
Building things that make me feel like I'm living in 2035. 🧾 https://teapawt.net/t.co/g9Ut8DVxXr 🔎 https://teapawt.net/t.co/Nn69C3T0OG 🎲 https://teapawt.net/t.co/2tY6cCrSkS 📒https://teapawt.net/t.co/AeSVkk31UN 💡https://teapawt.net/t.co/2Ee4SRoLf3
Philadelphia Area
Joined December 2025
- Tweets1.1K
- Following379
- Followers610
- Likes1.9K
Pinned Tweet
Meta Ray-Bans (native WhatsApp support) + @steipete's @openclaw + tmux = hands-free voice control for my whole system & all of my Claude Code and Codex agents
Agent status check:
"Check agent tmux sessions, show any awaiting approval...yes approve that command"
At a store, see something overpriced:
"Use the alexa skill, ask what's this price on Amazon?"
$30 cheaper.
"Tell Alexa to add that to my cart"
Smart home:
"In an hour, set the thermostat to 68 and turn on the living room lights"
Logging meals:
"Use the myfitnesspal skill and log my lunch: I had a 4oz chicken breast, a Mission wrap...etc"
At my desk:
"Create a GitHub issue on aptus: fix sidebar mobile keyboard touch handlers"
Phone stays in pocket the whole time. Smart home controlled hands-free from anywhere. Agent session statuses reporting and managed via voice. Tasks captured.
The frontier voice models have pretty limited code execution and integrations, but I have to imagine this is the direction they're headed.
For now, this feels like living in the future.
Buddy Hadry retweeted
Another example of "agent teammates" not being the right mental model.
This by Anthropic seems both right & a condemnation of many “agentic solutions” & harnesses being built at companies. Rather than a sprawl of specialized subagents built by different teams, you often get better & cheaper results using one agent to do it all claude.com/resources/article…
Buddy Hadry retweeted
I dug into this a bit. It looks like we're at highs, but not unprecedented highs for major systems failures.
We're seeing alarming headlines every day, but a lot of it is recency bias.
are you also struggling to safely give your agents access to passwords?
meet agent keyring (v2.0). open source (MIT).
approve passwords from your Watch, your iPhone, or your Mac, anywhere in the world.
v1.0 is live at github.com/joshp123/agent-ke… (approval only when your Mac is unlocked, touchID and Watch); v2.0 (designs below) will ship within a week or so.
I got fed up with my agents not having access to log in to things, but I don't want to YOLO and give them credentials to everything.
Agent Keyring lets them request access to any password - and you can approve or deny it, with a reason (often they ask for things that they shouldn't).
I've been using this for a couple of months now and it's completely indispensable for me.
You can let certain sites always approve (if you don't care about that site), or auto-reject others (if you don't want your agent ever getting access to it).
The password ends up securely in a file, so it never gets shipped to your inference provider.
I'm also working on a secure-entry mechanism like @bot and @t3dotcodes has - so your agent can even be prevented from accessing the password too.
It imports passwords securely from Apple Passwords / iCloud Keychain.
I know 1password and others have similar features - but this is free, self-hosted, and fully open source.
Would you use this? What features would you need to make it part of your daily workflow?
Buddy Hadry retweeted
Me to my new homepod mini:
Hey Siri, I have some devices I want to move from my office to another room, are you able to move some for me?
Siri:
there is a motion sensor in the office.
Me to my new homepod mini:
Hey Siri, I have some devices I want to move from my office to another room, are you able to move some for me?
Siri:
there is a motion sensor in the office.
Buddy Hadry retweeted
The moment the new Claude model hits and you have it redesign your project and think “wow, now this is it!”
And then see that design 10 times over the next two weeks on X
Every time
Buddy Hadry retweeted
pro tip: make your agent's scheduled daily brief end with launchable sessions.
mine fires first at 6am, then again every 2 hours.
3-6 new one click agent sessions each time based on what changed and what's come in since the last run.
pro tip: make your agent's scheduled daily brief end with launchable sessions.
mine fires first at 6am, then again every 2 hours.
3-6 new one click agent sessions each time based on what changed and what's come in since the last run.
Buddy Hadry retweeted
Elderly and can’t stand on one leg for 10 seconds?
About 1-in-7 people aged 50+ who can't die in five years.
The moment the new Claude model hits and you have it redesign your project and think “wow, now this is it!”
And then see that design 10 times over the next two weeks on X
Every time
Buddy Hadry retweeted
Every day I see a new tool/post get attention that showcases something Claude / Codex already do on their own.
Sometimes I think people need to stop, open a new thread and just ask first and see whether it actually required any new tooling.
Every day I see a new tool/post get attention that showcases something Claude / Codex already do on their own.
Sometimes I think people need to stop, open a new thread and just ask first and see whether it actually required any new tooling.
Buddy Hadry retweeted
artisto.so is now live 🎉
fully free for now. try to connect your agent through api and mcp
(pls do not try to do things manually that's not the point of the fucking app, it's a design tool for agents)
Buddy Hadry retweeted
I use LAN Mouse to seamlessly transfer my mouse and keyboard from my Mac to my Omarchy box.
Since voice dictation is such a big part of my workflow, I wanted the mic to transfer too.
Built a little tool that does that, works as a companion to LAN Mouse. Shows the mic transfer in the menubar as it happens.
LAN Mic, repo below
👇
Buddy Hadry retweeted
Replying to @nvidianewsroom
jensen will take some quality time off with his families

















