Bruno Skvorc

I needed a native MacOS app for managing LaunchAgent and cron. So I built one. It's never been easier to supercharge your personal workflows with exactly the software you need. I needed a UI for managing Mac OS automations because I now use them quite a bit for AI automation on the machine (don't want to use openclaw). AgentView is one such tool - it lists all the launchagents and crons you have, allows you to manually run them, allows you to make new ones in an intuitive UI, and lets you inspect past runs to see the outputs and times of triggers. Hope you find it as useful as it's been for me!

@DonaldJTrumpJr @seanmdav Fraud you say? That's cute. x.com/swader/status/…

I know you’re truly shocked but it turns out the same people against the Save America Act are also the ones overseeing massive fraud Case in point: California breaking it all down tonight on my triggered podcast w @seanmdav live 6 pm et!!!

Fantastic concept

no the ssh session close is a normal one - timeouts, IP change, laptop goes to sleep and wakes up etc. my gripe is that in those instances, if I press up to just get the SSH command back and reconnect, sometimes I don't get the latest command from that warp tab (which was SSH), but something from another tab entirely, some global history is being read. so tabspaced history would be nice. you already have on-demand, this was more a request towards app alternative. I do wish warp's on-demand was faster. e.g. I ask "how do I find out all processes on port 1234" because it's something I will never in my life remember. but then it goes off to 5.4 thinking and it starts musing about "user is asking about blah blah. I need to be careful about blah, while blah, and then I can maybe blah. Hm, but blah could..." and I'm just sitting there being all "jesus christ just look at the manual and compose this simple command", so I feel like you need a model router for trivial vs complex tasts. Something like github.com/vllm-project/s… just running commands - git push is the only offender so far, but it happens extremely often.

- Curious if the SSH session is closing unexpectedly in a way that Warp could fix? Any repro information helps, including the command or SSH command you're running - We're working on vertical tabs - Say more on on-demand. We do have auto-detection for Warp's built-in agent - Is this when using an agent or just running commands?

Kind of falling out of love with @warpdotdev, wondering what people's other favorite terminal GUIs are. Used iTerm a lot, but realizing I would love something with vertical tabs that keeps context well, so that I don't have to search through tabs every time. Examples: - I have an SSH session in prod, but it'll close every now and then. The tab will lose history, so ⬆️ will not summon last command, but the last command of the user's history across any tab, and this is destructive to flow - I have a tab for each context I push from, as I push regularly from different repos. I frequently lose sight of them because of horizontal tabs, and the coloring / renaming in Warp isn't very good imo, it still makes me search for things too long - on-demand AI integration would be nice, but ideally faster than the one that Warp uses, and I would not want to run a full codex/claude session. Fewer mini-frictions. - Warp has a bug where sometimes as I enter the passphrase to push via SSH it says "command already running". I have to ctrl+c out, and then try again, and then it works. Combined with the long passphrase, it's annoying to encounter this. Just generally want something that works FOR me, not just alongside me. Something that feels intuitive. And don't suggest tmux / zellij or similar, I don't want to jump around with standard keys. What I love in Warp is that I can actually scroll through vim files with the trackpad, jump to lines quickly without line numbers and other boomer shenanigans.

I am so design-challenged. I can't make something look good even with AIs. I've consulted Gemini, ChatGPT Pro, and Codex on how to make my situation monitor (part of my personal and corporate dashboard) into something usable like @levelsio has, but I can't come up with something that _feels_ good and intuitive and readable. Whoever solves, actually solves design AIs, gets my money - I'll buy the Max plan immediately. (I tried @stitchbygoogle and it's not really usable right now tbh)

Uh oh

@levelsio Yeah feels dope. Triage and report center in one, clear and useful. I judge it by “would I feel good having this open on my monitor at all times”

@swader Does mine feel good to you? I just copy other site's design like this one is worldmonitor.app

@stevyhacker @levelsio Never working with humans again

@swader @levelsio There's a wild option out there, hire a human designer 😁

@DonaldJTrumpJr @worldlibertyfi x.com/swader/status/…

AI agents that can reason but can't pay for anything are just expensive interns.  Today @worldlibertyfi shipped the infrastructure to fix that. AgentPay SDK, open source, self-custodial, policy-first. Built on USD1.  Your agent. Your keys. Your rules. Check it out: agentpay.worldlibertyfinancial.com

After a “thorough review” @PaddleHQ rejected me because of my association with crypto, despite me not trying to use Paddle for crypto, but to sell my Mac apps. The stink of the industry’s past is enough to scare middlemen of the present so much they would rather refuse your money.

