Engineer, consultant, advisor. Ex-Netflix, Twitter, Comcast. Author of “Push to Prod or Die Trying”, a guide to building real-world, high-volume systems.
Pennsylvania, USA
Joined April 2011
- Tweets 2,695
- Following 385
- Followers 725
- Likes 12,100
Pinned Tweet
🔥 interested in reading real-world stories about violating engineering principles, at scale?
look no further pushtoprodordietrying.com/
if you were building tools at a big company (call it a “platform team” if you wish), which would you choose?
1) build things that nobody asked for, but you think they’d find useful
or
2) only build things that people specifically ask for
you should have better problems tomorrow than you have today
and if that’s not the case, then you aren’t doing good architecture work
you may need to find new architects
or consider not having any architects at all
recently partnered with @GergelyOrosz to write "What is good software architecture?" for The Pragmatic Engineer:
newsletter.pragmaticengineer…
the core thesis is that good architecture work involves upgrading your problems
from the AWS incident summary:
> We have already disabled the DynamoDB DNS Planner and the DNS Enactor automation worldwide.
I wonder how the DNS records get updated now. manually tuning the weights every day?
Push to Prod or Die Trying: High-Scale Systems, Production Incidents, and Big Tech Chaos by Matthew Hawthorne @mhawthorne is the featured book on the Leanpub homepage! #DistributedSystems #AmazonWebServices #Aws #Testing #PersonalTransformation #books
Build high-volume systems.
Thrive through sociotechnical challenges.
Survive gruesome production incidents.
Find it on Leanpub!
One thing that’s annoying about being a software engineer is that you’re surrounded by people who have strong opinions about the stupidest possible stuff that couldn’t possibly ever matter even a little bit. This is your whole career. It’s just arguments like this one over and over and over and over again and they literally never matter. The years roll into decades and nobody ever stops talking about this stuff and it didn’t matter in 1980 and it didn’t matter in 1990 and it didn’t matter in 2010 and it doesn’t matter today.
Matthew Hawthorne retweeted
Craft Is the Antidote to Slop
"The devil’s oldest strategy is, of course, promising godlike creation without godlike effort."
Matthew Hawthorne retweeted
first, that’s such funny marketing “boss says just rollout ai” — good one.
second, check out early release of a book i wrote about ai evaluations. thank you!
Boss says “just roll out AI” with no details?
Raise your hand if you can relate 🙋
AI Model Evaluation by @LeemayNassery shows you how to test, tune, and trust your models before they hit production. It’s fresh in MEAP + 50% off with code nasseryML through Sept 4th: hubs.la/Q03DHV2D0
Matthew Hawthorne retweeted
In 2022, McKinsey was paid $55M to advise Warner Brothers to combine with Discovery.
From 2022-2025, McKinsey charged Warner Brothers Discovery $37M by advising the company to change HBO to HBO Max, then to Max, then back to HBO Max.
In 2025, McKinsey billed Warner Brothers Discovery an additional $63M to determine that Warner Brothers and Discovery should be separate brands again.
Warner Bros. Discovery to Split Into Two Companies variety.com/2025/tv/news/war…
editing chapter 3 of my book, just found the typo "insourmountable"
feels like it should be a real word
Matthew Hawthorne retweeted
A guy on my engineering team gave his standup update by just saying “it’s looking good over here” and then muted. I'm saying that forever now
