@OfficialLoganK@GoogleDeepMind So it goes as a replacement for 3.1 flash image as well, but flash image is on average 2-2.5 times faster. Is this just the rollout phase or 2.1 is just slower?
minio removed even their mc container from docker hub. So now we can't even have the outdated one from the "official" repo. It was such a good project.
@jackfriks@ClaudeDevs I have a telegram bridge service layer with a bot and a daemon that gives me abolity to restart the dead session any time + spin up new on demand and have a backup history in a telegram group or chat. Works pretty good
@levelsio@Nabil_Alouani_ I guess AI labs would tap in directly into all the news feeds, etc. They probably doing that already. And for the "expert" things there's gonna be just jobs to answer such questions. Similar to how even now there are gigs to train AI
@guyroyse It a solutrly does! I had an email a day before my vacation. So from some perspective the vacation absorbed that, but I'd also rather receive the news in a different manner. But everything is going to be great no matter what ๐
@GoogleCloudTech Would be great if someone from the data catalog team or a devrel/pm answer some issues/suggestions/discussions. I'm very much interested in the standard, but lack of communication is dawning. I've created a set of go-to okf skills as the repo only has opinionated ones
1/ Being on call when a prod incident hits at 2 AM is rough.
Distributed systems make it worse: nested traces, scattered logs, a dozen dashboards to correlate by hand.
So I build an agent to help me๐งต๐
#AgenticArchitect#GoogleAntigravity
13/ That's the blueprint: an autonomous SRE agent that scales the team, improves MTTR, and takes some of the on-call load off you.
๐ https://t.co/wNG3Tx85t3 โ full source + step-by-step CODELAB. Follow for more!