norootcause.surfingcomplexity.com
Student of complex systems failures, resilience engineering, cognitive systems engineering. Will talk your ear off about @resilienceinsoftware.org
1,012 posts
1,454 followers
522 following
Regular Contributor
Active Commenter
comment in response to
post
I’m half-crazy…
comment in response to
post
No no, I think we've learned our lesson and we will no longer be introducing bugs.
comment in response to
post
Bugs: just don't code them
comment in response to
post
It’s these existing breakages that haven’t bit you yet that you should be most worried about, because the risk is invisible to you.
comment in response to
post
Ouch!
comment in response to
post
I admit it, I have a problem
comment in response to
post
I think just getting to a consensus on that numerical rating for a priori risk awareness would be a fun exercise.
comment in response to
post
This skeet inspired by Vanessa Huerta Granda’s SRECon talk: “Learning from Incidents at Scale; Actually Doing Cross-Incident Analysis” youtu.be/Q69WND8YHag
comment in response to
post
Note: I am very pro incident retrospectives! It is the RCA approach to it that I criticize.
comment in response to
post
These are reasonable claims, but the past two decades of safety science research suggest that they don’t apply to complex systems. github.com/lorin/resili...
comment in response to
post
“We never imagined this sort of failure mode could happen.” Yep. The problem is not that you didn’t imagine it, but that you didn’t recognize that there are unimaginable failure modes.
comment in response to
post
AIs as workarounds for broken processes
comment in response to
post
I am deeply convinced that a lot of the things that people want AIs to do are things that either don't need to be done at all, or that do need to be done but have been done so badly for so long that people have given in to despair.
comment in response to
post
Exactly!
comment in response to
post
Not where I am
comment in response to
post
Because we live in the adaptive universe. surfingcomplexity.blog/2020/01/19/t...
comment in response to
post
I know, I know, but I blame Atlassian because of their ALL CAPS PROJECT NAME convention.
comment in response to
post
Be honest: which outcome is more likely?
comment in response to
post
Heck, JIRA engineer is a legit thing! A thing that I hope to never be.
comment in response to
post
Why remove pointless meetings when we can send bots to them instead
comment in response to
post
LLM-generated status updates that are consumed by other LLMs.
comment in response to
post
Watch for their book in 2029!
comment in response to
post
comment in response to
post
quality might go up without constant meddling from Reviewer 2 trying to get their papers cited
comment in response to
post
I need to start making my own LLM
so it can be like: “kevin what is your estimate?”
“high confidence of that? list subtasks for me”
“It will actually take {{ secret algorithm: X * 2 }} because of {{ we don’t care just make up reason }}”
comment in response to
post
Next up: LLM-generated promotion packets that are evaluated by other LLMs.