Matt Delmarter

@FlyaKiet @superset_sh Do you generally work in claude with /tui default? I find /tui fullscreen to be quite laggy still, but has less issues with text reflow.

We spent last week optimizing performance of @superset_sh's core piece before pushing for next big phase of the product. Big one is terminal scrolling. We were scrolling 1/3 speed of native Mac terminal. We rewrote a custom scroll handler to behave like a native terminal. Before, scrolling in Claude code jumped in chunks. Now, it tracks your finger. Small swipe moves a little, flick goes far. All smoothly. The same fix applies to vim, less, htop, opencode, and tmux, not just Claude Code. Open a new terminal in Superset to feel the difference! Credit to @bra1n_dump for reporting and @avimakesrobots for fixing!

@FlyaKiet @superset_sh Oh this looks nice. Side benefit - for a remote terminal there is no lag when typing 👏

Native Warp-like input now in @superset_sh terminals 🎹 CMD+I to activate @ to include files

@powl_d @superset_sh @orca_build Can’t wait to see it! 🔥

Bon, Sol d'OpenAI sort. Perso j'utilise Claude Code à fond via RC Remote. Je préfère largement @superset_sh à @orca_build . Le seul truc qui me manque pour pouvoir partir à la plage et continuer à bosser mais sur Codex proprement, c'est une app mobile Superset. Du coup, Emilien a une mission aujourd'hui : ship l'app Superset. 👀 Vu ce qu'il est en train de cook... ça sent très bon. 🔥

@FlyaKiet @saddle_paddle @avimakesrobots Congrats! 💍❤️🍾🥳

Just got married, here's what it taught me about B2B SAAS: It's more fun w/ your bros there

@bcherny Would be really nice if we could invoke /goal from a skill. As far as I am aware that is not currently possible?

Seeing a number of benchmarks showing Opus is the best model for long-running work. Five tips for running Opus autonomously for hours/days: 1. Use auto mode for permissions, so Claude doesn’t ask for approval 2. Use dynamic workflows, to have Claude orchestrate hundreds/thousands of agents to get a task done 3. Use /goal or /loop, to nudge Claude to keep going until it’s done 4. Use Claude Code in the cloud, so you can close your laptop (easiest way is the desktop or mobile app) 5. Make sure Claude has a way to self-verify its work end to end: Claude in Chrome browser extension for web, iOS/Android sim MCP for mobile, a way to start the full web server or service for backend work

@SebastianRoehl @focuskitapp Looks great! What are you using for your screenshots?

Added a cool changelog to the @focuskitapp website. I never kept track of changes in my other apps, and I kind of regret it. It’s fun to look back and see what you’ve shipped!

I am with you on this, have also been using Claude Code to generate html outputs - much easier to consume, visually fun to read. Documents become unique instead of overly standardised. In saying that, there is a world where you can convert markdown ⇒ html easily as well, to make easier to read plans/specs/reports/docs. There needs to be a workflow and tooling around it to optimise the advantages. I have been building a tool in this space which I hope to release soon.

@FlyaKiet @superset_sh Amazing work team.

@superset_sh 2.0 is now in open beta! For the past few weeks, we've been hitting physical limits on using Superset, my computer begins meltdown when running 20+ worktrees. Many of our users have 50+ (!!!) Since we're 3 ex-CTOs, the logical next step is to rewrite the whole app to enable cloud workspaces. And today that's available to the public 🎉

We’re introducing the Cursor SDK so you can build agents with the same runtime, harness, and models that power Cursor. Run agents from CI/CD pipelines, create automations for end-to-end workflows, or embed agents directly inside your products.

Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see. @eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook. (1/5)

@saddle_paddle @Superset Can we choose a default behaviour for clicks? Especially for urls, I prefer they open my external brower rather than inside superset.

Made it easier to access both internal and external links in @superset today use meta+click to open natively in superset, use meta+shift+click to open in the browser or editor of your choice!

Claude Opus 4.7 is available today on claude.ai, the Claude Platform, and all major cloud platforms. Read more: anthropic.com/news/claude-op…

@FlyaKiet @superset_sh Your team must be getting hungry over there Kiet! Stay strong 💪 looking forward to seeing v2 … and the mobile app 😀

Apologies for all the UX paper cuts recently in @superset_sh. We've been focusing on building v2 in parallel inside the app and sometimes functionalities leaks into v1 in an unintended way. The plan is to continue pushing for v2 and port over the good bug fixes into v1 when we see them. We will be better, we are so close.

Anthropic just released the most IMPORTANT chart in the AI labor debate. This comes from the company that builds Claude using data from 2 million real conversations. Here’s what it shows. The blue area is every task AI could theoretically do right now. The red area is what people are actually using it for. The gap between them is enormous and that gap is your career runway. Computer programmers are already 75% covered. Customer service reps, data entry workers, financial analysts, they’re next. But here’s what no one is talking about. The mass layoffs haven’t really started. Unemployment for exposed workers hasn’t budged. So what’s actually happening? Companies are closing the front door, hiring for workers aged 22 to 25 in AI exposed jobs has dropped 14%. The most exposed workers aren’t factory workers, they’re college educated, higher earning. 49% of US jobs now have at least a quarter of their tasks inside AI’s reach. That’s up from 36% just one year ago. And the red area on that chart, the real world usage is still a fraction of what’s possible. Every month, it grows a bit. Anthropic built the scoreboard and most people haven’t looked at it yet.

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

@AnthropicAI 93.9% on SWE-bench 🤯 When the model becomes so smart it can find its way into most systems - better use it to secure those systems first before releasing it to the world 🌎 smart move Opus → Mythos @claudeai #AI #cybersecurity

Vibe-coding to my Epic Soundtracks 🍿 playlist when @Spotify casually turned my progress bar into a lightsaber for a @StarWars track. lol. Well played 👏 #Spotify #StarWars #EasterEgg #VibeCoding

@FlyaKiet we stand with you. none of us are eating either. cloud ships or we all starve together 🪧

no one is eating at superset until cloud is shipped

Going to have to play with this and think about the impact on future projects / interfaces

@amorriscode Use worktrees for everything, @superset_sh makes them super easy to work with.

I was looking into worktree usage for claude code desktop. A surprising amount of you folks don't use worktress. I wouldn't be able to multi-Claude without them so I'm curious why you don't? How can we make them better or more intuitive for you?