@ivanfioravanti I was experimenting with this same setup on my m4 max and it is incredible. Although at some point I did thought that claude code is beemttwr but this is definitely better than what we we're doing 6 months back.
15:45 Clinical Trials conversation is 🎯
#1 challenge with running trials is finding qualified patients.
@HealthKey_HK, we actually ran the Lp(a) study Dave @EliLillyandCo discussed at 16:35. We processed 10,000s of patients, and found 216 qualified patients from a single clinic in <24 hours.
This technology was impossible to build until last year, and we solved it.
TIL that when you do "/context" in CC, it fills up your context window by at least 1k tokens, not sure if its just me bcz of my claude_md file or its just the output of the /context that fills up the context window but in either case its better to use statusline functionality
gpt-oss is out!
we made an open model that performs at the level of o4-mini and runs on a high-end laptop (WTF!!)
(and a smaller one that runs on a phone).
super proud of the team; big triumph of technology.
@coderabbi Yeah is great. I was taking more in terms of when you want to use LLMs to return you structured output you night want to use with json or xml format prompt that has the output schema.
Has anyone successfully built a claude code system where at the end of each session it can store the summary of the session and todos and then at the beginning of each session it can read last session's summary
I am trying to build one