In the 8 months since I started my current job, I've switched through 3 different directors, and my team has worked on 5 unrelated projects. All the decisions make sense, but the pace of change is crazy.
This is by far most chaotic time I've ever experienced in my 15 year career.
There are a few things happening at once:
You have CEOs and executives with "AI psychosis”, whose expectations about what's possible keep rising.
You have a lot more competition everywhere, and the (hopefully) great software you've built doesn't feel like a strong moat anymore.
Every company can do a lot more now. The optionality that keeps increasing means less focus, and more changes of direction, more decisions, more deliberations.
Layoffs everywhere (but the delivery expectations keep rising, as I shared in “your team got cut. Your scope didn’t”). Just this week, layoffs in: Wix, NetApp, Webflow, Meta, ClickUp. 100k+ tech jobs lost in 2026 alone:
There's the “great flattening” of middle managers (which even has a Wikipedia article). Jack Dorsey thinks for some reason that ideally all 6000 employees would report directly to him, and some people listen.
Anyway, quite a tough time for EMs, but that’s not news for you.
What I feel most people miss is that the succeeding right now as an EM requires much more non-technical skills than ever before:
More “AI skills” is not what you need to survive
Let’s look at each one of the causes of this chaos:
AI psychosis and rising expectations
If you haven’t received a message in the style of "I built this feature in an hour in Lovable, can we add it to our app tomorrow?", you probably will…
The solution is not to go crazy and try meet the impossible expectations. It’s to find the right ways to connect the non-engineering people to reality. I was told too many times by a senior leader (usually from product): “I still can’t understand, shouldn’t this take just days with Claude? Look at OpenClaw, a single person built it!”.
To handle these conversations you need to know what to say to whom, how to push back, when to wait. Getting your whole org aligned on what is realistic and what not is quite a challenge, it’s one of your most important jobs right now.
layoffs
By far the easiest way to find a job is through personal connection (if not, then on my side project, leadjobs.dev!). Engineers you managed who liked working with you, peers, previous managers (I found every job I had this way).
Of course you need some hands on experience in today’s market, I’m not saying to be completely behind. Still, if you spend years being a manager people want to work with again, you are much more likely to find a job.
The great flattening
Around 50% of the first-line managers reading this manage 9 to 15 people.
You can of course just ignore your engineers and let them survive by themselves, but then your chances of a happy and productive team are usually much lower. I strongly believe that every human being can benefit from a good manager who’s there for them.
Doing that for a large amount of people requires a lot of mental effort. Knowing who needs you when, who’s frustrated, who is not challenged anymore, what doesn’t work in the team dynamics, etc.
Yeah, AI can help here of course (like I wrote in “the EM memory crisis” article), but that’s just a small part.
Reporting up.
There's a good chance you report to a director who is busier than ever and with less attention to give you than ever before. Getting what you need from them, and involving them at the right moment, is not really a “prompting” problem.
You need to:
Prove yourself
Make sure they trust you
Have them take you seriously
Make sure they involve you in their decisions
Advocate for your team and your own agenda
And all that with minimal 1:1 time.
So if not more Claude, than what can I do?
There are 2 skills I try to master, and one very simple and effective method:
1. Improve at “compression” and “decompression” of information.
Compression is being able to take your experience/feeling/thought and compressing it into the right few sentences for the other person. Articulate yourself so the other side fully gets you, while in parallel:
Minimizing the cost (needing a few minutes instead of 1 hour meeting)
Avoiding confusion (understanding what exactly the other side thinks)
Decompression: being able to receive information effectively - get as close as possible to what the other side tried to transmit. Make the other side feel heard, and walk out of conversations feeling that both sides understood each other.
2. Learn how to persuade
Persuasion is a must skill right now:
We persuade executives to approve new projects.
We persuade our leadership to give us more resources.
We persuade reports to tackle tasks that may not excite them.
We persuade peer managers to lend us their help in cross-team projects.
We persuade our cross-functional partners to move deadlines when they might overwhelm our team.
…and many more examples
In an article last year I went over 5 effective methods:
The Nemawashi Method
The Decoy Pricing Method
The Reverse Psychology Method
The LMDTFY Method
The Engineered Serendipity Method
The main method to improve
It’s very hard to judge the improvements in the 2 skills above. The only effective method I found is to insist on high-quality feedback from senior leaders.
As I wrote in the curse of good managers (excuse me for the overload of references, this whole topic is dear to me), when you do your job reasonably well as an EM you often don't get useful feedback. Your engineers don't say anything, and your manager isn't involved enough in the day to day.
So in addition to trying to get the maximum from my own manager, I ask senior leaders directly. Very often their ‘bird-eye’ view can provide valuable feedback.
It can be small: after a recent short demo I gave the company, I asked the director of product for feedback on how my message was received (I was too unfocused, tried to cover too many things - awesome feedback!).
It can also be bigger. I recently asked a senior executive for feedback, and his first response was "you're doing a great job, keep it up." But then I focused him: do you feel I know how to push things the right way and get my agenda through? I felt like I started a bit on the left foot here. And he said, "I wouldn't say left foot, but definitely things you can improve, schedule some time for us when I’m back from US?"
YES. That was the goal (I’ll let you know once the conversation happens - a very recent example!).
Discover weekly
I got fired from my SWE job at Stripe a month ago. Loved this honest take about the departure, and especially about his experiences since then.
You’ll lose your job in 2027. A contrarian take by Elena Verna (Lovable’s head of growth). I don’t agree with many parts of it, but it was still a great read.
AI’s impact on software engineers in 2026: key trends. The pragmatic engineer is my favorite newsletter, and this is one more great deep dive.
