Manifesto

A public declaration of beliefs, principles, and intentions - what a person or movement stands for.

Manifesto

A manifesto is a public declaration of beliefs, principles, and intentions - a statement of what a person, group, or movement stands for and aims to do. It is aspirational and identity-defining, written to rally the like-minded and announce a stance to the world. The manifesto does not wait to be invited into a debate; it opens one. Its unit is the principle, not the paragraph.

The canonical structure of a manifesto moves from premise to principles to call. The writer opens by naming the condition or worldview that makes this document necessary - the thing being declared against or found insufficient. The body states the core beliefs: what the movement holds to be true, what it rejects, and what it commits to. The close is an exhortation - what the writer or movement intends to do, and what is asked of those who stand with them. This arc from premise to commitment to call distinguishes the manifesto from formats that argue, survey, or instruct.

Unlike an op-ed, which argues one clear position on a specific timely issue for a publication’s readership, a manifesto declares a whole worldview or program. Where the op-ed advances a single claim and then exits, the manifesto lays down a system of beliefs and says: this is who we are. It is broader, more foundational, and concerned with values and identity rather than winning a single argument on a single question.

A manifesto’s power comes from its willingness to commit fully. It names what it stands against as well as what it stands for, makes claims without qualifying hedges, and invites readers to choose a side. That commitment is also its risk: a manifesto that retreats into concessions or softens its principles to avoid alienating anyone has abandoned the form.

Canonical template

[Opening: the premise or critique - the condition or worldview that makes this manifesto
necessary. 1-2 sentences that name what is being declared against or what has been found
insufficient.]

[Worldview: the core beliefs that follow from the premise. Stated as principles, not as
evidence-based arguments. 3-7 items or paragraphs.]

[What we stand against: the alternative, the status quo, or the opposing force - named
plainly, without equivocation.]

[What we stand for: the positive commitments, stated without hedge or qualification.]

[Call: what the movement or writer intends to do, and what is asked of those who are with us.]

When to use

• Declaring a movement’s, organization’s, or person’s founding beliefs and values
• Announcing a new stance, mission, or philosophy to a public audience
• Rallying aligned readers around a shared identity or cause
• Establishing what a project or initiative stands for at its founding or relaunch
• Responding to a status quo the writer rejects by naming the alternative publicly

When not to use

• When the goal is to argue a specific timely claim for a publication’s readership - use an op-ed instead
• When readers need practical instruction, steps, or actionable guidance rather than a declaration of belief
• When the context calls for collaborative deliberation rather than a public announcement of position

Pairs well with

columnist, executive, resolute, confident

Often confused with

op-ed: An op-ed is a short argued opinion piece - typically 600 to 800 words - written for a publication’s readership to advance one clear position on a timely issue. Where an op-ed argues one specific claim and then exits, a manifesto declares a whole worldview or program: it is broader, more foundational, and concerned with values and identity rather than winning a single argument on a single question. The op-ed’s unit is the claim and its supporting evidence; the manifesto’s unit is the belief and its implications.

Tells

• Opens by naming the premise or condition the document is declaring against - the thing being rejected or found insufficient
• States beliefs as principles rather than as evidence-based claims that require proof
• Written in a committed voice that invites readers to join the position, not merely observe it
• Moves from worldview to commitment to call, not from problem to solution
• Ends with an exhortation or declaration of intent rather than a qualified conclusion
• Holds its position without hedging - no both-sides framing or qualifying concessions
• Names what it stands against alongside what it stands for

Anti-patterns

• Hedging the beliefs to avoid alienating any reader - A manifesto exists to declare a position without reservation; hedging turns it into a policy memo or a survey of views, and loses the readers who came for commitment.
• Arguing each principle with evidence and counter-evidence as though making a case in court - A manifesto states what the writer or movement believes and invites agreement; it is not a debate brief. Over-evidencing the principles shifts the frame from identity to persuasion and collapses the format into an argument essay.
• Treating the manifesto as an op-ed by anchoring the whole document to one timely issue and one publication’s readership, then appending extra paragraphs of principles - An op-ed advances one clear position on a specific timely issue for a named publication’s readership; a manifesto declares a whole worldview or program not tied to a single question. Anchoring a manifesto to one timely issue produces neither a proper manifesto nor a proper op-ed.
• Listing principles without naming what the manifesto stands against - A manifesto draws its force from contrast - what it rejects as much as what it affirms. A list of values with no stated alternative or critique reads as a mission statement, not a declaration.

Failure modes

• Conviction tips into undifferentiated proclamation - every sentence reaches maximum intensity so the document has no shape, and the reader cannot identify which principles are central - Name each core belief in a single declarative sentence before expanding on it. If the principles cannot be extracted and listed in three sentences, they are buried in rhetoric and the manifesto has lost its spine.
• Breadth collapses into a laundry list - many principles stated at equal weight, so the manifesto has no center and the reader cannot identify what the movement actually prioritizes - Limit core principles to five to seven, then draft a shortened version that cuts to three. If all principles feel equally essential, the manifesto has no hierarchy of commitment and will not be remembered.

Instruction

Write as a manifesto. Open by naming the premise or condition that makes this document necessary -
the thing being declared against or found insufficient. State the core beliefs as principles, not
as evidence-based arguments requiring proof. Name what the movement or writer stands against, then
name the positive commitments with equal clarity. End with a call - what is intended, and what is
asked of those who stand with this position. Do not hedge. Do not present multiple sides. A
manifesto that qualifies its principles has abandoned its form. The reader who is aligned wants
language that commits fully; the reader who is not aligned should at least know exactly where
this position stands.

Template

See the Manifesto template.

Related

Pairs well with

Columnist, Executive, Resolute, Confident

Avoid with

Matter of Fact, Instructional, Skeptical

Often confused with

Op-Ed

Examples

Whether the team should move to async-first standups

The Async-First Manifesto

Why we killed the daily standup, and what we’re committing to instead.

