@dvasishtha Many of the payers have mentioned on earnings calls more prior auth gaming. I assume those prior auth rates will be get beaten back down as payers adopt their own agentic prior auth adjudication/denial programs...
@daphnezohar The scale of the drugs does mean that this could be something like 5k new cases if 50M ppl are on it in the US? Sudden blindness sounds like a high impact negative even if the incidence is low..
@Borlaug_ I'd throw $WELL and $PACS in there as well. With the clinical genomics companies it does seem like you are betting on innovation and have to take a stance on which innovator wins? With nursing homes they can all win because demand will exceed supply and benefit all boats
@FundamentEdge A subset of the problem could be training an agent to make the same a priori decision you would make on a stock you know well. If I am Citadel's COO, I am running a ton of experiments seeing if we can get agents to autonomously make the same decisions as the firm's best analysts
@reidhoffman How do you tell a contact that you don't think an intro makes sense? What if the other person doesn't want to connect with your introduction (even if you think they should?) what do you tell the original asker?
@DimitryNakhla Why not incorporate further multiple compression from here due to rising competition from large players that may erode future economics?
@kevg1412 You missed the tall ships but maybe a few still in the harbor by Seaport. There is a sand castle competition this weekend at Revere beach that brings in international artists
@nikillinit I find this debate so interesting because presumably engineers already had individual performance goals related to business goals. Who cares if they use AI or not if they hit their KPIs. If AI helps an engineer hit their goals then I'm not sure why they wouldn't use it
@blueprintsmb22 If he is Millennial or GenZ I find that @sircalebhammer (shock and awe) or @ramit (appeals to people who think of themselves as smart) has the same effect on the younger generation...helps them get that moment of "I'm not poor, I'm just too entitled"
@nikillinit My understanding is that cowork consumes a lot more tokens than Claude code for the same task because the context window cowork is looking at is a lot bigger. So doing same task in Claude code is cheaper since the scope is more compact.