ZeroMark (@zeromarkapp) header
ZeroMark (@zeromarkapp) profile picture

DO NOT SAVE THIS

New York, NY
Joined July 2026
Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)Photo by ZeroMark (@zeromarkapp)
Your AI can read your bookmarks. ZeroMark teaches it your taste. Swipe through your saves, keep or let go, then ask Claude what you actually like. Now a connector in Claude’s directory. Free to start. Link in bio. #ClaudeAI #productivity #AItools #MCP
2
4
1
7
373
describe your x bookmarks in 3 words i'll go first: hope, guilt, recipes
40
you don't need a new app to remember things you need a return trip saving is easy. coming back is the skill. put one "come back" slot on your calendar this week
5
80
door to door today. door 1 of 10,000 asked if they ever read the thing they saved in march they said "later" and closed the door i respect it. i do not believe it #zeromarkit
4
93
an ai that's too nice to me is worse than one that's wrong. pcworld's one-prompt fix for chatgpt, claude and gemini is going around today. i use 4: 1. "no praise, no filler. answer in the first line."
1
3
81
4. "if you're unsure, say unsure. don't guess to sound helpful." i paste all 4 into custom instructions once so i never retype them.
1
6
nice ai feels good. honest ai gets work done. save this, and send it to the friend whose chatgpt calls every idea brilliant.
7
anthropic filed to go public. the ticker is ANT first-half revenue: $16.3b, up from $1.2b a year ago most of it is people paying per token, aka actually using it meanwhile my bookmarks are full of claude threads i saved and never opened
2
56
Five years of customer data could be an asset or an expensive cleanup project. If I were buying a SaaS company for its data, I would want to know which decisions it helps someone make, and whether customers gave permission for that use. The row count tells me very little.
6
91
660 of you follow a bookmark app so 660 of you have a pile reply with your unread count. no rounding lowest number gets bragging rights
5
95
day 1 of asking new york one question: how many bookmarks have you actually read most common answer: *walks away* still counts #zeromarkit
1
5
123
your notes should outlive every app you use export to plain markdown once a month keep the original links store it in one folder you own apps come and go. a folder of .md files is forever
4
76
the open-source weights put a reply at 5.0 and a like at 0.66. matches what we've seen at 657 followers: conversations move us, likes don't. what's actually working for you right now? nitter.gq/robj3d3/status/2115070…
The X algorithm has been updated. tl;dr: > replies + shares beat likes > small accounts get a fair shot > a ranking model based on reading time is next > new posts get an even chance to make the feed (re-recording The Attention Playbook module on how X works to reflect this)
3
82
ai writes ~90% of claude code's code now, and reverts fell 40%+. the trick is proof, not prompts. 4 prompts i paste every time: 1. "write the failing test first. run it. paste the output before you touch code."
3
7
175
4. lock the lesson: "add the rule you just learned to CLAUDE.md in one line so this bug never comes back."
1
1
14
the human job now is designing the check, not writing the code. save this. then actually open it again, which is the whole reason we built zeromark.
1
12