@seantheolane Gosh the more I think about it the more terrifying it is. Not only globally able to push software without SA intervention but signed kernel space software without SA intervention.
If they’re compromised or a rouge actor… game over. 😱
@seantheolane No doubt there is an issue with Q/A on CS side here.
However it’s just bonkers to me that a third party can just deploy software to the globe like this without SA’s interaction.
Bad actors wet dream.
Maybe I am a bit naive here. Correct me where I’m wrong.
But I don’t understand how this Crowdstrike thing could’ve rolled out everywhere all at once. Do sysadmins/IT departments have no control over their end devices updating?
@seantheolane I don’t understand why anyone in their right mind would allow a third party to fully control patch application — especially on critical devices.
@seantheolane Maybe I have a weird view but I’d think it would work like this:
- CS releases a patch
- Dashboard to manage devices allows sysadmins (SAs) to apply it
- SAs would apply to their test machines and validate
- SA’s would deploy to their fleet after validation
Sure, Crowdstrike absolutely should’ve caught this on their end. No doubt. But I don’t think that dissolves each company of their own responsibility here either. Especially when considering DR strats.
@AdamAshworth2@TeslaTruckClub@saradietschy Latency makes a crazy difference. Especially on websites. Gaming is highly optimized for low latency and has a ton of tricks to hide it. Whereas websites often make tons of serial requests and just feels bad when latency is high.
WiFi and desktop snaps since we are sharing 😁
@MackeyTech@TeslaTruckClub@saradietschy From my experience it is while it is under 100% load, I use 5G for my home internet and when playing games my ping stays at around 24-32ms constantly, I would obviously notice if it jumped up to 150 even briefly.
Mine is about 150ms while also downloading at 100%.
@AdamAshworth2@TeslaTruckClub@saradietschy It’s 428ms while under bandwidth load. Not sure when it transitions from the 39ms to that. Could be only when fully loaded or could be even under light load. Impossible to tell with the data provided.
It’s crazy how many bullshit jobs you encounter in a huge organization. Like I recently talked to a “network engineer.” Wait what? A whole job for this? Plugging shit in? Be serious.
@josevalim I’m curious, and apologies if this is naive, but how far from *basic* function type checking are we? Say:
:ok = func_which_returns_tuple()
Throwing a compilation error instead of runtime match error.
Exciting stuff so far. Can’t wait for more.
@JuvHarlequinKFM Appreciate the response. Looking forward to any follow-up’s you do when/if you get time to. It’s a very interesting subject and super cool demonstrations of the tooling available to you.
@MackeyTech Short answer: no. Someone else ended up building on the cleanups that I did and made some significant improvements. I keep meaning to go back and try again, but I took over protobuf.dev, which consumes most of my time.
. @JuvHarlequinKFM watched your talk on lock-free MPMC Q. Curious about a follow-up. Did you ever get to try those other ideas? Did they make it faster than the mutex impl?
In any case, great presentation! I really enjoyed it despite never having written any real cpp in my life