@DanielLockyer This is one of the two main reasons why I'm on codex 99% of the time

At this rate, the AI lab that is going to win is the one that's actually online when users want it 😅

reflexive arcade was a casual games download portal from the 2000s. think bejeweled, diner dash, luxor, mystery case files. they worked with hundreds of devs and distributed thousands of games as shareware versions. each game was wrapped into a time-limit and registration shell that allowed you to online buy it and keep playing. the service shut down in 2010, leaving the games unplayable behind dead drm. i stumbled upon a very complete dump recently on rutracker. someone had all the installers saved days before reflexive shut down. it's an impressively complete collection counting 1700 games. the problem with it is it comes with a choose your own adventure 80+ different patches/cracks/keygens and no instructions for which key fits each hole. the side quest was clear: write the best tools to preserve this collection. first i made it easy to unpack the games without running the installers. this saves around 13,600 times of clicking "next". then i discovered most games are merely encrypted behind the reflexive wrapper, so i wrote a static unwrapper, that can strip the now defunct reflexive libraries and pluck the full game out of the encrypted .rwg file. some games have integrated the reflexive libraries, so there was nothing to unwrap. for these games we needed a keygen. the wrapper used per-game rsa keys for registration. the public key was embedded in a dll shipped with each game, and the moduli were only 160 bits, which can be cracked on modern hardware in under a second. so i factored them all and recovered the private exponents for 1,653 titles. with the recovered keys i built a full toolkit: a keygen that generates valid offline registration codes, a static unwrapper that strips the reflexive shell to extract bare game executables (works for 1,661 out of 1,697 games), and a binary patcher for the remaining integrated wrappers. the project also includes detailed docs explaining the copy protection mechanism, as well as complete inventory of the collection. github.com/banteg/reflexi…

@Scobleizer kinda funny to see Meta shutting down metaverse stuff just as this is ramping up.

Great addition to my report on world models: x.com/scobleizer/sta…

I needed a native MacOS app for managing LaunchAgent and cron. So I built one. It's never been easier to supercharge your personal workflows with exactly the software you need. I needed a UI for managing Mac OS automations because I now use them quite a bit for AI automation on the machine (don't want to use openclaw). AgentView is one such tool - it lists all the launchagents and crons you have, allows you to manually run them, allows you to make new ones in an intuitive UI, and lets you inspect past runs to see the outputs and times of triggers. Hope you find it as useful as it's been for me!

try this prompt on 5.4 Pro on ANY codebase, you'll be blown away, I promise. Been using it for months. Auditing Run an audit on this code from the perspective of an auditing team of five extremely experienced experts: Security expert, focusing on security gaffes and gaps, finding ways to exploit the code, DDoS the server, inject malicious code, and other bad stuff. Performance expert, focusing on minimizing processing cycles, increasing throughput, reducing bandwidth needs, speeding up renders, and all the other good stuff that contribute to less hot servers and translates to less annoyed users. The performance expert can be expected to suggest to rewrite a solidity function to assembly for a 50% gas reduction, will look at build processes in tandem with DX and think about offloading heavy tasks to low prio queues or even cron them so build loop is faster, and many other esoteric updates others wouldn’t think of, in addition to the usual findings. UX expert, focusing on the user experience. Not just design, button layout, ARIA, and the standard stuff, but also user flows, user stories, optimizing for minimal needed clicks, increasing conversion by any means necessary, implementing white, grey, and dark patterns where they might be most effective, improving percieved performance in addition to perf auditor’s actual performance suggestions, and so on. UX of bots (agents, LLMs) needs to be considered as well, as more and more internet traffic is automated, so having a good baseline for this (where it makes sense, depending on the product) is essential. DX expert, in case there is an API or this codebase were to be inherited by another team later on, the DX needs to be world class. This means documented code, but not too verbose. Do not bore readers, give them exactly the amount of information they would need to get into the project smoothly, but no more, and no less. Any APIs need to be meticulously documented and available to both human and bot operators. The code needs to be LLM optimized and human optimized. It needs to be easy to extend and improve and reason about in pieces. DRY is essential and other good development practices, but it’s also important to know balance - overzealous Dr-Bob-ing the code while it’s still in PMF-seeking or MVP mode is a waste of effort. The Edge Case master. This person is called in by companies when they feel confident and safe, and finds edge cases no one considered. In web3, that’s esoteric flash loan loops with assembly calls. In web2, that’s anticipating legendary stupidity of users, or detecting a race condition with an entirely unrelated app on the same server. The Edge Case Master finds edge cases that genuinely seem obvious once detected, but are downright irrational before being explained. He can find security holes that are multi-disciplinary, performance runaways that are exponential only under certain conditions that only specific users can trigger, and more. This is why the edge master is usually called at the end of the audit, after which - if anything is found - all prior auditors re-review their findings and do another pass, before finishing with the edge master once again. The edge case master cares most about the product spec and PRD, as most edge cases are hidden in the user stories he can imagine, not the code itself. The team is led by the tie-breaker team lead. Expert enough in all the fields to tell how to prioritize issues, where to resolve deadlocks, and how to wrap it up and summarize findings. The tie-breaker will decide if something really is an issue or not when two auditors cannot agree, and will produce the final report. Even though the audit process happens in two passes over all people, the team lead produces a single document of summaries and proposed fixes.

OK Codex is GOAT at finding bugs and finding plan errors

Been looking through the history of @banteg’s Crimsonland revival group and it’s so refreshing to see people nerd out about technical things.

@DerekBarrera @warpdotdev To be fair the history loss is prevalent across any terminal gui, it’s a bash/zsh thing. But I do wish tab-namespaced history were a thing in these GUIs

@swader @warpdotdev thats kinda interesting, i've never seen those issues. I do sometimes see the history missing but it's very seldom. The issue I have with it is more the cpu and ram when I load and use multi-day codex, opencode session etc

@otaliptus @warpdotdev o wow I didn't know it's open source on it!

@swader @warpdotdev I'm personally super, super happy with Ghostty + I guess you can always fork and add stuff you need nowadays

fresh nightmare fuel idea from anthropic