This team ran on one assumption nobody ever wrote down: if something mattered, we said it out loud, together, at the same time. That worked when we were six engineers close enough in time zones to share a morning. It stopped working somewhere in the eighteen months it took us to grow into eleven engineers spread across Pacific, Eastern, the UK, and India, and we kept holding the same 9am Pacific standup anyway - a meeting that lands at 9:30pm for the three engineers who joined us from India. Fourteen minutes a day, and by our own count, about four of those minutes were worth anyone’s time. The rest was status spoken out loud and gone, with no record anyone could search the next time the same problem came back.

Nobody chose this on purpose. We just never stopped to ask who was still awake for it. We are stopping now.

What We Believe

Written is the default, not the fallback. If something is worth saying, it is worth writing down. A status update spoken once and never repeated is not a record, it is a rumor. We choose text first: searchable, re-readable, and there for whoever needs it at 6am or 11pm.

Every time zone is a home time zone. No clock on this team owns the calendar by default. Pacific is not the anchor just because it was convenient back when we were smaller. We design our rituals around the whole team’s day, not around whichever quarter of it is easiest to schedule.

Presence is not the same as work. An engineer who posts a clear, honest update from India before their own 10am cutoff has participated. An engineer who sits silent through a live call at 9am Pacific because the culture expects a body in the square has not, no matter how it looks on a calendar.

Unwritten knowledge gets solved twice. We watched three separate cases last quarter where someone burned over an hour rediscovering a bug a teammate had already found and mentioned once, out loud, in a meeting nobody could search afterward. That is not the cost of moving fast. It is the cost of moving carelessly.

Real-time meetings are for judgment, not status. We are not against meetings. We are against meetings that could have been three sentences in a channel. When we do get on a call, it is because something needs a live back-and-forth: a decision, a disagreement, a design that will not survive being flattened into bullet points.

What We’re Done With

We are done with the standup as theater: a status round-robin performed for the benefit of whoever is in the room, timed to the convenience of headquarters, forgotten the moment the call ends. We are done treating a person’s silence during their own middle of the night as a discipline problem instead of a design flaw. We are done measuring engagement by who was awake to watch a meeting instead of by what they shipped.

What We’re Choosing Instead

We choose a written update, every weekday, in #team-standup, from every engineer, posted in their own morning. We choose blockers that get a name and an @mention instead of a shrug in a room half the team never heard. We choose one protected weekly session for the conversations that actually need two people talking at once, and we choose to cancel it without guilt when nothing needs it. We choose to trust each other’s word in writing as much as we ever trusted a voice on a call.

Starting Today

Every one of us posts by 10am in our own local time: what shipped, what’s in progress, what’s blocked and who can unblock it. Whoever is on call reads the channel and chases every open blocker like it is theirs, because for that day it is. Thursday is protected for whatever actually needs a live room; if nothing does, we cancel by Wednesday afternoon instead of meeting out of habit.

We are giving this thirty days, not because we need more proof that the old way asked our India-based teammates to sit through a status meeting at 9:30pm, night after night, but because thirty days is what it takes for a new habit to feel normal instead of forced. If this is wrong, say so where the whole team can see it, not in a side channel that only reaches whoever shares your time zone. But show up for the thirty days first. Belief without practice is just a nicer-sounding meeting.

This is not a pilot we are quietly hoping fails so we can slide back to what was comfortable for whoever was already in the room. Async-first is not a trial we are running. It is who we are now.

Posted to the team wiki by the engineering manager, on behalf of a team that finally gets to start its own day on its own clock.

Designing a sustainable morning routine

The First Hour Manifesto

Most mornings, by the time my feet hit the floor at 6:30, my day already belongs to someone else. The phone is in my hand before I have said a word out loud, and whatever is waiting there - a Slack thread, a headline, someone else’s version of urgent - sets the terms before I have chosen anything. I did not decide this. I inherited it, the way anyone inherits a default nobody designed on purpose, and I am done treating it as fixed.

This is not my first attempt at the first hour. Three separate times in the past year I told myself the plan was simple: do not check the phone before your feet hit the floor. Each attempt lasted four to six days before it dissolved back into the old default. I also tried the opposite extreme once: a 90-day commitment, a long list, a big promise to myself. I quit on day 11. Both failures taught me the same lesson, and that lesson is the premise of everything below: a rule with nothing built to replace it is not a plan. It is a wish with a deadline.

Here is what I believe now, and what this repo exists to test in public, out loud, on the record.

1. Willpower is not a plan. Environment is the plan. A decision you have to keep winning every single morning is a decision you will eventually lose. The only discipline I trust is a room arranged so the wrong choice is not available at 6:30am. My phone charges in the kitchen now, face down, and that one fact does more work than every intention I have ever had.

2. The first hour is not neutral time. It is the load-bearing hour. Whatever the first 60 minutes decide, the other 23 tend to follow. Waking up is not the start of the day. Deciding what wakes you up is.

3. Small and boring outlasts big and impressive. A 90-day plan is not more serious than a 30-day plan with a review date already on the calendar. It is a longer runway to quit on, unwatched, before anyone notices. I trust the version I can finish over the version that sounds better out loud.

4. A private intention can be renegotiated. A public one cannot, not quietly. I am writing this down and posting it where my spouse, the friend who agreed to be my accountability partner, and anyone who finds this repo can see whether I kept it. That is not a gimmick. It is the actual mechanism.

5. Water, light, and movement come before any screen. Every time. No exceptions written in the moment. Not because the order is sacred, but because an order that bends on a bad morning is not an order. It is a preference, and preferences lose to phones.

