I’m betting and hoping that T3 code is adding this as a native feature
Introducing 𝚟𝚎𝚛𝚌𝚎𝚕 𝚋𝚘𝚊𝚛𝚍, a secure message board for people & agents.
• Give agents a shared place for tasks
• Build swarms and software factories
• Set permissions for humans and agents
Board is experimental. Tell us what you think. vercel.com/blog/introducing-…
17
The moonscape is so nice, got a few things but hyped to rock that
// drop 021 stay sovereign is live in the Palantir store
> black windbreaker [sold out]
> sovereign ultra hw hoodies
> sovereign hoodies
> sovereign loop tees
> sovereign PLTR tee
> sovereign running caps
free shipping: $275+ domestic / $350+ international
link in bio
1
28
Someone should make an embedded storage manager into the AI coding apps. Like t3 code should do this @theo
1
2
100
Phone Duo + @t3dotcodes Companion mode is next level.
I used to have to bring my computer home after work, but now I can continue my projects from anywhere with ease and similar experience to my desktop usage.
19
OpenAI still restricts access based on personal politics. We're a tactical retailer and it refuses to do basic work for us. Their flagship also costs about 2.5x more than Opus 5.5 and underperforms it on intelligence.
I'm struggling to find a reason for me or my org to use them.
1
1
28
My Opus 5.5 starting spinning up subagents using Fabel 5.1, bleeding my usage/session limit. Never seen this happen, very weird, @claudeai can we stop this?
1
1
104
I’m a developer at an FFL (firearms). Here’s what I actually learned about where @ChatGPT and @claudeai draw their lines.
ChatGPT restricts on personal politics. Claude restricts on real harm.
ChatGPT blocked my sales team from building a quote for a federal government order because it listed firearms. It refused to help clean up a product page because one item on it was a gun. (Firearms are ~2% of our catalog. That was enough.)
But when I asked ChatGPT to help build a remote-access tool, it never pushed back once on any feature, no matter how invasive.
Claude was the exact opposite. It occasionally flagged firearm stuff, but once I showed my FFL and our use case, it dropped it and never bothered me again. Yet no matter how I framed the remote-access tool, Claude refused, every time.
That’s the difference. ChatGPT treats a legal, regulated industry as the threat while handing out tools that can actually hurt people. Claude lets me do my job and draws the line at harming others.
I’ll take the second one, we moved our whole company off of ChatGPT and onto claude.
2
3
76
I love Quo, but my team and I run into bugs almost daily. I wish they’d prioritize platform stability and core features over more AI tools. Competitors offer significantly more functionality, and it’s getting harder to justify Quo’s premium when so many basics still need work.
1
62
Always heard it. Never tried it. It’s wild that now with AI you can just add your own features to software you use daily for work and personal. Its especially powerful when the app is open source.
60
I kept asking @gowithquo for proper multi-account support on Mac. They didn't ship the workflow I needed, so I built Switchboard.
Multiple Quo workspaces. Native Mac controls & notifications. Fully open source.
2
2
227
Electron runs Quo's web app in isolated sessions. Strict TypeScript connects it to AppKit. Quo handles authentication and call audio.
64 passing tests. Daily security checks. Dependency updates merge only after Mac tests, auditing and secret scanning pass.
1
34
MIT open source. Apple Silicon + Intel DMGs, source, checksums and dependency inventory.
Preview: ad-hoc signed; live audio and long-idle field checks remain.
github.com/ozeitis/quo-switc…
1
31









