My first iOS app is live!
I got tired of losing track of which prompts actually worked across different models and platforms. So I built PromptFolio. https://t.co/FWjdQpNb6v
I have this tool Claude and I made: just type "twit" in your terminal followed by the URL of the X post and it saves it to your location of choice (mine is Obsidian): https://t.co/7LTBd8Q5yP
luckily bookmark rot is an easy problem to fix now
here's how to turn every X bookmark you've ever saved into a second brain your agent has full context on:
1. export your bookmarks. i use twitter-web-exporter (free userscript) or the BookmarkSave extension. you get one file with every bookmark + the full text + the author + the link
2. drop that file into a folder. if you already run an llm wiki / obsidian vault, drop it straight in so your bookmarks join the rest of your knowledge
3. point your agent at the folder (claude code, codex, hermes, whatever you run) and tell it: "read this export and turn every bookmark into its own markdown note with the original link and a couple of topic tags"
that's it, your agent has read all of it.
now you can ask "what have i saved about pricing" or "pull everything i bookmarked on claude code" and it answers across the whole pile
takes maybe 10 minutes
after that they actually get used, and every new bookmark folds into the same brain instead of rotting in a tab you never open again
To animate something like this (show the tape wheels spinning, the tape in window reduces on one side, grows on the other, the exposed tape maybe shows a reflection at times to make you think it's moving)- what would be the software of choice? AI or manual. 🙏
So now I'm doing what I should've done from the start — two weeks of manual testing across all three agents, just me at the keyboard, tracking every session. Does it checkpoint? Does it restore? Real data, not automated test harnesses.
Update on Agent Time Machine: the snapshot and restore engine works. Agents checkpoint, agents restore, the MCP tools are solid across Claude Code, Codex, and Gemini.
Before I go deep, I want to know:
• Does this problem resonate?
• Would your agent benefit from self-awareness tools?
• What am I missing?
Roast it, improve it, or tell me it’s dumb.
#buildinpublic
The challenge I didn't expect: agents don't follow instructions every time. They read them, understand them, can quote them back — and still skip the checkpoint because the edit "felt simple." This isn't a tm bug. It's an LLM reliability thing.
@LyalinDotCom and @OfficialLoganK Sorry to bother you- hope things are well. I don't know where to turn for help with Antigravity. I'm an Ultra subscriber but every single prompt fails. Anyone you can offer to help?
Ok- love you. bye.
@jocarrasqueira@antigravity@_anshulr@geekyouup I repeatedly can’t get it to work- even as an ultra subscriber- it gives me an error and just says to retry the prompt. just wish I could get help
Before I go deep, I want to know:
• Does this problem resonate?
• Would your agent benefit from self-awareness tools?
• What am I missing?
Roast it, improve it, or tell me it’s dumb.
#buildinpublic
So Codex and I are building Agent Time Machine:
Not a human tool that wraps AI commands.
An MCP server that AI agents USE.
Agents checkpoint themselves.
Agents inspect their own diffs.
Agents can roll back if they drift.
Agent-first. Human-second.
Monologue for iOS is here.
You talk 3x faster than you type.
Prompts to Claude. Brain dumps to Notion. Grocery lists. Midnight epiphanies. First drafts that sound like you. Now in your pocket.
Follow, like, RT—50 annual subscriptions are up for grabs.