Entirely Foreseeable AWS Outageshttps://rys.io/en/182.html
Uncategorized
12
Posts
5
Posters
1
Views
-
@rysiek I 100% agree. Any sane #Sysadmin/#DevOps person would far prefer a deterministic tool to a non-deterministic tool. Science has everything to do with finding that which is reliable and deterministic, and can be trusted, used as a basis, used as a foundation to build higher, more trust-worthy things. It all comes back to trust at the end of the day - *not* saving time learning difficult things, *not* eschewing doing things "the hard way", *not* taking shortcuts.
@Owl Eyes Yes, but also a bit no.
On one hand you want your tools to be deterministic and predictable. On the other hand, people aren't deterministic either, and they are the most valuable tools 🤔
Maybe in time these agentic tools become predictable enough to be used as sysadmin's grease monkeys, but my experience so far is that I can't let it do its job without my constant vigilant supervision.
@Michał "rysiek" Woźniak · 🇺🇦 -
Entirely Foreseeable AWS Outages
https://rys.io/en/182.htmlOnce you strip away the marketing hype, agentic systems like Kiro AI are just automation tools.
The difference between Kiro and regular infrastructure management tools is that the latter are deterministic. They can be tested, analyzed, and bugs can be reliably, provably fixed.
That's just not the case with agentic tools. They are by their very nature non-deterministic. And that's the last thing a systems engineer should want.
@rysiek This is why I think “AI first” approaches promoted by some (many? all, nowadays?) companies are based on lack of understanding of the systems they develop. It should be opposite: algorithms first. Use machine learning if there’s no better deterministic algorithm to accomplish a specific task. That’s how, I guess, systems would work if engineers decided on their design more often instead of managers dictating solutions to them.
-
@Owl Eyes Yes, but also a bit no.
On one hand you want your tools to be deterministic and predictable. On the other hand, people aren't deterministic either, and they are the most valuable tools 🤔
Maybe in time these agentic tools become predictable enough to be used as sysadmin's grease monkeys, but my experience so far is that I can't let it do its job without my constant vigilant supervision.
@Michał "rysiek" Woźniak · 🇺🇦@hans I agree people are precious and unique - and also super non-deterministic.
I think AI can be most effectively used to stimulate new thinking, and make bold forays into new-to-oneself domains of knowledge. The LLMs can reveal lots of new terminology, and possibilities one didn't think to search for before. But then one has to take one's time, verifying the claims (often wrong), and synthesize a truth for oneself, going beyond the hallucination of the LLMs. So I think of LLMs/agents more like *discovery* tools, at best, but much less something to build solid foundations with.
-
@Owl Eyes Yes, but also a bit no.
On one hand you want your tools to be deterministic and predictable. On the other hand, people aren't deterministic either, and they are the most valuable tools 🤔
Maybe in time these agentic tools become predictable enough to be used as sysadmin's grease monkeys, but my experience so far is that I can't let it do its job without my constant vigilant supervision.
@Michał "rysiek" Woźniak · 🇺🇦> On the other hand, people aren't deterministic either, and they are the most valuable tools
Whoa, okay, maybe let's start by not calling people "tools".
You have a massively complex system like AWS infrastructure. You have engineers who are not "deterministic" in the sense that software is deterministic, managing it.
Why on Earth would you want to complicate your life and take on loads of risk by adding another layer of random non-determinism in there? Makes no sense.
-
@rysiek This is why I think “AI first” approaches promoted by some (many? all, nowadays?) companies are based on lack of understanding of the systems they develop. It should be opposite: algorithms first. Use machine learning if there’s no better deterministic algorithm to accomplish a specific task. That’s how, I guess, systems would work if engineers decided on their design more often instead of managers dictating solutions to them.
@aemstuz mostly agreed, but I'd replace "AI" (which is a marketing term) with "machine learning" (which is much less of a marketing term and more of a technical term).
