The interesting thing about export controls is that it means only U.S. Citizens can use the system. Work on the system. Build the system.
It means they’ll need to hire Americans.
@dikshabasu classic Carnatic community stuff haha
I spent a year learning mridangam under Thiravarur Vaidyanathan — then performed at the Cleveland Thyagaraja Aradhana
13 years of vocal before that lmao — the Bay Area scene was even more so like this
@agupta skip the terraform completely IMHO with a best in class Infra Eng agent :)
Check out our agent's results from @MSFTResearch's SRE/Infra Ops bench here: https://t.co/Y1AmLcvLYf
@agupta@MSFTResearch publishing our results against Claude Code, Codex etc this week - we're well ahead of all frontier coding agents on public infra/SRE ops benchmarks!
@agupta skip the terraform completely IMHO with a best in class Infra Eng agent :)
Check out our agent's results from @MSFTResearch's SRE/Infra Ops bench here: https://t.co/Y1AmLcvLYf
@nikunj regardless - the conclusion on 'judgement' being the defensible piece that compounds tracks
"The place where new judgment gets entered, where people and agents approve and correct and reverse each other. That is the part a rival can’t migrate, at least not easily."
@nikunj i think market incumbents are just now catching up to the conversation - they were training their own models to start
now they've discovered harnesses, etc
@yoheinakajima@ActiveGraphAI the log centric architecture — esp the derived state graph and append only event log — is very similar to our agents architecture
super cool to see others thinking about this
“laid back” is what high agency ppl look like from the outside when they’ve correctly identified which games are worth playing & simply declined the rest.