2026-10-14 08:00:00
Every few months someone declares that management is dead. Big Tech is in the middle of a Great Flattening, and there are companies that have a single Director managing 80 engineers and a fleet of agents with nobody in between.
I think it’s the opposite. Management is becoming more important as AI changes how businesses work, and the job itself is shifting faster than most managers have noticed. Let’s talk about what’s changing and how to deal with it.
Here are the biggest themes of what is going on:
Two things are going on in the collaboration world.
First, people are becoming less collaborative. AI has this magical way of convincing everyone that they are auteurs who just need to be unblocked so they can realize their vision. Less sardonically - people spend all day collaborating with an LLM, and it’s making them less collaborative with humans; the sycophantic, optimistic nature of LLMs are megalomania fuel that even the most even keeled people are susceptible to.
Second, the need for collaboration is rising. Producing stuff got cheap. Reviewing, integrating and coordinating that stuff did not. A recent Harvard study of ~700 companies found that AI coding agents pushed pull request volume up 23%, and the time to get a PR reviewed and merged up 49%.
This turns your company into a factory where every station is running at full output and nothing ships. Everyone does their job, and nobody else is doing theirs. This leads to the “cleaning up the Airbnb after a party” effect, where each person individually feels like they’re doing more than their fair share, while everybody else isn’t pulling their weight.
The answer here is unsexy - you must have crisp, well-defined expectations of collaboration. That means things like target SLAs on response times for high volume interaction modes (like pull request reviews). These metrics need to be measured and staffed, and most importantly made someone’s actual problem to solve.
People moved so slow in the beforetimes (say, 2022) that you could course correct investments in real time. Nowadays if you don’t actually know what your goal is in advance, you might spend $100k in tokens before realizing nobody asked for, or wants, this feature.
The answer is again unsexy - make sure everything has a clear line of sight to business value. Metrics should be coming back in a big way, after years of people cynically citing Goodhart’s law as a reason not to have any.
The catch is you have to measure outcomes and not activity. Amazon built an internal leaderboard ranking engineers by AI usage, and people promptly started pointing agents at pointless tasks to climb it. Goodhart would be proud. Measure things as viciously close to actual things that get reported on a 10-K as possible.
Managers are confused about how to manage performance in the age of AI. Should the bar be raised? Did you do it or did the AI do it? Etc.
The short playbook is:
What does a raised bar actually look like? For a senior engineer, it means the project that used to take three people now takes one person with a clear spec. It means “I’m blocked on the frontend team” stops being an acceptable answer for a small change you could make yourself. It means the design doc shows up with a working prototype attached.
Ownership cuts both ways and you own what you ship. “The AI wrote that part” is not an excuse for a bug, the same way “the intern wrote that part” never was.
People have gotten a lot of credit over the last stretch just for doing more stuff with AI. Before 2023, there was a strong correlation between volume of output and being good at your job. AI has completely shattered that correlation; it’s probably now close to 0. It’s time to rebaseline and be clear that more and higher quality work is expected, and set expectations that the baseline will rise with every model improvement.
Can docs be written with AI? What counts as slop?
New cultural standards are needed in the AI era, and there is strong debate on what is right and what is wrong. In my opinion, many of these debates are emotional. If you want one rule to follow it’s this - ignore the genesis of output, criticize the content of output.
“You used AI” is a tone deaf and fast dying criticism. Set a culture where criticizing outputs on their own merits, not on their origin, is the norm. This will help people actually understand and articulate what good/bad looks like, and also avoid an anti-AI bias that is increasingly ludicrous.
AI multiplied everyone’s output. Management is the job of pointing that output at something that matters. None of the fixes here are exciting. SLAs, metrics, a higher bar and clear standards are mostly what good management always was. The difference now is that when you skip them, you find out faster, and it costs a lot more.
2026-09-30 08:00:00
As part of writing Stay SaaSy we speak with and advise a fair number of startups, and one of the most common topics of conversation is how many people they should hire.
The most important factor to consider when expanding a team is that the usefulness of incremental hires is not linear. Sometimes an additional person helps a lot. Sometimes they don’t help at all, due to structural overlap or the quality of the hire (and keep in mind that for some roles, quality is much less reliable). The key point is that the utility of a marginal hire varies dramatically depending on role and whether you’re crossing a key breakpoint in terms of team size. The value of a new hire can even be negative.
And there are a few important breakpoints to consider on the journey of growing your team. These might seem a bit obvious, but the important realization is that at these breakpoints hiring decisions are much more important, so if you keep them in mind as hard thresholds you’ll make better scaling decisions.