I stand against the morning as performance: the version that gets filmed before sunrise, that requires a cold plunge and a leather-bound gratitude page, that turns the first hour into one more thing to be impressive at before 7am. That is not discipline. That is content, and it fails for the same reason my willpower-only attempts failed - it depends on a feeling that is strongest for about a week.

I stand against the idea that one missed morning is a verdict on my character. It is one data point out of 30, and 30 data points are the only honest way to know whether a routine survives an actual life: an actual cold, an actual bad flight, an actual Tuesday that arrives already exhausted from Monday.

I stand against doing this alone, quietly, where failure has no witness and success has nothing behind it but my own say-so.

I stand for a first hour that is boring on purpose: water, light, 15 minutes of movement, 10 minutes with a pen and paper, phone last. I stand for a 30-day window with the review date already fixed, so the design gets amended by evidence and not abandoned by mood. I stand for logging the failures next to the successes, in the same file, in the same handwriting, because a log that only records the good days is not a log. It is a highlight reel. I stand for asking the people closest to this - my spouse, my accountability partner - to say plainly whether it is working, instead of assuming silence means yes.

So here is what I am committing to, starting now. Thirty mornings, in this order, phone last, no exceptions I did not write down in advance. Every one of them gets logged in this repo, including the ones I lose. I will not quietly change the terms alone. If the design is wrong, I revisit it in the open, with the people who signed off on it, on the date already set for that purpose, not at 6:30am on a morning I would rather stay in bed.

If you are reading this because your own mornings feel like they belong to someone else, I am not asking you to copy my hour or my four steps. Copy the method: design the room so the right choice takes no willpower, choose a window short enough to actually finish, decide in advance who gets to see whether you kept it, and then keep it, or log honestly that you did not. The first hour is available to be decided on purpose. Mine starts now.

Choosing Postgres vs DynamoDB for a new service

The Boring Infrastructure Manifesto

Ana Rivera, on behalf of the notification-service team - Lattice Notify Engineering

We almost adopted a second production database this month for reasons that sounded airtight in a design review and stopped sounding airtight the moment someone asked who would carry the pager for it. We did not adopt it. But coming that close, on a team that already knew better, told us this was not a one-time lapse in judgment. It is a pattern, and until now we had never written down what stops it. This is that document.

What we believe

We build for the team we have. Eight backend engineers hold a four-person on-call rotation for everything Lattice Notify runs in production. Every system we add is a system one of those four people operates half-asleep during an incident, not a system judged once in a design review and forgotten. We size our architecture to the roster we have, not the team a future deal might buy us.

We design for the load we can measure. Five hundred thousand notification events a day is real traffic, arriving now. Ten times that is a scenario tied to a Slack partnership that has not closed. A scenario earns a plan and a number to watch. It does not earn a vote equal to the traffic already hitting our servers.

We put a number on every “not yet.” Five million events a day, sustained, is when we reopen the DynamoDB question for the notification service. We do not relitigate a decision from memory in a hallway conversation six months from now, and we do not pretend the decision is permanent either. We write the number down once, and we honor it when it arrives.

We price reversibility before we commit. Outgrowing Postgres costs us three to six weeks of migration we can see coming and plan around. That is the number that let us commit with confidence. We do not choose a direction whose cost of being wrong we have not already named.

We say the other side’s best argument out loud. Marcus built the strongest technical case in the room for DynamoDB’s fit with our access pattern. We said so, in the same meeting where we chose Postgres anyway. A team that cannot state its opponent’s best argument has not tested its own belief, it has only repeated it to itself.

Every decision gets one owner and one document. ADR-0023 exists so Priya, Marcus, Sam, and Jordan all know exactly what we decided, why, and what has to happen before we decide differently. A decision that lives only in a meeting is a decision the team will make again, worse, under more pressure, with fewer of the people who made the first case still in the room.

What we stand against

We stand against choosing infrastructure for the design review instead of the pager. We stand against sizing systems for a deal that has not closed while the load that has already arrived waits for a plan. We stand against decisions that live in one meeting and evaporate the moment someone changes teams. We stand against treating “best tool for the access pattern” as a complete answer, when it is one input into a larger one. We stand against collecting a datastore for every access pattern until the on-call rotation is defending four unfamiliar systems at three in the morning instead of one familiar one.

What we stand for

We stand for operational simplicity as a requirement we weigh first, not a tiebreaker we reach for last. We stand for a documented threshold attached to every deferred decision, not a future argument replayed from scratch. We stand for one owner and one written record per decision, so the team six months from now inherits a reason, not just a result. We stand for naming the argument we did not take, on the record, so the decision can be reopened honestly when the world changes, not reopened out of frustration or forgetting. We stand for choosing the tool that turns our team’s existing competence into an advantage over the tool that turns that same competence into a rebuild.

What we’re committing to

Starting now, every datastore or major infrastructure decision at Lattice Notify Engineering gets this same treatment, not only the ones that happen to involve a database. A written threshold. A named owner. An ADR that outlives the meeting where the decision got made. The counter-argument, stated in full, on the record.

We are asking something specific of the people closest to this one. Sam and Jordan: when you build the notifications schema and wire the on-call dashboard to watch it, treat the five-million mark as a commitment we will act on, not a number we wrote down to end a meeting. Priya: when you lock this Friday, hold the ADR as the actual contract, not a summary of what got said. Marcus, and every engineer who brings the next exciting tool into a design review: make the boring case first, out loud, before you make the exciting one. If the boring case cannot survive being said out loud, it was never the boring case. It was habit, wearing the boring case’s clothes.

We are not only choosing Postgres today. We are choosing, from today forward, to be a team that can tell you exactly why.

Telling stakeholders a committed feature is being cut this quarter

Late Beats Broken

Issued the week we tell four customers that Insights will not ship this quarter.

