Blake Xu (@BlakeXu_) header
Blake Xu (@BlakeXu_) profile picture

Founder | AI | Digital Marketing | Growth Marketer 🏠 https://teapawt.net/t.co/t2dem135Ie 🧾 https://teapawt.net/t.co/F0mwtkhydh 🏭 https://teapawt.net/t.co/WQrLad056f

Joined September 2026
Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)Photo by Blake Xu (@BlakeXu_)
What have you shipped that people use in a way you never expected? Those stories are usually better than the original pitch.
1
4
76
Small teams have an advantage when they can make a decision and ship it today. That advantage disappears if they spend the week copying a big company’s planning process.
2
7
181
More builders should publish the unglamorous details: How they handle refunds. What happens when an API goes down. How they restore a backup. That’s the part I’d bookmark.
1
4
140
If your app generates something, let people edit it. If it takes an action, let people inspect it. If it changes something important, give them a way back.
6
15
280
A surprisingly good customer interview question: “What happens if you do nothing about this?” Sometimes the answer is: nothing. Good to know before spending three months building.
4
7
131
You don’t have to beat the biggest product in the category. You need a good answer to: “Why would this specific person choose yours?”
9
5
154
Developer tools should make the first error easy to understand. The happy path gets someone to try your tool. The first broken setup decides whether they stay.
3
6
219
The cheapest model per token isn’t necessarily the cheapest model per completed task. Count the retries. Count the corrections. Count the human cleanup.
3
7
176
I’m more interested in what a founder decided to stop building than what they added to the roadmap. That’s usually where the customer conversations show up.
6
9
153
A product metric worth tracking: How long until a new user gets something they can actually use? Account creation doesn’t count.
6
1
10
208
AI-generated interfaces need an editor. Someone has to ask why a basic settings page has six cards, three gradients, and an inspirational headline.
1
1
7
254
There’s a difference between users loving your product and users loving the fact that you do everything for them. You’ll find out when you try to take a weekend off.
2
1
8
294
Your competitor having more features can be good news. Find the customers who are tired of navigating all of them.
10
1
16
360
What’s an AI workflow you’ve kept using for more than a month? Not a cool experiment. Something you’d actually be annoyed to lose.
4
1
9
202
A useful portfolio project: Build a webhook receiver that handles duplicates, retries, and out-of-order events. Then explain your decisions. I’d rather read that than another stack list.
5
9
286
Being able to build anything is a strange advantage. It makes saying no to a bad idea much harder.
6
1
15
237
An agent that can spend money needs a budget, a receipt, and a stop button. I’d want the same from a human assistant.
3
6
259
The best onboarding might be a completed task. Before the tour. Before the checklist. Before the “tell us about your role” screen.
4
9
321
A cancellation form that says “too expensive” leaves out the useful part. Too expensive compared with what? Another tool? Doing it manually? Leaving the problem unsolved?
3
6
154
Founders: what convinced your first customer to trust you? Especially interested in the answer if you had no audience, no funding, and no recognizable logos.
6
7
192