Wayne Haber

Engineering Leader | Built it, ship it, repeat | x-FloQast, x-GitLab, x-SecureWorks

[Wayne Haber @WayneHaber]

Top 3 things that Murphy's law indicates: 1. Do backups AND test restoring them. 2. Don't deploy on Fridays or before holidays. 3. Don't make important decisions when hungry, angry, anxious or tired.

[Wayne Haber @WayneHaber]

Do What Chris Rock And Julia Child Do Everyday https://t.co/0E09QPokOT

[Wayne Haber @WayneHaber]

Organizational Complexity Audit https://t.co/3fDtlJWzeZ Organizational complexity rarely shows up as a bad org chart. It shows up as meetings called to spur action that an email could have coordinated, as work that stalls because nobody's quite sure who owns it, and as teams optimizing for their own numbers in ways that quietly make the company's numbers worse. The damage lands in two places, and both of them walk out the door.

[Wayne Haber @WayneHaber]

I’m #hiring an engineering manager for an AI platform team in Hyderabad. If interested, please apply! We are at the forefront of AI in the legal space, building agents and processing hundreds of millions of documents.

https://t.co/6g5SrEdmiw

[Wayne Haber @WayneHaber]

A stat from a recent engineering leadership roundup stuck with me: teams that invested in real technical mentorship saw roughly a 30% jump in retention over two years. That tracks with me.

Engineers rarely leave because the problems got too hard. They leave because they stopped growing, or nobody was paying attention to where they wanted to go next.

The most useful thing I do most weeks isn't technical. It's noticing what someone is reaching for and clearing a path to it: the stretch project, the specific feedback they can act on tomorrow, the room to lead something and maybe stumble a little. Build the people, and they build the thing.

Mentorship isn't something you bolt on when the roadmap is calm. It is the roadmap's insurance policy.

[Wayne Haber @WayneHaber]

Four Lessons From a Week on a New Platform https://t.co/hdEHnAotjT

[Wayne Haber @WayneHaber]

Gartner now expects more than 40% of agentic AI projects to be scrapped by 2027. I also saw an essay about the "95% illusion" — a workflow that looks flawless in the demo, then quietly comes apart once rare failures start compounding across a dozen steps.

The teams that win the next two years won't be the ones with the flashiest demo. They'll be the ones who instrumented the boring stuff early: observability across every hop, honest error budgets, and a clear view of where a human still needs to stay in the loop.

[Wayne Haber @WayneHaber]

LangChain's State of Agent Engineering: ~89% of teams instrument their AI agents; only ~52% actually evaluate them.

Treat evals as a first-class part of reliability, not a research project you get to later. Decide what "good" means before you ship, measure it continuously, and tie it back to the customer outcome.

Trust is often earned in the boring, repeatable middle.

[Wayne Haber @WayneHaber]

Hi @PortkeyAI . Your "book a demo" page has not been working since yesterday - and I would like to book a demo. It looks like your calendly account is not active. Please DM me to set up a time to talk.

[Wayne Haber @WayneHaber]

Keep your eyes wide open, approach everything with circumspection, don’t accept any truth without deep thought, expose and eradicate half truths and demagoguery, learn to wonder at the beauty of the world around you and, above all, think!—about everything. —Richard Feynman

[Wayne Haber @WayneHaber]

@GergelyOrosz Everyone's arguing about which model wins. Wrong question. Model + framework + language + architecture evolve together, and the teams that tune the whole stack for human + AI collaboration will pull ahead.

[Wayne Haber @WayneHaber]

Goodhart’s Law: When a measure becomes a target, it ceases to be a good measure.

The measure should be a signal. The target should be the value the work creates.

Recent example: Token usage is a signal. The target should be faster time-to-value on new features.

[Wayne Haber @WayneHaber]

The Ninety-Ninety Rule: The first 90% of the code accounts for the first 90% of development time. The remaining 10% of the code accounts for the other 90% of development time.

My take: Starting is easy; doing the planned known work is easy; discovering the unknown work early and planning for it is hard.

[Wayne Haber @WayneHaber]

Parkinson’s Law: Work expands to fill the time available for its completion.

A reason to relentlessly prioritize and set both reasonable and stretch goals.

[Wayne Haber @WayneHaber]

Hofstadter’s Law: It always takes longer than you expect, even when you take into account Hofstadter’s Law.

[Wayne Haber @WayneHaber]

The Bus Factor is the team members who, if hypothetically hit by a bus, would cause the project to be in serious trouble.

[Wayne Haber @WayneHaber]

The Peter Principle: in a hierarchy, every employee tends to rise to their level of incompetence.

My read on it: When you get promoted, you can't just do your old job better. The responsibilities and the expectations change. If you don't change with them, the Peter Principle catches up to you.

The clearest example: your best engineer becomes a manager and keeps writing all the code. They were promoted for the thing they now have to stop doing.

Rising isn't the trap. Refusing to grow into the new role is.

[Wayne Haber @WayneHaber]

The Ringelmann Effect: Individual productivity decreases as group size increases.

The larger the group, the more coordination is needed, and the less effort each person may feel they need to contribute.

[Wayne Haber @WayneHaber]

Dunbar’s Number: There is a cognitive limit to the number of stable social relationships one person can maintain, commonly cited as around 150.

Intimate relationships (~5) Trusted collaborators (~15) Close working relationships (~50) Stable social connections (~150)

[Wayne Haber @WayneHaber]

My take on Little's Law: The more people on a project, the more communication and coordination, which in turn reduces overall team throughput. Right-sizing teams and the scope/accountability of each team is key.