Every roadmap breaks a promise it made in good faith, sooner or later. The standard response, at most companies we’ve worked for before Meridian, is to manage the breakage quietly: soften the language, leave the old date technically alive on a slide somewhere, ship a smaller feature under the original name, and let the gap get explained in a footnote if it gets explained at all. We are rejecting that approach, in public, this week, with a real example instead of a hypothetical one.

Here is what we believe instead.

1. A commitment belongs to the people it was made to, not to the team that made it. When we cannot keep one, they hear it from us directly and first, in specific terms, not through a status page that quietly changes color or a release note three paragraphs deep. Four customers were given a date for Insights. Those four customers are being contacted directly this week, not left to find out secondhand.

2. We do not ship the appearance of a feature to protect a date. A dashboard without the saved views and scheduled reports our customers were specifically promised is not a smaller Insights. It is a different, worse product wearing the name. Offered that choice, we miss the date instead.

3. A deferred commitment gets a quarter and a real date, not a “soon.” Insights moves to Q1 2027, with a target date of March 13. We are naming it in public because a claim that can later be checked is the only kind worth making now.

4. Partial value now beats total silence until later. Where we can put something real into someone’s hands while the full thing gets built, we ship it, and we say plainly that it is a bridge, not the destination. The CSV export shipping before the end of this quarter is not a smaller Insights either. It is not Insights at all, and we are not letting anyone describe it that way.

5. The team whose tradeoff caused the miss owns the conversation. A mandatory, regulatory-driven billing migration, not a discretionary one, took the engineering capacity we had set aside for Insights. That tradeoff was ours to make. We do not hand the explaining to sales or customer success and step back from it.

6. A directional date is labeled as one until it is closed. We are naming Q1 2027 before the capacity behind it is fully locked, and we are saying so plainly instead of letting an early date sound more settled than it is.

What we stand against is the whole quiet playbook: the roadmap slide that never updates, the rescoped feature still marketed under its original promise, a delay that only shows up in a changelog nobody reads, and “more details soon” used to move our own discomfort about the date onto the person who is waiting on it. None of that is exactly dishonest. It is designed to protect the team that broke the promise instead of the people who were counting on it, and we are done building it that way.

What we stand for is the opposite of each of those, held at the same time: the falsifiable date over the comfortable one, the smaller true thing shipped now over the bigger vague thing promised later, and our own names on the tradeoff that caused the delay, not just on the plan for fixing it.

So, plainly: the CSV export ships before the end of this quarter, built by Dario Reyes’s team, the same engineers who just finished the billing migration that displaced Insights in the first place. Jordan Park is contacting each of the four affected accounts directly this week. Insights is committed for Q1 2027, March 13 is the date we are asking to be held to, and we will put the next update in writing the moment anything about that changes. If you are the one delivering news like this, on any team, not just ours, take the standard for yourself: give the real date, name what people get in the meantime, and put your name on both. We would rather be caught missing a date we announced ourselves than be trusted for a promise we never meant to keep at full strength.

• Maya Chen, Product Lead, on behalf of the Insights team

Getting a new engineer productive in their first two weeks

The Guided Pairing Manifesto

Most engineering teams hand a new hire a login, a wiki link, and an unspoken test: figure it out, and we will decide later whether you belong here. We have watched that test run its course before - three or four weeks of quiet self-navigation, a first contribution that arrives so late it no longer feels like an introduction, a new engineer who is technically productive and still unsure whether anyone would notice if she asked for help. We do not accept that this is simply how onboarding works. It is how onboarding fails, slowly enough that nobody has to admit it.

What we believe

1. Belonging is engineered, not discovered. A new engineer does not stumble into feeling like part of the team any more than a service stumbles into passing its health checks. Both are built, deliberately, by people who decided in advance what working looks like.

2. A name beats a document, every time. A wiki page cannot tell a new engineer which failure mode is real this week and which is a once-a-year edge case nobody has hit since launch. A named buddy can, because a buddy can be asked. Documentation is backup. It is never the plan.

3. The first real contribution is a decision the team makes, not a milestone the new hire has to earn. We scope it before her first day arrives. We do not wait for a new engineer to volunteer for real work, and we do not mistake silence for a lack of readiness.

4. Verification is not optional. “Should be done” is not the same as done. A checklist with no name attached to each line is a checklist nobody owns, and an unowned checklist is where new engineers go to get quietly stuck.

5. Protection is not the same as delay. Keeping a new engineer off the on-call rotation for her first 30 days is not a lack of trust in her judgment. It is sequencing, competence first and pager load second, and we will not apologize for choosing the order.

6. The cost of doing this right is real, and hiding it is a failure of planning, not a kindness to the sprint. A buddy who loses thirty to forty percent of their capacity in week one is not being generous. They are doing the job, and sprint planning should say so out loud instead of absorbing it as a rounding error.

What we stand against

• The read-and-absorb plan that treats a wiki space as a substitute for a person.
• Silence as a design choice: the two, three, four weeks a new engineer spends unsure whether her question is worth someone’s time.
• Measuring a new hire’s first month only by what she produces, as if belonging were a side effect that either shows up or does not.
• Treating this sprint’s velocity as more important than whether the person joining it ever becomes part of it.

What we stand for

• A named buddy from day one, accountable for a checklist that is not complete until it has been verified with the new engineer directly.
• A first real change, scoped and waiting before the new engineer’s first morning.
• Guided orientation before independent work, not instead of it.
• A check-in that asks about more than blockers.
• A name in the deploy log before the end of week two.
• Budgeting the buddy’s time the way we budget any other real cost: in the open, on the sprint plan, not absorbed in silence.

What we intend to do

Two weeks ago, Priya Rao started at this desk with a login, a checklist, and a person, Arjun Nair, whose job was to be asked questions. Three days ago, her first change shipped. No one had to guess when she would be ready. We decided that before she started.

