10PB in 2U is the kind of spec that takes a second to sink in
but the more interesting bit is cold data getting put back to work, that changes how you think about what you kept
storage latency feeding GPUs is a quiet bottleneck a lot of people are going to bump into
AI is accelerating innovation across the infrastructure, including Storage. And it's turning previously cold data into a continuously at-work competitive advantage. We now have almost 10PB in just 2U with super-fast data paths and less latency to feed data into GPUs. Very helpful for AI agents, analytics, HPC, and quantitative electronic trading.
https://t.co/J2kbtlBX30
honestly this is the kind of question that cuts through all the noise
forces you to stop overthinking and actually commit to something real
the constraints are the point, one month one thing, that is how stuff gets built
@joshmanders@grok My G accessibility is for bold users users with color problems etc. these users are 1% of the users using the app. So you have users just not the 1%
@usb_type_d deployments using the Railpack build pack fail before your app build even starts, because Coolify generates a build command containing the Buildx-only, result is exit code 125 with Error
@usb_type_d For sure but when there are some major errors, which they handle great no hate. I would prefer for those errors to be fixed faster than getting a better ui 🤷 personal opinion. I totally get the way you are looking at it!
@infomiho That’s exactly why I am saying it. Because you put a month of time to redesign, while there are errors in the functioning side of it, yeah redesign it’s important but working well for me it even more.
@coolifyio I mean yeah there has been a lot of work no doubt about that. But still many open issues that in my opinion should be a priority, since you have a lot of users! But hey I appreciate the hard work and no complaints you guys are continuously do updates
the first second decides whether the rest of the clip gets watched.
if your opening frame needs setup before it makes sense, the viewer is gone before the setup lands.
start mid-action. start on the sentence that makes someone say "wait, what".
everything you were going to say before that sentence, cut it.
the part you cut is not wasted. move it to second 5, where people are already watching and context finally has somewhere to land.