Pinned Tweet
I’m thinking in public about AI and Bitcoin because they’re reshaping how society and business work — often faster than we can process.
This isn’t about predictions or hot takes. It’s about slowing down, refining my thinking, and learning out loud.
I’m here to learn, ask better questions, and share ideas in progress.
Most enterprise demos fail in the first 10 minutes because Presales jumps straight from "Customer Problem" to "Product Feature."
A value-first Presales motion forces 4 steps in between:
1. Reframe technical friction as operational drag
2. Agree on the required future-state capability (before naming UI)
3. Prove that capability with one tailored workflow
4. Connect the proof to a baseline KPI
ROI spreadsheets don't close CFOs. A tight Value Narrative does—the spreadsheet just confirms the math.
1
30
I used to be intimidated of how to start because building took so much effort and didn’t have the time or resources. Now I’m struggling with so many options of how to build various approaches to my idea and inspire how to package. Simple kits for volume use ? Or coaching and consulting higher ticket or full on consulting projects. Or some combination. Evaluating all models at the moment.
15
When more than one AI helper is working at once, write down who decides what.
One line per piece of work: which helper is doing what, and what stays your decision.
13
With AI, build something small for someone else to use.
Pick one person and one problem of theirs. Aim for a first version they can try, not a finished product.
1
14
When AI finally gives you a draft that's right, save its shape.
Ask it to record the structure and layout for documents of that kind. The next one starts from the template, not a blank page.
25
Pick one thing you've avoided because you don't know how.
Ask your AI to walk you through it one step at a time. Then do the first one by hand.
12
If it's a must, it goes on the calendar.
Take the one thing you keep meaning to do and give it a time this week, like any other meeting. The time AI gives back disappears into the queue unless you name where it goes.
11
I had never made a video in my life. I committed to delivering one anyway.
The first one was slow. Every step by hand, and plenty of it wrong.
One habit changed that. Before I closed each session, I asked the AI to write down what we'd learned: the steps that worked, the ones that didn't, the standard for next time.
The second video started from those notes, not from zero. So did the third.
Don't write your process down afterwards. Have the system keep track as you go.
12
Find the step where everything waits on you, and take yourself out of it.
It's one of the first things I look for. Start by listing what sat waiting for you this week, then pick the most routine item on the list.
That's the one to hand off first.
16
Before AI takes on a task, have it ask you questions.
I add one line to a request now: "Before you draft anything, ask me what you need to know."
Try that line on your next one. It's one of the small habits I actually use, and it costs one sentence.
14
A quick test of how you're using AI.
When you sit down to do something you've done before, does the AI already know how you do it? Or do you start from a blank chat and explain it all again?
If it's the blank chat, you're learning the same thing twice.
1
24
Building with AI comes with a dilemma, and it's worth being realistic about it.
The opportunity to keep building never stops. The engineering, the design, the content: none of it needs to rest. So you work late. You get up early because there's another project you could start. You try to keep pace with something that doesn't get tired. That needs to be managed, and it has to be sustainable.
Then comes the context switching. You start one piece of work and let it run. While it's running, you start another. The first comes back with questions, so you answer those. There's a gap, so you start a third. Now all three come back at once, and you're working out where you were on each one.
Very quickly you can overload yourself. I hit that wall.
This is what more capacity feels like from the inside. You really can take on more than you could before. And the limit moves from the work to your own attention.
What I've changed so far: I use the system itself to run routines that help me manage it. I'm still figuring it out. It hasn't been fully solved.
If you're at the point where you can run several things at once, let's be pragmatic:
1. Notice when you start something new only because the last thing is still running.
2. Decide what brings you back to each one.
3. Protect the rest the AI doesn't need and you do.
1
14
When using AI and a result isn't right, I don't quietly fix it myself.
I say what's wrong and why, and I ask for it again. That's the job when you're directing instead of doing.
Next time a result misses, send it back with the reason.
18
The most useful thing I've done with AI wasn't a prompt. It was a commitment.
I promised to deliver something I had no idea how to make. Then I got curious and asked AI how we could build it with the tools I already had.
Commit first. Then get curious.
1
44
Expertise can start with the skill right next to your job. Then it keeps widening as you see what's possible.
20
When I need something like a one-pager now, I don't start by writing it.
I describe what good looks like. I describe the edits I want, the tone of voice, the kind of output. We go back and forth until it's where it needs to be.
Then I ask the AI to save that structure and layout as the template for documents like this one.
That's a different job from doing the work. It's giving clear instructions, delegating more, trusting more, and giving short, direct feedback. More orchestration than execution.
Next time you'd normally start typing the thing yourself, describe what good looks like first.
19
I had never made a video in my life. I committed to delivering one anyway.
The first one was slow. Every step by hand, and plenty of it wrong.
One habit changed that. Before I closed each session, I asked the AI to write down what we'd learned: the steps that worked, the ones that didn't, the standard for next time.
The second video started from those notes, not from zero. So did the third.
Don't write your process down afterwards. Have the system keep track as you go.
Building with AI comes with a dilemma.
The engineering, the design, the writing: none of it needs to rest. So it's easy to keep going late into the night, or to start a second project while the first one runs. Then a third. Then all three come back with questions, and you're trying to remember where you were on each.
Very quickly you can overload yourself. I hit that wall.
More capacity is real. It also has to be managed, and the thing to manage is your own attention. So before you add the next project, ask what will bring you back to the first one. For me that's routines the system runs. I haven't fully solved it.
27
I committed to delivering a video before I knew how to make one. I'd never made one in my life.
I knew that if I committed, I'd figure it out. So I got curious. I asked AI a plain question: how could we build this with what I already have?
I assumed I'd need other tools. It turned out I already had access to what I needed.
Curiosity, commitment, and a bias toward action. Pick one thing you'd normally pass on, commit to it, and start by asking how.
25