This is not a pilot anymore. It is how Backend Services onboards every engineer who joins us, and we are done treating it as optional. If your team still hands new hires a login and a wiki link and calls it a plan, we are not asking you to take our word for the alternative. Name a buddy before the next person’s first day. Scope their first real change before you know whether they will need help finishing it. Put the cost on the sprint plan where everyone can see it. Then watch what happens in two weeks.

We will not go back to guessing.

Mei Chen, on behalf of Backend Services, Northlane Systems

Writing to thank a mentor who shaped your career

Put Them Forward Before They’re Ready

A manifesto on developing people, and the ten-year-late thank-you folded inside it.

Most organizations develop people by subtraction of risk. They wait until the track record already exists, then hand over the scope that would have produced it, and call the waiting patience. I do not believe development works that way, and I did not arrive at the belief through a training program. It was given to me, without my consent, ten years ago, by my mentor, Dana, who put my name forward for a project I had not earned and then refused to take it back when I tried to return it to her in a moment of real panic. I am done letting that gift sit unacknowledged. This is what it taught me to believe, stated plainly, in public, where a debt like this belongs.

People become capable of a role by carrying it, not by being certified for it beforehand. Competence proven in a simulation does not transfer to the room where the stakes are real. It has to be built there, under the actual weight, or it is not built at all.

Staying close is not the same as staying in control. A mentor who sits in the meeting and says nothing is doing something entirely different from a mentor who runs the meeting through someone else’s mouth. The first is building a leader. The second is building a puppet with good attendance.

Correction belongs after the room has emptied. Said in front of the people whose respect a report is still earning, even a small correction becomes a public demotion. Said afterward, in private, it is only information, and information does not cost anyone their standing.

The exposure of the bet belongs to the person who made it. If the project fails, the question the organization asks is who vouched for this person, not why this person was not ready yet. A mentor unwilling to carry that question was never really offering the opportunity. She was only offering the title.

A pattern that stops with you was never actually learned. It was only performed. If you cannot hand it to someone else and watch it work in their hands, you were imitating your mentor, not practicing what she taught you.

I stand against the co-lead arrangement that grants a title while keeping every real decision in more senior hands, because it teaches nothing and protects everyone except the person it claims to be developing. I stand against the stretch assignment with an invisible ceiling, the kind that looks like trust from a distance and is actually supervision with a longer leash. I stand against correcting people inside the room where they are still building their nerve. I stand against mentors who wait for all the risk to disappear before they will hand off real work, and then call the waiting rigor.

I stand for nominating someone before the case for them is fully made. I stand for staying close enough to be asked a question and not one inch closer. I stand for absorbing the reputational cost of the bet myself, in full, without ever mentioning that cost to the person who is the reason for it. I stand for correcting once, afterward, in private, and only the part that actually needs correcting. I stand for stepping back on the day I am no longer needed, without making an announcement of having done so.

I am not proposing any of this for adoption. I am telling you it is already what I practice. This year I nominated someone on my team for something she did not yet feel ready to lead. I have stayed close. I did not take it back, not even in the week she tried to hand it to me. It is going the way it is supposed to go, which is to say it is hard, and it is hers.

Somewhere in the middle of watching that happen, I recognized whose hands I was using.

Dana, you did this for me a decade ago and never once asked me to notice what it cost you. I am noticing now, in the open, where it can be read instead of assumed. This document is the thank-you I should have written years ago, and it is also the only repayment a debt like this accepts: not gratitude spoken in private, but the pattern kept alive in public, in someone else’s hands, on purpose.

If you recognize yourself anywhere in this, the ask is not complicated. Find the name of the person who put you forward before you were ready. Say it out loud, somewhere it can actually be heard. This practice does not survive on being felt. It survives on being named.

• Sable Marchetti

Reflecting on keeping a discipline of rest

One Day in Seven: A Manifesto for Rest

I spent years treating my calendar as a scoreboard and rest as the moment the scoreboard stopped updating. A productive person works until the work is done, and the work is never done - I did not invent that sentence, but I lived inside it without ever once holding it up to look at. Every rest day I tried collapsed within a few weeks, because I kept questioning my schedule and never questioned the sentence. The schedule was never the problem.

I no longer believe rest is what is left over after the work. I believe it is what makes the work worth doing, and I am done keeping that belief private.

What I believe

1. Rest is not a lighter workday. It is a structurally different kind of day, and calling it “lighter” is how the six days quietly absorb the seventh.
2. One day in seven is either kept completely or it is not kept. There is no “mostly.”
3. The discomfort of stopping is not proof that stopping is wrong. It is proof that stopping is unfamiliar, and for a long time I treated those as the same thing.
4. A rested day cannot be judged by what it produced, including whether it produced good rest. Scoring the rest is still working.
5. A practice that collapsed once has not failed forever. Mine ran six weeks, fell apart under a deadline, and sat unused for eleven months before I picked it up again. The restart counts for as much as the streak.
6. What goes unanswered today gets answered tomorrow, and the world absorbs the delay far better than the mind expects it to.

What I stand against

I stand against the idea that availability is the same as value, and that a message deserves a faster reply than a person deserves an unclaimed hour. I stand against the quiet accounting habit that tallies whether a day of rest was good enough to justify what it cost, the background process that keeps running even on the days I hold the practice, checking whether I am allowed to feel good about it yet. I stand against rest that has to be earned before it is taken, and against the deeper insult of rest that has to be justified after.

What I stand for

I stand for one full day in seven, Sunday, that starts the night before when the laptop closes and the phone goes into a drawer. For letting a message sit unanswered from Sunday until Monday morning without composing the reply in my head while it waits. For treating the first anxious hour as weather passing through, not as a signal that something is wrong. For six days that carry a different quality of attention because the seventh was actually kept. For trying again after failing, without needing the new attempt to look clean or unbroken.

