@theo Please preserve the scroll position on the diff/file viewers while swapping between threads and windows on the side panel, it's really annoying to have to scroll it back to the place you want every time
@theo Could you please start preserving the scroll position on the diff viewer? It's always painful to switch threads while reviewing and having to manually scroll back to the point you were before 😭
@theo Is there a place for me to give a few feedbacks on the t3code experience?
I love the inbox solution and I think you guys are the best on a few things but there are a few others that still annoy me 🥲,
@kentcdodds@cursor_ai This “dumb zone” is way less dumb these days, that’s one of the reasons GPT 5.6 and Fable 5 can work for hours uninterrupted and generate good results.
Matt pocock says he is happy with opus 4.8 medium effort, in that case this zone might be a bit more present🤷♂️
Obrigado Deus, o Senhor é maravilhoso, a minha primeira Copa Do Mundo com a maior seleção do mundo, só tenho que agradecer a Deus por esse momento, agora é representar a NOSSA NAÇÃO, VAMOS COM TUDO JUNTOS SEMPRE! @CBF_Futebol 🇧🇷
@noahzweben Apparently Anthropic think we're all fools. You're basically cutting all the subsidization for third-party tools (since you clearly don't offer just $200 in credits on the cli/desktop usage) and framing it as a nice addition 🤡
@lucas_montano Although they changed the default effort, the models were being worse even if you set the effort to max and that's what they explain in the rest of the article... the post you retweeted is a bit sensationalist.
@lucas_montano But this is common knowledge. You decrease the effort to gain speed.
In the article, they're just summarizing the timeline, this was not the cause of the "dumbification" of claude code. They're just outlining that it was not a good decision to default to medium.