kennethops

I like to give my graces to people and the companies who are typically not trillions of dollars. Have an incredible amount of resources that many countries would like to have, with fewer of the obligations. I'm going to chalk this up to its intended

As a previous SRE at Cloudflare, I'll never shit-talk fellow SREs at big companies.

The level of scale and complexity a big tech SRE has to deal with on a constant day-to-day is a very imbalanced proposition. A lot of people, in my experience, are not fully comprehending.

You have to be a jack-of-all-trades and a master of all.

These tools are data collection mechanisms to help train these better models. I'm working with some folks to figure out a way to put a layer between the harness and the models to have better control of what data gets sent to and from the model itself and the harness.

We need to start to come up with vastly different ideas other than make more damn datacenters when we have a ton of computers on the edge. Look I wont say I have all of the answers, but we have got to start to have a conversation before this companies just use the laws to lock everyone else out.

I wrote this to start a conversation and as I get a better idea ill try to share that as well

imo it will slowly turn into where people run their own AI

I have been deeply getting fascinated by local LLMs and more sovereign AI, so I tried to capture my thoughts rawly. I will be following up with some more technical details, but I would love to hear what people think around this. I tried to have a little fun as well.

My fiancée has a PhD in biomechanics. A few Fridays ago we were winding down with some wine, and she said something that turned into 3 hour long conversation around where we think a lot of this technology is going. We tried our best to capture it here, would love to hear everyone's thoughts. It got my brain fixated on a few things as well

do you know if they did this to it?

https://research.google/blog/turboquant-redefining-ai-effici...

I love the idea of building competitor to open weight models but damn is this an expensive game to play

Yeah, software development life cycle.

>You cannot understand everything.

I 100% agree with this in a individual person sense, but in a humanity sense someone does understand linux very deeply and is very intentional on how they change it which to me is how I gain trust in it.

does trust change when the entire SLDC is AI?

a comment I cannot stop thinking about is "we need to start thinking about production as throw away" Which is a wild thought to think about when I think about on my career. We had so many dbs or servers that we couldn't touch because they were a special snowflake

I have been doing most of my writing over there...honestly I hate it. So thank you for the feedback.

Cries in defense contractor nosies. (I have)

I wanted to touch on this point but then this post started getting WAY too long.

>the only change AI brings is that even first person developers don't know anymore what the fuck they just deployed.

This to me is going to make your first point 100x worse in every damn way

>Honestly, the post itself reads very generated, very rage bate. I have so much more faith in us and our hobby/industry than this blog post.

Don't get me wrong. I think the future is very bright for software. I have friends who are scientist and biomedical professionals and I am excited to see what they are able to do with the powers of software where they don't need to care about syntax and can only lead with their intentions.

The rage bait part is mainly my frustrations manifesting. As an SRE my annoyances came off a little bit when it comes to how fast developers are shipping vs how fast our rails can maintain things.

I am mainly seeing across a lot of my engineer friends and mentor who I respect deeply. They are using swarms of agents to build crms, small business and run their homelabs.

Im not a writer. Just a guy with an opinion

I tried to write it more stream consciousness

>The real problem isn't lack of tools - it's that the knowledge is ephemeral. This 100% the problem. This is why we are trying to capture business context and attach it to the infra itself vs just keeping it in docs.

>How are you handling the drift problem? Auto-discovery polling, change events from cloud providers, or something else?

We built a pretty awesome approach to handling the drift problem. We do a combination of indexing, change even capture and then user behavior. So if a user is looking for a information we pull the live value first.

but think about the shareholders!

Woah what are you doing?

That is what I figured it would be, but you never know anymore with the amount of acronyms thrown around nowadays.

Sorry what is an ER diagram?

This has been my experience as well. imo documentation feels like one of the few areas that AI can be good at today.

But how do you ensure the .md file is able to see all of the details of the infra?

Im honestly looking for both. I haven't found a vender to do this well for just humans nor am I seeing something that can expose this context, read only, to all of the ai agent coding models

I will check that tool out.

So far the feedback has clustered around a few themes:

People want it to be significantly more proactive over time, things like root cause analysis, security-style probing, or guided investigations rather than just visibility.

There’s interest in going deeper on telemetry and using it to surface higher-level insights, not just raw data or links out to other tools.

A lot of people ask whether it can eventually write to environments. The direction that’s resonated most is doing this first for new or greenfield environments. For example, going from a prototype to a production-ready AWS setup in a more agentic way. For existing environments, trust and safety are still the gating factors.

My takeaway is that read-only context earns trust first, and write access has to be very deliberate and staged.