Better Communication Makes You Better at AI
I’ve said before that working with AI is mostly just good communication. That idea came up again listening to Wes Kao on Lenny Rachitsky’s podcast. The episode is framed as a conversation about executive communication and influence, but it ended up being one of the best explanations I’ve heard for how to work well — with people or with language models.
Even if LLMs didn’t exist, communication would still be the skill I’d bet on. It’s consistently been the biggest lift in my career. It’s the thing that multiplies everything else: hiring, product, leadership, writing. If you want to get better at building things, start with how you explain your thinking.
My favorite takeaways:
-
Use MOO (Most Obvious Objection). Before you share anything — an idea, a message, a doc — pause and ask yourself: “If someone were to disagree with this, what would they say?” Then handle it upfront. Wes emphasized the value of making this a habit. The key is figuring out simple ways to keep frameworks like this in front of you — whether it’s a note in your prep doc or a quick check before you hit send. If it’s not easy to reach for, you won’t use it.
-
When things don’t land, start by looking at yourself. If a Slack post, a deck, or a pitch didn’t get the reaction you wanted, don’t blame the audience. Ask: Did I lead with the right message? Was the core idea clear? Did I jump into logistics before getting alignment? That habit of reflection is useful well beyond communication — it shows up in leadership, product decisions, and relationships. It’s an underrated mindset.
-
Concise doesn’t mean short. It means clear and efficient. Wes pointed out that the reason people struggle to be concise is usually because they haven’t figured out what their point actually is. You can’t cut to the chase if you don’t know what the chase is. She gave the example of telling a story you’ve told ten times — over time, it gets sharper. But at work, you’re usually explaining something new, something you’re still processing. That’s why real concision takes prep. You need to do the thinking first.
-
If you’re asking someone to read it, you should have read it. Attention is fragmented. Everyone is overwhelmed. If you’re asking someone to process your message, it’s on you to make it worth their time. That means reviewing what you wrote, trimming it down, and making it clear. It’s table stakes.
-
Prepare for meetings. You need to have the meeting before the meeting. Don’t treat it like improv. Go in with a view. Know what you want to say and why it matters. If you don’t have a goal, you’re wasting people’s time.
-
Communication compounds. Small messages aren’t throwaway tasks. Slack threads, status updates, emails — these are the building blocks of how people experience working with you. You have to put yourself in the recipient’s shoes and ask: how are they going to read this? What’s the takeaway? Then revise it until it’s clear. It’s not about checking something off your list. You need to make it about building trust over time.
-
Bullet points can hide unclear thinking. Wes made this point well: bullets and formatting can make your writing look structured, but it might just be fragments strung together. If you try to turn each bullet into a full sentence and struggle, that’s a sign you don’t fully understand your own thinking. She emphasized that bullets can feel fast and polished, but often they’re just a shortcut around the hard work of organizing your thoughts. Don’t mistake formatting for clarity.
-
Sales before logistics. You need to be aware of what you’re trying to do: are you getting alignment on the idea, or coordinating execution of the idea? If you skip the sales step and start with process, you’ll lose people. It’s a sequencing problem. If people don’t understand why something matters, they’re not going to care how it works. Wes gave a great example where someone was asking execs to fill in a form, but led with deadlines and formatting instead of the why — which meant no one engaged.
-
Sometimes people don’t ask what they mean. Especially in cross-functional settings. You have to peel the onion. Someone might ask about a metric, but what they’re really asking is whether a project is off-track. Don’t just answer literally. Figure out what they actually mean and respond to that.
-
Feedback should drive behavior change. Not catharsis. If your goal is to feel heard, vent to a friend. If your goal is to help someone improve, strip your feedback down to what will help them do something differently next time.
-
Delegation without clarity is just abandonment. Wes shared a framework: comprehension, excitement, de-risk, alignment, feedback. Most failed handoffs break on one of these. If the other person doesn’t understand the task, isn’t bought in, doesn’t see what success looks like, or isn’t getting feedback — you’re setting them up to fail.
-
Designing your work around your strengths is underrated. Wes talked about this in the context of starting her own business. But the idea holds even if you’re in-house. If you can shape your role around the narrow slice of work where you do your best thinking and create the most value, it gets a lot more sustainable.
Final thoughts:
If you’re serious about getting better at prompting, building, or collaborating with LLMs, you don’t need another AI course. You need to get better at explaining things. Go read On Writing Well. Write a post about something you learned. Try saying something clearly, then try saying it better. These tools are just amplifiers. Your thinking is still the foundation.
This podcast episode is worth your time. These were just the takeaways that stuck with me, but there are probably 20 more useful insights I didn’t even get to. You can listen to the full conversation here: