_davide_

They recently released a huge bump of their AI package, they probably wanted to differentiate between the older and newer version...still they could have renamed the plugin, or at very least, mark older comments as "for an older version"

I actually pushed a fix for the soundcard upstream back in kernel 5.13, it had been working great. And yeah, don't expect a Wacom level pen, it doesn't come with tilt. I never trust/tried sleep on Linux(could corrupt the FS), but everything is working great on my side.

Anyone surprised that Autodesk is citing blender in a non-negative light?

The RTX 4080 should be capable of ~40 TFLOPS, yet they only report 2,160 billion operations per second. Shouldn't this be enough to reconsider the benchmark? They probably made some serious error in measuring FLOPS. Regarding the fact that CPU beats NPU is possible but they should benchmark many matrix multiplications without any application synchronization in order to have a decent comparison.

I really like the trend of making software companies liable for security incidents. Way too often they get away without any penalty.

Did I ever? :')

...

May I suggest a tool built for application level portals instead?

https://github.com/build-trust/ockam

One binary, easy to use, no ssh getting stuck! (yep, I work at Ockam :)

I'm waiting for AI bubble to burst to buy some of those at the fraction of the price :)

What about memory? the real price of threads is the stack.

Even when perfectly optimized, it wouldn't be enough to handle serious workloads.

Is anyone surprised?

I have a couple of M40 with 24 gb on a desktop computer, i had to tape the pci-e to "scale it down to 1x". It's ok for interference and play around with small trainings, but I can barely infer a quantized version of 70b parameters. Training anything bigger than 3b parameters in human timescale is impossible. Either you scale it down or ask for a sponsor. It's frustrating because IT had always been approachable, until now...

Everyone keeps thinking current GPT models will improve to be superhuman, it won't. It's trained on human data. Much like alpha-go had to drop completely the concept of learning from human plays since it was stuck at a local minimum. Once they started to train with adversarial networks, it evolved a well above the previous local minimum (with an extra order of magnitude of computation). So, don't expect much more from the current generation of AI.

After this, I completely lost faith in humanity :'D

I never interacted with systems older than a 286 but I'm actually curios about them. Maybe you could "lease" these computer, by virtualizing the keyboard input and stream the screen back via a simple webcam, some curios and universities could rent them for education purposes. Still I can't see any decent revenue, let alone driving a business.

indie entrepreneurs?? this is not an hobby! either you are all in or do something else! prepare to fail! you should fail, if you are not failing you are not pushing yourself enough, and it's painful for everyone.

I saw the evolution from a tiny 3 people company to a 150+ and growing, my boss and friend went from one struggle to another, as elon musk put "it's like staring into the abyss while eating glass". It's that hard and gets emotional right away.

indie entrepreneurs? influencers? really? it's an insult to everyone that put their soul into it, it makes my blood boil. There is no half measure, if you think this way you won't succeed.

gmail has a ridiculous anti-spam, it should use AI and cool stuff but it just ignore most of the email just by IP... My stack is also perfect but the provider i use sometimes ends up in a "light" blacklist due to some spammer not being blocker swiftly.

check your ip here: https://multirbl.valli.org/lookup/

zimbra - fully open source GPLv2

mail/contact/cal/task/briefcase

http://zimbra.org https://github.com/zimbra