I swear I always get terrible results, and it seems to burn a frickton of tokens going crazy with sub agents and workflows.
I get far better results from simply prompting Claude to review the diff of the current branch against origin/main.
Do others have the same experience? How are you using Claude to effectively review code?
Or is it skill issue?
Please educate me.
I bought some tickets today and opted for ticket insurance. After completing the purchase in the TicketMaster app with my PayPal account, TicketMaster deceptively prompted Apple Pay to complete a purchase for "ShipmentProtect.com", which I thought was the ticket insurance and so I tapped to authorize the transaction.
It turned out this is something completely unrelated from the ticket insurance and I have no idea what it's for. It seemed like some kind of scam embedded via third party advertising, in the promotions that TicketMaster advertises after you complete a purchase. It utilizes dark patterns to get users to pay thinking that it's for the ticket insurance.
Why is it TicketMaster embedding a scam like this in its checkout flow?