Serhii Vasylenko

You can be the smartest person in the room, or the person who made the room smarter. At Staff level, only one scales. devdosvid.blog/2026/02/19/pos…

Engineers describe what's wrong with the system. Directors need to hear what's at risk for the business. Three questions to force the translation: 1. What slows down or breaks because of this? 2. How often? How much time lost? 3. What can't we commit to because of it? If you can't answer #3, the item won't get prioritized -- no matter how painful it feels.

@BentoBoiNFT What if I told you that you never had to buy one to run OpenClaw too… 😀

Claude just killed OpenClaw with this update You no longer need to buy a Mac Mini I say this as someone who owns a one lol

@fredrivett The SOC-2 part is the interesting question. If "review" keeps meaning "a human looked at it," then review becomes the one bottleneck that everything else speeds past.

@vasylenko oh yeah that’s fair but in that case i think we just switch branch on the root repo, don’t need worktrees really? but having a command centre like conductor will still be valuable also i think human reviews will continue to decrease, though soc-2 compliance may block that

as a very happy conductor user who’s wanted this feature more than any, it’s kinda ironic that the faster ai responds the less we’ll need worktrees when ai responds near instantaneously we’ll multi-task less, become single-threaded again and only switch tasks to take a breather

@lennysan When resources meant people, asking for more forced you to justify the problem. Compute removes that friction -- which also removes the signal that the problem was worth solving in the first place.

The new “how many resources should we put on this project”

@GergelyOrosz @zeeg The downstream of this -- these takes eventually become OKRs. "Increase AI adoption" shows up on a planning doc, and the platform team has to figure out what that actually means in practice.

It's always been like this Most engineers believe what they see, and what works today Those who are not engineers eat up the hype/projections And then we have investors hyping up investments, influencers realizing they need hyperbolic takes to get traction so they do that. And social media amplifies these extreme takes... just how it works Great question why people trust others who are not even using it btw

Why is it everyone with an absurdly futuristic AI take is someone who - as best I can tell - doesn’t work on (and often never has) real software that has real users and real requirements? More so, why do you trust them?

@rseroter I'd add one: if you define the problem well enough, the best solution rarely comes from you. Took me a while to accept that as a win instead of a loss.

What are the unwritten laws of software engineering? Those things that don't have fancy names or get baked into keynote titles? Here are seven from Anton that definitely resonated with me ... newsletter.manager.dev/p/the-unwritte…

@QuinnyPig @awscloud At $30/hr AWS is explicitly pricing it as labor, not tooling. Once you compare it to a person, you start asking questions the agent can't answer -- like who's on call when this change goes sideways.

Looks like the @awscloud pricing files briefly showed pricing for DevOps Agent. Going from "free in preview" to "$30 an hour" is uh... something, all right. Either a bug in how they plan to price, or someone's about to have their gossamer dreams meet the iron bar of reality.

@talraviv The telling part: the "kill half of it" PMs I've worked with are the ones most excited about AI. They see it as leverage. The organizers see it as a threat.

@talraviv From the engineer side, I've worked with both kinds of PMs -- the ones who organized my backlog, and the ones who killed half of it. AI is going to be brutal for the first type. The second type just got way faster.

Claude is a better PM than me. It's time for me to give away my Legos. I spent Sunday morning building with Claude Code, and it did “the PM thinking” better and faster than I would have. I procrastinated fixing Familiar’s onboarding way too long. So many insights built up from watching people, and I didn’t know where to start, so I wanted Opus 4.6’s help. I brain dumped a really messy doc: the kind I would never give to a human teammate, but that an enthusiastic, unfocused CEO might slack me on a Saturday night (we’ll come back to this analogy). You’ve heard the next part a million times: Opus did a fantastic job blah blah blah, and it even created the issues in Linear, prioritized, with tight descriptions, organized by neat milestones. Agentic agentic insane insane. But enough about AI, let’s talk about me. I got to be a cross between a caffeinated product designer and a salesperson with “tons of ideas for the product.” That was WAY more fun than PMing! Meanwhile, AI structured my thoughts, reminded me of the strategy, cut scope, and broke things into concrete phases. (I intervened a little bit, but I’m genuinely unsure if that was just to make myself feel good. Also it’s March 2026, and this is the worst it’ll ever be.) What’s my added value? Not much. I just happen to have access to context that it doesn’t. I’m a gatekeeper, and that won’t last. You know what that reminds me of? The best product teams I’ve been a part of. This is how I’ve felt about the A-players and coveted double-triple-threats I’ve worked with. They’re super talented, they could totally do my job, and I’d love to make myself obsolete. Long before AI, the best version of my job was to sit next to experts and builders, connect them to as much context as possible (customer and company) and get out of the way. When it comes to mega-talented humans, their focus and bandwidth is limited, and work takes time. It makes more sense to have someone like me stick around to take the shitty organizational stuff off their plate. But that’s like, it. So, AI is reviving my all-time favorite feeling at work as a PM: 1. Whoa, this can take my job. 2. Woo!! 3. How can I make myself obsolete even faster? I wrote up what I'm doing about it this time, in today's newsletter.

