AI-Driven SRE Tools. Please stop.

I’ve decided to stop evaluating AI-based SRE tools.

After evaluating many systems and finding them all crap, I’ve decided it is a waste of time for me to evaluate any more. If any company creates a successful product, I’ll look into it then. Otherwise I won’t waste my time.

In other words, I’ve decided it’s more efficient to be a “late adopter”. I’ll let other people do the evaluating (waste their time) for now.

I define “Success” as “the company has been EBITA profitable for 12 months in a row”.

If you aren’t a “success”, please don’t call me.

(I wrote a longer explanation in 2018 and nothing has changed.)

Here’s an example message I got on LinkedIn the other day:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

I’m building an AI tool to reduce the manual correlation
work during incident investigation (logs, deploys,
infra changes).

We’re working closely with a few teams to shape it and
looking for senior SREs who want real input into how this
evolves.

Would you be the right person to drive something like this
internally, or should I connect with someone else on your team?

Good for you, sir. Have fun wasting their time. When someone has a product that is more than a “parlor trick” that works for small data sets, it will be front page news in every tech journal.

I’m sure you’ll find many willing test subjects. Not me and definitely not today, my friend.