Hacker Newsnew | past | comments | ask | show | jobs | submit | farixco's commentslogin

The problem is that you're not only putting yourself at risk when you don't update.

You're putting everyone who you've talked to at risk. I don't know about you, but I prefer not having to worry about whether I'm communicating with someone whose installation can easily be pwned by any halfway incompetent attacker.


It's the same when I install a update that I not personally security reviewed. Sorry, thats not a argument.


  > a update that I not personally security reviewed
Great, can you give me a summary of the updates for the Linux Kernel, Android Kernel, iOS kernel, libssl, and all the drivers that updated this week on my arch machine?

  > Sorry, thats not a argument.
Neither is pretending you're reviewing hundreds of thousands of lines of code a week.

This is Hacker News man, some of us actually understand how computers work.


This has 100% been my experience, even with in-person shopping.

You ask for an Arduino, and the follow up question is: 'genuine or generic?'.

I don't think the Arduino trademark is that valuable, it's already well underway genericization.


The NPU is somewhat usable on RK3588(S(2)) boards, as long as the kernel you’re using is recent enough to have the latest NPU driver.

For a quick bring-up and implementation, take a look at this repo: https://github.com/Pelochus/ezrknpu

I have a Radxa Rock 5C (on Armbian) and it runs various models rather well. Trying to convert models is kind of a pain, though, so watch out.


Works great for me on Frigate. I am doing object detection on 3x 4k streams and it’s only 20% utilized.


Has anyone found a good table of tokens/s comparing different cards and systems (CPU, GPU, NPU etc.)?

Also: recent Geekbench should have NN tests: have you found dedicated comparison pages, and is which processor (C/G/NPU) is used clear?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: