Cryptocheese

@define_app Im i dire need of a better mail app. But if you think im going to give you “my” email as a precondition to see anything at all at your website… you are wrong.

Reading emails side by side sounded like a tiny feature... Turns out it's one of those rare things that feels unnecessary, until you use it. You move through conversations faster, keep context clear, and reply with far less back-and-forth between tabs.

@CoachDanGo Simple: if you can VISUALLY identify the mess and it ANNOYS you, then you should do smt about it. Ofc one should be “judged” id the annoyance treshold is very high. But so should be one who is annoyed by smt that is not perceivable. Health risks are basically inexistant.

As an Asian guy I never understood why people walk with their shoes in their home.

@nihil2501 @mitsuhiko Cus there are trade offs for people who work with photo/video editing. And as per usual, theybwere making opinanted choices

@mitsuhiko why on earth did apple force everyone to look at gloss for decades when it has always been obvious that matte is so much more amazing. stuff like this kills me

Nanotexture means working from the balcony works!

@NtTestAlertX @paddi_hansen Yeah, i do see the instrumental value in executing 80% preference even if it is “no good”. Otherwise it’s difficult to maintain trust in institutions. And without it everything will become a shit show rather sooner than later

the "transformation" of my country was a deeply corrupt process with institutional rot. we still have PRLisms in our institutions and the corruption perpetrated by the same circles. everyone was in on it, so it's really a crappy example. but yeah, 80% can be "wrong", tho imo it is not in these matters, and with the polarization and tribalism in todays politics I'd say if something is an 80% issue it is even less probable they are wrong, but if they are they should still get to find out the consequences of being wrong. live and lear

A quick update on the infamous EU “ChatControl” 🇪🇺 What a turn of events in EU tech policy: from potential mandatory mass scanning of data (“ChatControl”) → to even voluntary scans losing their legal basis (for now). Just months ago, fears were growing around mandatory scanning of private communications in the EU (incl. pictures and videos). Now, talks between the EU Council (Member States) and the European Parliament have collapsed - and the result is a complete reversal. As of April 3, even voluntary scanning of data by platforms loses its legal basis under EU privacy (ePrivacy & GDPR) rules, as the temporary exemption was not extended. A striking example of how fast EU tech policy can turn - and a big win for European privacy advocates.

@NtTestAlertX @paddi_hansen It matters to the extent what patterns of political thinking and action it enables. Pretty important to be picky about that imo

@CryptoCaseus @paddi_hansen I don't really care what Rossignol said /s (it was not the same thing anyway, just sounding similar) Hitler drank water, does that mean I shouldn't drink water? Okay, bad example, I'm dehydrated.

@NtTestAlertX @paddi_hansen What people want is indicative of what may be useful to do, but the ultimate virdict should be based on the current best available justification for a policy (and not so much peoples immedeate preference).

@CryptoCaseus @paddi_hansen Well there is such a thing as an 80% issue, and the eurocrats repeatedly go against the 80+% issues. I would say things that are a 80% of higher consensus are "the will of the people".

@NtTestAlertX @paddi_hansen “The people’s will” is a toxic rousseauan, go read books. Rousseau would be much closer to “far left” than the current eu structures. You just say words to vibe.

Arguments for ChatControl - far left autocrats have wet dreams of north koera - corporations want more data to feed to AI and ad networks - some irresponsible parents want to hand their kid a tablet and ignore them Arguments against ChatControl - erosion of last vestiges of privacy (and GDPR incompatible at that) - rampant identity theft and other security issues - complete control of information with orwellian targeted suppression and censorship - destruction of democratic processes sure, I'm such a "populist", cause I won't "think of the kids", you "slowthinker". I think our "leaders" think about kids way too often. And on the wrong kinds of islands.

@trq212 Dudes, but why do you need to rely on tg’s api for that? Why can’t you make it read the history by reading the “channel” json saved locally?

read here on how to setup Telegram: github.com/anthropics/cla… and Discord: github.com/anthropics/cla…

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

@Palanikannan_M Looking forward!

thanks! means a lot! yeah i live in tmux these days (okay 3 years now), so made it pretty + useful, and I'm working on multiple worktrees at the same time so needed this bad! will make the ux super simple, the onboarding etc, and share my tmux config as well for your reference for the "pretty"/"not ugly af" part

pov: you tried cmux and realized tmux was right there the whole time all i needed was "is my agent done?" and fast switching. not a new app. runs inside ghostty. same tmux shortcuts. freakishly fast

@NtTestAlertX @paddi_hansen The will of the people as totalitarian/populist a concept as it gets. There is plurality of individuals with competing wills, you, “freethinker”

@paddi_hansen I'm sure this is still far from over. These autocrats think the will of the people is something to overcome, not accept.