The call

I am keeping this day. Not as an experiment with an end date, not as a reward for a good week, but as a standing commitment that does not have to be re-justified every Sunday morning. I am not asking anyone to schedule their own version. Scheduling was never what was missing. What is missing is the willingness to actually stop, and no calendar invite manufactures that on its own.

If you are running the same accounting habit I have been running, quietly tallying whether your rest is worth what it costs, stop the tally before you stop the work. Take the day. Do not measure it against last week’s. Tell me if you kept it, or tell no one at all. Either way, keep it.

Marking a long-serving colleague's departure

We Will Not Do This Again

A manifesto for institutional memory at Crestfield Operations Issued by Carolyn Marsh, Operations Lead, on behalf of Crestfield Operations - June 30, 2026

In the two months before Howard Thayer’s last day, we tried to pull twenty-six years of judgment out of one person’s head. We got the vendor contacts. We got the incident sequences, the escalation paths, the four utility contacts that had existed nowhere but Howard’s own records. We did not get all of it, and no amount of exit-interview time was ever going to close that gap. This is not a story about Howard. It is a story about a department that let a single point of memory sit in one person for twenty-six years and called it normal operating conditions.

We are done calling it normal.

We believe institutional knowledge that lives only in one person’s head is not knowledge the department actually has. It is knowledge the department is borrowing, on terms it does not set and a timeline it does not control, and the bill comes due the day that person gives notice.

We believe documentation is not a parting gift. It does not belong in someone’s last two months, assembled under deadline by people who are simultaneously losing a colleague and racing a clock. It is a standing practice on a standing cadence, or it is not a practice at all - it is a scramble with better paperwork.

We believe mentoring is a function of the job, not a personality trait we happen to notice in hindsight. When someone spends two decades quietly keeping colleagues from expensive mistakes, that work has a name. The name belongs in how we describe their role while they are still doing it, not in the tributes we write once they are gone.

We believe a filled requisition covers a role. It does not cover a person. We will stop letting a completed hire stand in for an honest account of what this department loses when someone like Howard leaves, and we will say so plainly to the people above us, not just to each other.

We believe naming a gap out loud is a sign this department is paying attention. Pretending the gap does not exist, because admitting it feels like admitting failure, is the only actual failure available to us here.

Here is what we are done tolerating: heroics treated as an operating model, where one person’s memory quietly catches what our systems miss and we call it a strength instead of a risk. Waiting for a resignation letter to discover a single point of failure we could have found on any ordinary Tuesday. Mentorship that goes unacknowledged for twenty years and gets a round of applause only at the goodbye party. A filled job posting mistaken for closure.

Here is what we are committing to instead: the quarterly knowledge capture practice adopted under ADR-0047 is now how this department runs, on a cadence that is not optional and does not wait for anyone’s notice period. Mentoring is now an explicit expectation held by named senior individual contributors, whether or not it comes with a title. We write things down while the person who knows them is still here to check our work, not after. We tell leadership what a departure actually costs in the same conversation where we tell them the position is filled, and we do not let the second fact quiet the first.

So: every team lead’s quarterly entry is due on the cadence, not on the day someone gives their notice. Everyone who was mentored by someone senior on this team, say so, in writing, while it still describes something you can point to instead of something you are trying to remember. And to whoever leads this department after this document is old news: measure what this team knows the way you measure what it ships. For twenty-six years those were treated as the same number, and nobody checked until Howard cleaned out his desk.

The next person who gives this department twenty-six years will retire from a department that already knew what it had. That is the only kind of goodbye we intend to give from here on.

Marking the team shipping a hard, long project

Nothing Broke. That Was the Work.

A companion to the Project Halyard retrospective, and louder than a retrospective is built to be.

The retrospective we filed for Project Halyard says, in its list of what did not go well, that the work was largely invisible to the rest of the organization. That line is accurate, filed correctly, and not enough. A bullet in a retro template will not change how this company sees fourteen months of work that produced zero incidents and zero features. This is the same point, said louder, in a document built for saying things without hedging.

Fourteen months ago we chose to rebuild checkout as a system running alongside the old one, after two earlier attempts to fix it in place had already stalled and been abandoned. The old system has sat in read-only archive since the cutover. The new one has carried this company’s two highest-traffic periods without degrading. Not one customer ever noticed the changeover happen, and to most of the org, that looked like evidence that nothing had happened at all. Nothing happening is not nothing. It is the hardest outcome in this business to produce and the easiest one to overlook. Here is what we believe, now that we have done this once and intend to do it again.

What we believe

A system is proven by the load it carries, not the review it passes. We did not trust our own rebuild until real checkout traffic was moving through it, cohort by cohort, starting at one percent. A design document earns a checkmark. Only production earns trust.

We keep two ways home until one of them has earned the right to be the only way. The old checkout stayed live and fully maintained for fourteen months, ready to take traffic back at any point. We did not decommission it until the new system had carried at least two peak-load periods without incident, exactly as we said we would. That cost real engineering time we could have spent elsewhere. We would spend it again.

A launch date that moves because someone found a real bug is the system working, not the system failing. Dani Rowe called the hold on the March launch when the pressure to hit that date was real and the bug behind it was not yet fully run down. That call cost three weeks. It was correct, and it was not the last hold we called.

We ship when the system is ready, not when the calendar says so, and we do not call that a tradeoff against speed. Holding that line through two slipped dates was the hardest discipline we kept, because nothing rewards it in the moment. Fourteen months is the receipt for building safety and speed at once, not an excuse for how long it took.

