ai coding has given me way too much confidence
iβll start building something that wouldβve seemed completely impossible to me a year ago, get surprisingly far, and then immediately come up with another project
kimi deliberately made its new model worse at hacking before releasing the open weights
i actually respect the decision, but it's kinda crazy that we're reaching a point where AI companies have to make their models less capable on purpose
never thought we'd get here this fast
over 20,000 people apparently preordered those little AI robot ducks after Jensen Huang showed them off
i've been following AI breakthroughs every day and somehow a tiny duck that does backflips is one of the most convincing products i've seen
π¨BREAKING:
1Password is rolling out a way for Grok Bot and Meta Muse to log into your accounts without ever seeing your passwords.
ChatGPT Work, dots and Claude Duet are coming next.
you approve the login yourself.
this is how AI agents should've handled passwords from day one.