-
@aemstuz mostly agreed, but I'd replace "AI" (which is a marketing term) with "machine learning" (which is much less of a marketing term and more of a technical term).
@rysiek Great point. I’m gonna edit the toot.
-
Entirely Foreseeable AWS Outages
https://rys.io/en/182.htmlOnce you strip away the marketing hype, agentic systems like Kiro AI are just automation tools.
The difference between Kiro and regular infrastructure management tools is that the latter are deterministic. They can be tested, analyzed, and bugs can be reliably, provably fixed.
That's just not the case with agentic tools. They are by their very nature non-deterministic. And that's the last thing a systems engineer should want.
@rysiek How do you funnel suckers, I mean, errr... investors, into "deterministic tools we've been using for decades", though?
-
> On the other hand, people aren't deterministic either, and they are the most valuable tools
Whoa, okay, maybe let's start by not calling people "tools".
You have a massively complex system like AWS infrastructure. You have engineers who are not "deterministic" in the sense that software is deterministic, managing it.
Why on Earth would you want to complicate your life and take on loads of risk by adding another layer of random non-determinism in there? Makes no sense.
-
@hans I agree people are precious and unique - and also super non-deterministic.
I think AI can be most effectively used to stimulate new thinking, and make bold forays into new-to-oneself domains of knowledge. The LLMs can reveal lots of new terminology, and possibilities one didn't think to search for before. But then one has to take one's time, verifying the claims (often wrong), and synthesize a truth for oneself, going beyond the hallucination of the LLMs. So I think of LLMs/agents more like *discovery* tools, at best, but much less something to build solid foundations with.
@Owl Eyes That's my experience so far, indeed. But the new knowledge that my tool has brought me, is more often than not wrong.
It has no problem telling me to configure settings that don't exist, even after I tell it they don't exist.
So far, it has been an interesting journey, but hasn't saved me any time. In fact, I have spent a lot more time, because I had to explain to my agent what I wanted, and go on a wild goose chase afterwards to check its solutions.
But I kind of expect that in a few years, these tools will become good enough to actually help. -
> On the other hand, people aren't deterministic either, and they are the most valuable tools
Whoa, okay, maybe let's start by not calling people "tools".
You have a massively complex system like AWS infrastructure. You have engineers who are not "deterministic" in the sense that software is deterministic, managing it.
Why on Earth would you want to complicate your life and take on loads of risk by adding another layer of random non-determinism in there? Makes no sense.
@Michał "rysiek" Woźniak · 🇺🇦 I'm a sysadmin myself, so I can call them tools 😇
But agree: at the moment these agentic tools aren't good enough to be trusted with massive, complex tasks. But I would be surprised if that would remain the situation for long. -
undefined oblomov@sociale.network shared this topic on
Gli ultimi otto messaggi ricevuti dalla Federazione
-
“la solidarieta’ non e’ reato! la resistenza non e’ terrorismo”. il collegamento con le mobilitazioni per i prigionieri politici palestinesi
@anarchia
AGGIORNAMENTO SABATO 21 FEBBRAIO. ORE 16.35. Ci siamo collegati in diretta con il presidio davanti al carcere di Rossano Stazione (CS) dove abbiamo raggiunto
-
bocchino teme la querela per il suo libro "Giorgia meravigliosa stupenda inarrivabile premier"!
-
i say as the fucking thing reboots again...
-
so the DMA bug I have been chasing for the past almost-year on my thinkpad?
turns out it was a typo in the DTB file. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/arm64?id=b38dd256e11a4c8bd5a893e11fc42d493939c907
-
@downdetector they're back.
-
@lashman site seems down from here?
-
@sassdawe @lashman I would not recommend that route for this use-case, it's got a ton of sharp edges and you'll spend more time figuring out if there's a bug in X11 forwarding or your app than actually working on your app.
a VM is much cleaner and far better for a development workflow like this.
