@PawelHuryn Hello, you mean that the usage amount for Claude at $200 is 20 times, rather than the online claim that the weekly quota is 20 times and the monthly quota is 10 times?
@thsottiaux 6. While I acknowledge that this is an excellent product and concept—one that could truly become a groundbreaking innovation if executed correctly—it is currently frustrating. Even basic user interactions are poor, failing to provide any sense of reassurance or control.
@thsottiaux 5. After using it for a day, I still find it less reassuring and convenient than simply executing tasks directly via Codex. It is unclear where to adjust many settings, resulting in a very poor user experience.
@thsottiaux 4. You have to constantly monitor the system when approvals are required. If you don't, you risk missing critical approval requests, giving DOT a valid reason to halt operations and achieve nothing.
@thsottiaux 2. Claude, however, is different—it can at least grasp the intent behind such instructions. When I return after a few hours, I find it has progressed to the expected stage.
@thsottiaux 1. DOT is truly disappointing; it is marketed as running around the clock, yet it frequently halts on its own due to approval requirements. Even when fully authorized and explicitly tasked with achieving a goal during deployment, it will arbitrarily stop and demand permission.