𝗪𝗵𝗮𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝗲𝗱 𝘁𝗵𝗶𝘀 𝘆𝗲𝗮𝗿 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 𝘁𝗵𝗶𝗻𝗴𝘆 𝗮𝘁 𝗔𝗻𝘁𝗵𝗿𝗼𝗽𝗶𝗰? 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗰𝗼𝗺𝗶

𝗪𝗵𝗮𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝗲𝗱 𝘁𝗵𝗶𝘀 𝘆𝗲𝗮𝗿 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 𝘁𝗵𝗶𝗻𝗴𝘆 𝗮𝘁 𝗔𝗻𝘁𝗵𝗿𝗼𝗽𝗶𝗰? 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗰𝗼𝗺𝗶𝗻𝗴 𝗻𝗲𝘅𝘁?

The week(s) after re:invent are typically the time, where a pile of very good content videos pile up at my end and I have both a hard and fun time to get through this.

Today the “𝗔𝗪𝗦 𝗿𝗲:𝗜𝗻𝘃𝗲𝗻𝘁 𝟮𝟬𝟮𝟱 - 𝗪𝗵𝗮𝘁 𝗔𝗻𝘁𝗵𝗿𝗼𝗽𝗶𝗰 𝗟𝗲𝗮𝗿𝗻𝗲𝗱 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 𝗶𝗻 𝟮𝟬𝟮𝟱 (𝗔𝗜𝗠𝟮𝟳𝟳” ended up on the top of the pile and turned the time on the rowing machine in the gym to a fun learning experience. After all it’s a level 200 session - should be able to follow it during a casual workout, right?

❤️ Great talk by Cal Rueb - I loved not just he content but his casual story telling style. After all learning from real experience is the best - and certainly Cal’s experience from building Claude Code are a good place to start from .

💰 Quite some stuff in the tall will get to it some time. Loved e.g. the statement that absolute cost per token might not be the only dimension to look at but also factor in token efficiency per use case - how many billed tokens a model actually spends to get your task done, so a pricier-per-token model can come out cheaper per solved task.

🏚️ @Cal Rueb, the slide on the first screenshot broke me kind of. I was sure that this must have been Claude, but then you mentioned that is is not. I know get it, but took me a long time that things with timestamp don’t need to be ordered by timestamp on the slide.

🏭 𝗛𝗮𝘃𝗲 𝘆𝗼𝘂 𝘀𝗲𝗲𝗻 𝘁𝗵𝗲 𝘀𝗲𝘀𝘀𝗶𝗼𝗻? 𝗪𝗵𝗮𝘁 𝗶𝘀 𝘆𝗼𝘂𝗿 𝘃𝗶𝗲𝘄?  Cal is pointing at a talk be his colleague Alex on Long Horizon coding which made it on my viewing list.

@Cal Rueb: What happened to the Claude Code project building entire time at re:invent based on visitors GitHub issue? Frankenstein or the next big thing?!

🎬 The session itself: https://www.youtube.com/watch?v=TledrLrVUQI

🎬 Quite a bunch of breakout Sessions on YouTube: https://www.youtube.com/playlist?list=PL2yQDdvlhXf9gdFFBcDPUHAJS7kkIkIet