Sometimes you just need a single expert on a team who gives you a key capability. There are several different functions where just having one individual will make your whole group much more productive:
For these 0-1 hires, the most important thing to do is not screw it up. A great designer will let you move 10x faster at 10x higher quality; a bad designer will mess up your whole product. A good data person will set up your foundations to scale for years; a bad data leader will set you up for overhauls every 24 months in perpetuity.
But often, confusingly, adding a second person in a function makes things a lot worse. For example:
In a few other cases, adding another person can make things a lot better. For example, it’s often very helpful to have 2 sales team members (classic post on this), or 2 support team members, because they’ll keep one another honest about pace. Sales and Support are highly dependent upon product-market fit to be successful; as a result, both sales reps failing or both support reps struggling often means that the product is insufficient; whereas only one failing usually points you to what the correct bar should be for that role at this moment in time.
For these types of hires: don’t wait too long to hire your second person, especially if the first one is doing well.
Most people cannot manage more than 9 people effectively. The canonical advice on how many people you can manage is 7 plus-or-minus 2 (unsurprisingly, 7 is the median amount of items that you can store in your working memory).
Once you cross into the zone of 7 => 10 direct reports for one manager, it’s very common to see things fall apart. Typically, the dominos fall in this order:
The general rule of thumb: aim to load up all of your managers to their maximum carrying capacity of ~7-10, as each incremental hire will increase velocity, reduce key-person risk, and expand coverage. Past that, tread extremely carefully.
There is a movement going on right now to load managers with significantly more direct reports than before – in some cases that we’ve heard of, 50+. I think that it’s unlikely that advances in AI will be able to dramatically increase the number of direct reports that one person can responsibly manage. Ultimately, even if we have futuristic technology, the neurological hardware that we use for tracking relationships (an essential part of management) is still firmly rooted in 2026 evolution. AI might increase managers’ ideal maximum carrying capacity to 12 or 13; I personally doubt that it will increase it to 50.
Transitioning from ~140 to ~160 employees takes you across the famous Dunbar number, which sits at around 150 people. In essence, the Dunbar number is the maximum number of people in a group that still allows for family-like cohesion. That is, a group of 149 people can operate as a family while a group of 151 people needs to operate more like a corporation (or an infantry battalion, or a bunch of people on a cruise ship, or a prison). Coordination and communication are very easy below the number; they’re very hard above it.
We can debate the exact scientific reality of the Dunbar number, but I will say two things:
In very broad strokes, below the Dunbar number marginal hires typically speed you up because they add resources without dramatically increasing coordination overhead. They aren’t free in terms of overhead, but they’re relatively cheap. You can still yell across the office, or text people, and everyone kind of:
Beyond the Dunbar number, things get crazy and you need to start architecting your organization more like the Navy and less like a sports team.
Generally, my recommendation to companies as they grow past 50 people is to aim for a hiring trajectory that lets them stay below 150 people for as long as they can. This is especially true with advances in AI that give team members more leverage. Past 150 people you are simply going to get less productivity out of each marginal hire, and a team of 149 armed to the teeth with great tools can carry even a fairly complex business very far.
Additionally, once you’ve crossed 150 people, you can and often should hire faster. You’re now fully going to need a heavier structure of internal communications, operations, and other functions to keep everyone rowing in the same direction – you might as well just invest since you’re already incurring the full coordination penalty.
2026-09-20 08:00:00
To have a positive career you need to be appropriately recognized and rewarded for your work.
But there are a lot of ways that credit and/or rewards can get missed. Often it’s something unintentional, like your boss or some other executive simply didn’t realize what you accomplished, or didn’t understand the right way to compensate you for the value you’re adding. Sometimes it’s more nefarious, and someone is trying to take credit for your work or otherwise keep you down. Either way you need to drive defensively - taking actions that allow you to make forward progress even if people around you are negligent or malicious.
All of the techniques below can be approached in a benign way, but I realize that they may come across as cynical. There is a less cynical set of takes at the end as well.
Many organizations basically only know how to value you from what you announce publicly. If your role doesn’t own an explicit topline number for the business, it’s shockingly easy for people (including your boss!) to judge your work performance almost entirely on what you share with them.
This truth often gets lost in tech because many of the most important roles have visible and quantitative outputs (particularly engineering and sales, but also product management, design, and marketing roles that drive pipeline). For the rest, however, you’re basically relying on the last email you sent that your boss saw.
As a result, you need to be visible in a way that isn’t obnoxious or wasteful. To do this, you should pick 1-3 important, public announcements that you always make. Some examples:
Also, in many cases you have to send or speak the note, not your team. “They’ll know it was me” – they likely won’t. Nobody else is paying attention to you, and nobody else cares about you. Own the message, push it gently down people’s throats, and let your team own a different sub-message if they want to.