The months nobody can see are not lesser work. Shadow mode. Canary ramps. A contested regression-freeze call that Sam Wickfield held on June 9, when every hour of delay felt enormous. A dual on-call rotation that Dev Okonkwo and Marcus Ferreira carried for fourteen months with no real relief window. None of that produces a screenshot. It is the reason the cutover was boring, and boring was the entire point.

What we stand against

We stand against the big-bang cutover with no way back, chosen because a real rollback path is expensive to build and easy to skip when nobody forces the question. We stand against treating a slipped date as an incident to explain rather than a save to credit. We stand against a planning process that only makes room for work it can see, so months of shadow testing lose the argument to a feature with a screenshot. And we stand against calling this kind of work invisible, as if invisibility were a property of the work itself and not a choice about what the organization bothered to look for.

What we stand for

We stand for a rollback path that stays staffed and funded until the day it is formally retired, not a diagram that claims one exists. We stand for naming the people who made the boring outcome possible, out loud: Dev Okonkwo and Marcus Ferreira carried two live checkout systems for fourteen months so that carrying one badly was never a risk anyone had to take. We stand for crediting a hold as loudly as a launch, the way Dani Rowe’s call in March belongs next to the cutover in June, not underneath it. We stand for a planning process that can fund months of nothing going wrong on purpose, the same way it funds a launch week.

What we’re taking into the next one

Project Halyard is not entirely finished. The legacy checkout comes down for good on July 14, once the archive window Sam Wickfield’s freeze bought us runs out. We are writing this now because the next hard, multi-month, low-glory project in this company has not been named yet, and whoever draws it should not have to invent this doctrine from nothing, the way we did.

To whoever picks up that project: keep the old path funded until the new one has carried real load, not until a demo goes well. Name the number that earns your trust, the way we named one percent as a start and two clean peak-load periods as the bar for calling it done. Build the on-call relief rotation before the parallel run starts, not after four people carry it alone for a year. And to whoever funds it: fourteen months of nothing going wrong on purpose is not nothing. Fund the next one like you believe that, before an incident makes you believe it the hard way.

We ship when it is ready. Not when the calendar says so. That is the whole doctrine, and we intend to run it again.

• Yuki Tanaka, on behalf of the Halyard team

Arguing a public position on return-to-office

Where Work Happens: A Declaration

A declaration from the Work-Location Policy Working Group, published for our own people and for anyone outside this company still relitigating the same fight.

Every return-to-office debate happening right now is being run as a war with exactly two acceptable outcomes: everyone back in a chair five days a week, or nobody ever required to share a room again. We reject the premise. A workplace is not obligated to choose between control and abdication, and we are done pretending those are the only two shapes a policy can take.

We spent two quarters inside this argument before we understood what we actually believed: not a compromise position, not a survey of what each camp would tolerate, but a worldview about what presence is for and what it is not for. We are ready to state it plainly and hold to it.

We believe:

• Presence is a tool for coordination, not a proxy for commitment. A chair filled five days a week proves nothing about the work done in it.
• Flexibility without structure is not trust. It is abdication wearing trust’s clothes, and the people who built their lives on top of it deserve better than a placeholder nobody actually decided.
• Talent has no commute radius. Every policy that pretends otherwise is a talent policy in denial.
• A commitment made during hiring is a commitment, not a draft. We do not get to discover, a year later, that “flexible” meant “for now.”
• Trust is built in scheduled proximity, not enforced by constant proximity. A room shared on purpose, on a known schedule, builds more of it than a room that is either always full or never full.
• A policy that hedges to keep every camp comfortable ends up defending nothing. We would rather be argued with than be vague.

What we stand against. We stand against the binary itself, the idea that the only honest positions are full mandate or full abdication. We stand against “flexible” used as a placeholder for a decision nobody actually made, the arrangement that let ambiguity harden into policy by default while two groups of real people built incompatible expectations on top of it. And we stand against calling an honest middle ground a compromise, as though landing somewhere other than an extreme were a failure of nerve instead of a decision made on purpose.

What we stand for. Two anchor days, Tuesday and Thursday, when everyone who can reach a room is in one. Three days that are actually flexible, not flexible with an asterisk: no approval required, no case made for permission. A line held for a full quarter before anyone is allowed to renegotiate it, because a policy that bends in its first month was never a policy. Presence used to coordinate, never to surveil.

Here is what we intend to do. We will hold Tuesday and Thursday the way we said we would, not the way “flexible” quietly stopped meaning flexible. We will build the mentorship structure new hires need to learn this place in two shared days a week, because we are not calling a known gap solved before we have closed it. And we will publish our reasoning, not just our conclusion, so that other organizations relitigating this same fight do not have to start from an empty page.

Here is what we ask of anyone who stands with us. Protect the anchor days from becoming optional, one quiet exception at a time, the way the last policy drifted into nothing. Show up on the two days that matter so the other three can stay genuinely, defensibly yours. And if you are a leader still choosing between the two extremes on offer: stop. There is a third option, designed on purpose instead of inherited by accident. We are living proof it holds.

We did not split the difference. We built something else entirely, and we are done being quiet about it.

Announcing a new product to an outside audience

The Tidemark Manifesto

Why we built a roadmap tool, and what we refuse to build next.

Every roadmap meeting we have ever sat in got decided the same way before anyone opened a scoring spreadsheet: whoever spoke last, loudest, or most recently won the room. Nobody in that room was lying. Nobody was working from the same set of facts either, and a prioritization framework cannot fix that on its own. We do not accept that this is simply how roadmaps get made. Tidemark opens to the public today, June 30, 2026, and we are saying so out loud.