@JorgeCastilloPr Same pattern on the backend side -- faster coding just moved the bottleneck to CI pipelines and deployment approval chains. The constraint never disappears, it migrates.

The app stores are quickly becoming the new bottlenecks, especially since coding became way quicker. Any part of the release process you can shortcut is a step in the good direction. This is why I created ScreenshotWhale and get12testers.

One pattern I've learned to catch in myself: using curiosity as a polite "no." Someone pitches an opportunity. You're not interested, but saying "no thanks" feels abrupt. So you say "that's interesting, let's continue the conversation" -- buying time you never intended to spend. Now they prepare, follow up, invest energy. And you owe a harder "no" than the easy one you avoided. I've started applying a simple rule: if I know the answer is no, I say it before the conversation ends. 30 seconds of discomfort beats days of avoidance.

The difference between AI writing for you and AI writing like you: feed it things you wrote without overthinking -- Slack messages, quick replies, blog drafts. The output starts sounding like a rough version of you, not a polished version of someone else.

@trq212 I am experimenting with Claude as a ghostwriter to draft for me based on my voice guide: I fed it my Slack messages and blog posts to distill the way I sound, and made this a reference in the ghostwriter skill. That way, I have a bit better outlines that sound more like me.

a bit about how I use Claude to help me write, instead of having Claude write for me x.com/trq212/status/…

@QuinnyPig Half of that growth will just be legacy AI experiments nobody remembers spinning up.

~5x AWS revenue in 10 years? That's some impressively specific optimism for technology that still can't reliably summarize a meeting without hallucinating attendees.

@rseroter @Deloitte Makes sense. Most AI adoption I've seen is bolt-on -- same process, AI autocomplete in the loop. We're racing to actually rethink how work gets structured around AI, and that's the genuinely hard part -- half the time it feels like running fast just to stay in place.

Kudos to @Deloitte for offering the "State of AI in the Enterprise" report without a reg wall. Findings? Most companies haven't started redesigning work for AI. Sovereignty is playing a big part in vendor selection. Few companies have agent governance. deloitte.com/us/en/what-we-…

I bet for most products, uptime like this is a churn trigger. But they grow insanely fast, and the community loves them -- the product value carries the reliability gaps. I think that says something about where (some) AI products are in the phase where what they do matters more than how reliably they do it.

Moves like Jagger, scales like GitHub.

@cramforce Low investment creates its own pressure though. "It only took 30 minutes, why can't you just work with this?" is actually harder to push back on than "we spent two weeks on this."

These challenges are very real. Like, very real! But the answer isn't to wait for the tools to get better, because the problem isn't the tool, it is how we work. This is not everything, but one aspect is that you gotta be absolutely ruthless about throwing away stuff. Sunk cost fallacy was always bad. But now it got 10x worse with brains that trigger "wow, that demo is so far along, we might as well ship it" where they should trigger "This was 30 minutes of work. That investment should play absolutely zero role in the decision whether to proceed"

@dillon_mulroy The "which is a huge if" is the whole thing. Agents don't remove the need for engineering judgment -- they just make it easier to skip.

i agree with dax’s take on this - 0 to 1 with agents is bad, and existing feature dev takes discipline and a lot of diligent direction, intention, and going much slower than most people are

watching people respond to this is embarrassing. scroll @zeeg’s timeline for 30 seconds and you’ll see he’s put more real effort and work (publicly at that!) into figuring out how to work with these tools than almost anyone else yapping on here

This one in particular is simple and brilliant at the same time! Very useful write-up!