Other rules: Whatever you’re sharing has to actually be important, or else you’re like those free newspapers full of ads at the coffee shop. And if you send out the update, you are on the hook to really own it and answer questions. Which is kind of the whole point.
If you’re going to have a high-visibility job, like most execs, you can never let yourself get punked: having someone challenge you in front of a public audience and letting it slide.
The number one way that people punk you at work is by lying about you and what you did. Common examples include:
The problem with getting punked is that it makes people think they can mess with you anytime they please. As a result, every time that someone lies about you, you need to politely (or if necessary, not so politely) push back and correct the record. The ultimate goal: You want to be a corporate rattlesnake – spend your whole day chilling in your cave (never lie about others), and if anyone comes after you, rattle once for a warning and then bite them in the face in front of their friends and loved ones.
Companies operate on an ultra-mild version of prison rules. Life operates on an ultra-mild version of prison rules. Unfortunately, that means that sometimes you need to lean way into conflict when someone challenges you in public. This works unreasonably well 99% of the execs who will lie about you are bullies who fear actual confrontation.
There’s a certain personality type that acts very unprofessionally if they think that you got to your position through luck. If someone is undeservedly doing “better” than them, they’ll behave strangely at minimum and in some cases actively sabotage them in order to settle some sort of score.
Doing “better” is personal and can be whatever metric they’ve chosen, but it’s typically some combination of title and lifetime earnings / current compensation, moderated by age. As in, I was a VP with $500k at your age, and you’re a COO making $2.5m.
A few situations that I’ve seen increase this risk dramatically:
The best solution is to simply never let someone who thinks that you got lucky manage you, or control your title / compensation. That’s a pretty hard variable to control, so some of the better secondary ways to avoid this issue are to:
Lastly, you need to maintain leverage at all times. Leverage mainly comes from one place: your ability to walk a valuable asset (your labor) out the door. As a result, for better or worse, it’s usually best to both be very good at your job (obviously) and then make it clear that you could walk, and that you even consider it from time to time.
There’s a balance here. You don’t want to be the dude who’s constantly throwing a tantrum and threatening to quit. So here are a few concrete steps you can take to keep em a little bit nervous. All of these ideas assume that you’re a good-to-top performer in good standing:
I also recommend that you regularly take an interview or two to build your own confidence. You want that “I’m wanted” glow on you at all times.
Of course, if you’re a leader at a company (or the CEO), there are different takeaways. You don’t want your team to feel that they need to follow this advice and drive defensively all the time – you want them driving normally, on the same highway, ideally in unison, as fast as they possibly can. And there are ways to facilitate that:
2026-08-16 08:00:00
As companies get larger, scheduling a decision starts to take longer than making it. Decisions that used to get alignment in an hour now “can only happen three weeks from now, once Michael and Melissa are both back.”
The problem goes deeper than big decisions. Daily operational decisions on things like on-call can become impossible to make at all. Teams that need rapid alignment end up in Slack threads that linger and die before end of day, right before the same incident happens again overnight.
The answer is blindingly simple: book recurring time that exists only for as-needed, cross-functional meetings.
I’ve seen this exact setup take an organization from Slack-thread purgatory to crystal-clear daily alignment on operational next steps.
While the meeting format is simple, keeping it alive takes skill. The necessary conditions are:
It’s often the only slot your most senior people all hold at the same time. On canceled days, it becomes prime real estate for ad-hoc 1:1s.
The irony in many cases is that people often have “Do not book” private blocks on their calendars, they’re just at separate times. This allows people to just sync up a portion of those for collaboration.
2026-08-12 08:00:00
Imagine that you’re in a board room, locked in an intense negotiation. Maybe you’re selling your company, maybe you’re signing an NFL contract, maybe you’re fighting a legal case. During a negotiation, there’s a lot of meaning to what side of the table you’re on. Most significantly, everyone on the same side of the table is on the same team – we might not like each other, we might not always disagree, but we’ll close ranks and look out for one another if needed because we’re representing our shared goal more than ourselves.
The cardinal rule of meetings with your direct reports is that you are always on the same side of the table. From the moment you walk in to the moment you leave, every single thing that your team does is a reflection and extension of you:
(The only exceptions are for cases that are truly insane, like your direct report assaulting someone or getting naked on a table. In that case you’re of course permitted to physically defend yourself; you don’t need to get naked and assault people alongside them)
Your team isn’t going to be perfect. If you have a large enough team, from time to time they will mess up while you’re in the room. But when this happens you can either handle it well, or poorly.