Here is what we believe, on launch day and after it:

• A roadmap is only as honest as the evidence underneath it. Weigh requests by revenue, by frequency, by strategic fit, whatever your team has agreed matters, but do that work on one shared set of facts first. A framework cannot rescue evidence that was already in dispute. It only dresses the dispute up as math.
• Small teams deserve the same rigor as teams with a dedicated research function, without needing to build one. A team of three people who own a roadmap should not have to hire a researcher, sign an enterprise contract, or wait a quarter to know what their customers have actually been asking for.
• A roadmap only one person can see is not a roadmap. It is a private opinion with a deadline. Ranking the work is half the job. The other half is a link any stakeholder can open and trust without requesting an account first.
• The workflow comes before the feature list. Gather, rank, share - each has to work cleanly on its own before we add anything else to it. A long list of integrations does not fix a planning meeting that opens by re-litigating whose feedback is real.
• Understanding a product should never cost a sales call. A prospective user should be able to see exactly what a tool does before anyone asks for their calendar, their card, or their patience.

We stand against the private spreadsheet every team keeps and nobody names, the one whose sole purpose is settling arguments about what customers actually said. We stand against planning meetings that spend their first twenty minutes reconciling whose list is accurate before any real decision gets made. We stand against software priced and built for organizations with dedicated product operations teams, then sold down to teams of three who will pay for infrastructure they will never fully use. We stand against demos gated behind a pitch, where seeing the product requires first surviving the sales call meant to sell it to you.

We stand for one connected, ranked, shareable view, built from every source your team already uses: the support tickets, the chat threads, the ticket trackers, the survey exports, the notes from your last customer call. We stand for a free plan open to any individual, no time limit and no card required to start. We stand for a 20-minute walkthrough anyone can book directly, with no sales call attached to it. We stand for pricing anyone can read on the page: free to start, $29 per month for up to 15 seats, and a custom plan reserved for organizations that actually need SSO and audit logs, not offered as an upsell to teams that don’t. We stand for shipping the core loop (gather, rank, share) before a single integration gets added to it.

Tidemark is open to any team starting today at tidemark.io. If your last planning meeting spent its opening stretch establishing whose feedback was real before anyone decided anything, we built this for you, and we are asking you to try it. Start free. Book the walkthrough if you want to see the product before you commit to anything. Bring the spreadsheet you have been keeping to settle arguments about what your customers said. We intend to make it unnecessary. Hold us to that. Do not just take our word for it.

---

Renata Okafor, Co-Founder and CEO Marisol Veen, Head of Product Tidemark, June 30, 2026

A personal year-end reckoning with a difficult year

The Early Word: A Manifesto

In March, the Meridian initiative dissolved when the funder that had backed it walked away, and eleven people who had given eighteen months of volunteer work got a thank-you that did not come close to matching what they were owed. By August, Celeste and I had stopped speaking, after six years of closeness had drifted into a silence I spent the spring calling something gentler than what it actually was. A message from Theo has sat unanswered since April. Three different relationships were owed an early word this year, and in every case I delivered a late one, or none at all.

I do not believe anymore that managing how a hard thing looks is the same as managing the hard thing. I believed some version of that for most of this year, and it cost me a coalition’s trust, six years of ordinary friendship maintenance I let lapse, and eight months of a message I was fully capable of answering the week it arrived. I am done confusing the story I tell about a situation with the situation itself.

What I believe

1. A hard truth belongs on the timeline the facts set, not the timeline that is easiest to deliver. Waiting for a better moment to say a true thing is a way of hoping the thing stops being true before I have to say it.
2. Optimism that requires withholding a material fact from the people who trusted me is not optimism. It is a story I am telling myself first and everyone else second.
3. Giving someone space and avoiding a conversation I do not want to have are not the same behavior, and I do not get to use the name of the kinder one to cover for the other.
4. What a commitment costs when it fails is not proof the commitment was a mistake. It is only proof of what full commitment costs, and those are two different findings. This year I nearly let myself believe the second was the first.
5. Silence is a decision. It does not stop being a decision just because it did not feel like one while it was happening.
6. Naming what I do not understand about a loss is more honest than manufacturing a lesson I have not earned. Not every hard year converts into a growth story, and I no longer need mine to.

What I stand against

I stand against narrative management dressed up as loyalty - the instinct that kept a coalition of eleven aligned around optimism when some of them needed a harder truth weeks before I gave it to them. I stand against calling distance independence, because independence sounds like a choice I made on purpose and distance sounds like what it actually was. I stand against letting an unanswered message get harder to answer with every week I leave it, and then treating the difficulty as a reason instead of the symptom it is. I stand against reframing a genuinely bad year into a tidy account of hidden gifts. Some of what happened this year was simply a loss, and I do not owe anyone, including myself, a silver lining before I am allowed to move forward.

What I stand for

I stand for delivering the hard version of an update on the date the facts require it, even when a softer version is available and no one would catch the difference right away. I stand for telling the people who have trusted me with their time what I actually know, before I know how the story ends. I stand for full commitment to the next project and the next relationship, on the same terms as before, because this year is evidence of what full commitment can cost and not evidence that it should be rationed going forward. I stand for a written account, not a summary, for anyone who gave real time to something that did not work out. I stand for answering the message before it is comfortable to answer, not after.

The call

Theo gets an answer in January. Whatever discomfort that conversation holds is smaller than nine months of silence has already cost. The Meridian coalition gets the complete account in February - what happened, what the decision-making looked like from the inside, what I would do differently - not a paragraph and a lessons-learned bullet, because eleven people do not owe me the easy version of their own year. I am not starting the next large initiative until both of those are finished and at least one relationship I let go under-maintained has been actually repaired, not just intended to be.

If you are one of the people who might work with me again, or might trust me with something that matters to you, here is what I am asking in return. Tell me the hard thing early, before I make it easy for you not to. Do not let me manage you with a version of events I have not fully told myself yet. Hold me to the two dates above, and to every date after them. I would rather hear that I am repeating this pattern than be allowed to repeat it quietly. That is the whole of what I am asking anyone to give me for what I am committing to here.