claude code, it was awesome and I just wanted keep building, the dopamine hit after shipping is really good. But later, when I needed to inspect the code manually, and I realized there were a lot of trash/dead/unoptimized code. I started drowning in the mess I generated. It is good until you need manual changes.
I currently use it for 3 things, all of them are cron jobs that runs in the morning.
1- Check hacker news for AI related news and compile a digest
2- Check new YC companies in the P26 batch
3- Check ladybird and summarize the commits in the last 24 hour.
LoL. So funny! Tbh, my L*nkedIn feed is full of this kinda crappy posts. I tried to fight the algorithm, unfollowed people etc. But I think there is no way of winning.
Input: I’ve eaten banana
Output:
I’m thrilled to announce that I’ve successfully completed the consumption of a high-potassium superfruit today.
This experience reinforced the importance of consistent fueling and maintaining a growth-oriented nutritional strategy to optimize peak performance. It’s not just about the snack; it’s about the energy and resilience we bring to the table every single day.
How are you fueling your journey today? Let’s connect and discuss sustainable energy for high-impact results!
Freezing the browser at every step is a very good approach. I am also working on an agent browser. It uses wireframe snapshots instead of screenshots to reduce token cost.
https://github.com/agent-browser-io/browser
I’ve been thinking about how AI desktop clients like ChatGPT and Claude are expanding beyond chat. With MCP, MCP Apps, and now WebMCP, they could start interacting more directly with services and websites.
I gathered a few thoughts about what that might mean for browsers and wrote a short post about it.
Looks Cool! Navigation is good overall. But It would be great if I could use ArrowUp/ArrowDown to scroll in scrollable areas and focus on buttons as well.
A few hours ago. I am learning a new language (rust) so I disabled auto complete and AI suggestions. Otherwise AI becomes more of a distraction than a booster.
But for languages I am more confident (like JS and dart), I usually don't write code and only review it.
If it is a throw-away project (like I am only experimenting something or just need something for one time), I don't event look at the code and only focus on the outcomes.
As the article states, it is a "wild experiment". I wouldn't let AI control anything serious end to end. Also if Claude really becomes JIT, it is going to be an expensive one.