Here are examples of what to do when someone on your team is messing up. These approaches avoid over-defending your team, which would shield them from autonomy and consequences, but also avoid turning on them as part of the angry mob:
Finally, you should commit to your team that you will always be on their side of the table in the meeting. A simple statement like “if this gets heated, I will help to navigate the conversation” sends the type of unambiguous signal of support that people often need to hear.
For a real-world practical example, I once had some mid-level team-members (Senior PMs / Engineering Managers) who were asked to present their product to a very sharp, incisive, but (potentially) intimidating board member. A few steps that I took:
Sports teams have not forgotten the ancient rules of teamwork even as the corporate world has totally abandoned it. Watch a press conference after a team loses and it was obviously one guy’s fault: teams will go well out of their way to avoid blaming their teammate, no matter how obvious their failure was, because they know that blaming your teammate doesn’t absolve you. And that’s in spite of the fact that professional sports teams are famously savage about managing performance, reliably cutting or trading players without warning.
There are practical reasons to close ranks. Blaming people both eliminates any hope of them redeeming themselves, and more importantly instills a culture of fear that impacts future performance. It’s the curse that keeps on taking.
It’s also simply a matter of honor. People leave their team’s side of the table out of a misguided grasp at self-preservation. But people aren’t that dumb. If someone messes up, that reflects directly on their manager too. By staying on your team’s side of the table, even if it’s uncomfortable, you’re showing that you have honor. And in the long run, the best people only want to work with other honorable people, because they’re honorable themselves.
2026-08-05 08:00:00
Everyone knows that lying to your team is bad. Only a complete sociopath would argue that lying to your team is a positive, happy activity. But there’s a particular form of lying that many of the most experienced, compassionate, and honorable managers indulge in all the time: overpromising, and in particular overpromising career rewards. When managers are under pressure, and especially when they’re under pressure to retain a high performer, overpromising is one of the most common tools to reach for. The industry is rife with promises that someone will get a promotion, a new role, a large bonus, 5 more days of vacation, or any number of goodies, large or small.
Overpromising carries a terrible risk/reward tradeoff. Here’s how it goes down:
At this point, you have a very serious personnel problem on your hands that is almost certainly much worse than whatever you had at the start.
How overpromising actually goes
Decent managers routinely overpromise because if you’re an optimist with a can-do attitude, overpromising doesn’t feel like lying at all. You tell someone “I’m going to get you that promotion,” and you really do want them to get promoted. And you’ll be extra tempted to utter a promise that you can’t quite guarantee because you’re in one of the most stressful situations that a manager can face: a high performer on your team is distressed and upset, and you want to give them the confidence to stay engaged on your team. So you say the words.
But the words are a lie. You often can’t guarantee that you’ll get them promoted. Perhaps you are a department head, and you control all of the budgets and promotion decisions within your group, so you feel confident promising Jeff the promotion to Director. But what if:
All of these things can and do happen.
Even though overpromising doesn’t feel like lying, it will absolutely be interpreted as lying by your team. Contrast it with a much more mild form of deception: omitting the whole truth, for example by failing to mention that you’re fielding acquisition offers. Despite the fact that this comes from a more deceptive place (you technically really are intending to deceive your team in order to change behavior), people are generally much, much more understanding of merely unrevealed plans. But any untrue statement leaving your mouth, no matter how well-intentioned, runs the potential of causing a huge problem.
The solution is simple: Never, ever overpromise. Never let happy promises leave your lips until you are absolutely, 100% certain that you will be able to follow through on them beyond a reasonable doubt.
Of course, you don’t need to run your team like some sort of psychological experiment where you never share anything positive about the future. The way to navigate the landscape feels weird at first, but you should caveat any forward-looking career promises like you’re a paranoid lawyer and/or running down the list of scary-sounding disclosures on a pharmaceutical advertisement. I have literally said “I’m gonna try really hard to do X, but I need to add a disclaimer because I don’t ever want to make a hard promise I can’t keep.”
Here are some common fixes to lines that feel totally normal to say, but run a very real danger of overpromising:
Avoiding overpromising doesn’t mean that you can’t be sincere with your team. If you feel an emotion, like the fact that someone is really great and you hope they get promoted, it’s usually okay to show it within reason. Emotions that you really feel are authentic by definition, so they don’t run the risk of letting a lie leave your lips.
Your word as a manager is your bond. And if you want to be a good manager you can never break your word, especially to people you manage.
Management is an asymmetric relationship where you hold the power to enable someone to put food on the table, so your word is sacred. It’s better to be honest and painfully disclaim future looking statements than to risk being an optimistic, confident liar. And in the long run your team and your blood pressure will thank you for it.