@AdvicebyAimar He is not the same league as Jobs. But it’s not about claude code. Why bs?

comparing the guy who vibe coded on top of claude code to steve jobs is crazy.

@SpatiallyMe That looks neat, but y’all guys who may want it for this (and most part of other usecases), should really learn to use your keyboards. They already have all keys you may possibly need

Open Mac apps on your iPhone with my new app choclift 🍫 After extending my Mac screen with Vision Pro, I wanted even more connectivity between devices so I explored the advantage of having a separate user interface just for app management. It seems simple but it actually allows for a lot of new workflows and use-cases as you might think. Everyone works differently so we’re building this app as a platform that will support many different workflows, beyond just opening apps. Stay tuned as I’ll be sharing more in the coming weeks.

@maria_fibonacci Yes, and you shall be swift! Otherwsie, your moral high ground might sink to the level of your Dubai skyscraper’s foundation. Which is a migrant labor camp ;)

Oh well, I guess it's time to go back to iTerm2, then 😛

@MeganTStevenson But it's a piece of software, not a script?

This is an absolutely massive speed increase compared to the various "highly recommended" skills I'd tried previously.

@martyamark I feel you. Hacky solution: If you use any coding agents, ask it to setup hammerspoon interceptor for this. You can set up any behaviour for when you click on these links. I.e. copying them to the clipboard.

Can we make mailto: links illegal? Nothing ruins my day faster than clicking "Contact Us" and watching my mac launch the mail app and immediately crash.

@terry_b_davis @alxfazio Yep, as someone with only Alfred as a "poweruser" background to roll into cli was muuuuuuuuch eithier than setting up an mcp

@alxfazio The agent will just install the skill and cli itself Anyway, installing clis is much less effort than installing MCPs

mcps may be dead, but no normie is going to install a cli tool + a skill just so the clanker knows how to drive the cli and invoke the skill every time they need to use the tool

@marzeaned @ArtemXTech What the actual hell. I think this is the most thoughtful piece of engagement I've ever saw on this platform.

This solves a real problem in the dumbest possible way, and I mean that as a compliment. Tabs are genuinely terrible for multi-agent work. Anyone who's stared at "Claude, M1, M2, M3" knows the feeling. So the core insight -- named workspaces you can read and send to programmatically -- is solid. But let's talk about what's actually happening here, because the article buries the lede. The three commands -- list, read, send -- are a control plane. That's it. cmux is a lightweight control plane built on terminal multiplexing. The article dresses it up with Obsidian dashboards and daily notes, but strip that away and you have: an orchestrator that can enumerate agents, observe their state, and dispatch work. That's the same architecture every serious multi-agent system lands on eventually. The question is whether terminal screen-scraping is the right transport layer for it. Here's where I'd push back hard: "reads the screen" and "sleeps for 15 seconds" is brittle as hell. You're parsing terminal output as your communication protocol. What happens when an agent prints a progress bar? A multi-line error? A prompt that wraps differently on a narrower terminal? Screen-scraping works until it doesn't, and when it doesn't, you get silent misreads that corrupt your orchestrator's understanding of what's happening. Compare this to structured communication -- agents writing JSON to files, posting to an API, or even just appending to a shared log with a known format. The orchestrator doesn't have to guess what it's looking at. It reads a status field. Done. No 15-second sleep. No hoping the output has settled. The Obsidian dashboard layer is clever for a single human operator but doesn't scale past one person. The moment a second person needs visibility, or you want to run this headless, or you want programmatic alerts when something stalls -- you're rebuilding the dashboard as a real application anyway. Obsidian Bases is doing query work that a database does better. What I actually respect about this setup: the session-as-artifact pattern. Every workspace producing a session file with goal, progress, outcome, and definition of done -- that's genuinely good practice. Most people treat agent sessions as ephemeral. Making them first-class documents you can review, comment on, and link together is the right instinct. I'd steal that pattern and put it on a real backend. The workflow loop -- daily note, sessions, spawn, verify, comment, relay -- is also solid in concept. Having the orchestrator read your plan and derive intent is better than most people's approach of manually typing instructions into each agent. But "the orchestrator reads this plan and understands your intent" is doing a lot of heavy lifting in that sentence. How well does it actually parse freeform daily notes into actionable agent tasks? That's where the magic is, and the article hand-waves past it. Bottom line: if you're drowning in Claude Code tabs today, cmux is a genuine upgrade. Named workspaces with programmatic access beats anonymous tabs every time. But this is a v1 personal workflow, not an architecture. The terminal screen-scraping, the 15-second sleeps, the Obsidian-as-database -- all of that has a ceiling, and you'll hit it the moment your agent count or task complexity goes up by 2x. The ideas are right. The implementation needs a real transport layer.