Your Slack didn’t used to feel like this. There was a time when a message arriving in a channel meant something had happened that someone thought you should know. Now it means Billy from sales has had a thought, an AI bot has summarised a thread you weren’t in, someone has shared a document without context, and three people have reacted with 👀 to something that required a decision.
The cost of producing content has collapsed to nearly zero. Drafting a message, generating a summary, writing a weekly update: all of it takes seconds with a model standing by. This doesn’t always improve communication. It often results in a cacophony of unbearable noise.
When attention becomes the scarce resource, the people whose output is worth reading get more visible. Not louder. Not more productive by volume. Just consistently worth reading.
The compounding effect
Think about the last person at work whose messages you always opened. You didn’t think about why. You just knew that when they sent something, it was worth two minutes of your attention.
That reputation is built slowly and lost fast. One too many context-free links. One AI-drafted summary that turned out to be wrong. One message that generated three clarifying threads and resolved nothing.
In a low-volume environment, that kind of noise is forgettable. In a high-volume AI-saturated one, it should be career-limiting. When there are a hundred things arriving, the person who reliably adds signal gets to the top of the mental priority queue. The person who doesn’t gets scrolled past.
The pattern plays out differently at each level, but the logic is the same. As an individual contributor, people stop double-checking your numbers because you don’t get them wrong. Trust builds. You get brought into the next important conversation because you were useful in the last one. As a manager, you absorb the chaos on behalf of your team: they stay heads-down because you filtered. As an executive, if yours is the synthesis the CEO trusts, your function survives the next reorg. If yours requires translation every time it arrives, it gets reshuffled.
The compounding works in both directions.
What signal actually means
Knowing more AI tools is not the answer. Posting more is not the answer. Being the person who always has an update is definitely not the answer.
Leah Tharin wrote about this last week with a label I have found useful: information discipline — the ability to produce and route high-signal content consistently. Four practices:
- Maintain a personal context document with the contested numbers and the humans worth talking to.
- Update existing documents rather than creating new ones.
- Send fewer messages with more metadata: what’s the ask, why now, what action is required.
- Protect deep work fiercely.
What strikes me about each of these is how much they require judgment rather than effort. Drafting a message is easy. Deciding whether to send it is hard. Creating a new document is easy. Finding the existing one and updating it instead is hard. AI makes the easy parts trivially easy. It does nothing for the hard part.
Tharin writes: you’re not a servant of AI. AI should be a servant of you. That only holds if you verify what AI produces before it leaves your hands. Skip that step and you are not a signal source: you are a relay for noise.
The management implication
If you lead a team, this goes in two directions.
The first is your own signal. In a world where every report can generate polished output with minimal effort, the quality of your synthesis is what distinguishes you. If you are passing along AI-drafted summaries without reading them, you are not adding leadership: you are adding latency. Your team can tell. So can your peers.
The second is the signal you’re looking for in others. The standard interview question about AI is something like “how do you use AI in your work?” The more useful question in 2026 is: can you show me the last thing you sent that required genuine judgment? Not the most impressive output. The one where the judgment mattered.
The people worth developing aren’t the ones who produce the most. They are the ones who filter the best: who absorb the noise before it reaches the team, who ask the clarifying question instead of routing the mess downstream, who update the canonical document rather than creating the fifth version of a strategy deck that contradicts the previous four.
I wrote recently about what AI hasn’t replaced in product management: taste, context, relationships, and the courage to say no. All of those are signal. They are also exactly what gets buried when the channel fills up.
Asymmetry
The AI flooding the channel is not a neutral force. It tilts the environment in favour of people who were already good at judgment and against people who were relying on volume.
If your professional reputation rested on being the most responsive person in the room, or the one who always had a draft ready, that advantage is gone. The AI is faster.
If your reputation rested on being the person whose read of a situation was worth waiting for, that advantage just got larger. Because now there are a hundred reads arriving every hour, and yours is the one people trust.
The signal problem is not really about technology. It is about what you choose to stand for when anyone can generate anything.
Stand for something worth reading.
Leave a Reply