{"slug": "the-machine-that-finally-wake-up", "title": "The Machine That finally Wake Up", "summary": "A developer building JAZZ, a Linux desktop environment for AI engineering, documented a series of failures encountered while developing the system, including filesystem corruption that broke the bootloader on a dual-boot Windows/Linux laptop, a cloud RTX 4090 rental that arrived without the promised software environment, and seven consecutive failed clean installs from the public repository. Benchmarking the same code on the laptop's integrated graphics versus the rented RTX 4090 showed roughly 155 milliseconds versus 2.4 milliseconds, about a 64x speedup.", "body_md": "I had rebooted the laptop without really thinking about it. Just a normal restart. I was expecting the usual logo, a few seconds of waiting, then the desktop.\n\nNothing happened.\n\nNo logo. No error. No helpful message. Just a black screen.\n\nThat was the point where I stopped treating it like a minor problem.\n\nThis was not a spare machine or a test box. It was the laptop I actually used. My Linux setup was on it, Windows was on the same disk, and so were the files I cared about.\n\nFor a few minutes I just sat there going through the obvious questions.\n\nWhat had I changed?\n\nWhat had I broken?\n\nWas Windows still there?\n\nWere my files still there?\n\nThe slightly embarrassing part was that underneath all of that was another thought:\n\nI had built this system because I wanted something reliable.\n\nAnd now it would not even boot.\n\nThat feeling has come up more than once while building JAZZ.\n\nJAZZ is the Linux desktop I have been working on for AI engineering: running local models, building applications, testing them, red-teaming them and generally spending too much time in terminals.\n\nBut this is not really a post about the feature list.\n\nThe README can do that.\n\nThis is about the parts where it went wrong.\n\nThere were quite a few of those.\n\nThe corruption I never saw coming\n\nThe black screen eventually turned out to be filesystem corruption.\n\nThe shared partition used by Windows and Linux had developed cross-linked clusters, and over time that had damaged the bootloader.\n\nThere was no warning.\n\nNothing gradually got worse.\n\nThe machine worked, and then one day it did not.\n\nI booted from a USB drive and chrooted into the broken system.\n\nThat was one of those Linux moments where you feel both completely in control and slightly terrified because you are typing commands directly into the thing that currently refuses to start.\n\nI rebuilt the boot files manually and ran a filesystem check.\n\nThen I waited.\n\nThe check came back clean.\n\nI rebooted.\n\nThe machine started.\n\nI remember actually exhaling when the login screen appeared.\n\nNobody else was in the room, which probably made the reaction more honest.\n\nPaying by the hour to find out whether I was right\n\nA few months later I had a different problem.\n\nI wanted to know whether the AI side of JAZZ actually behaved properly on a serious GPU.\n\nMy laptop could run the code, but integrated graphics were not going to tell me much about real performance.\n\nSo I rented an RTX 4090 in the cloud.\n\nThat introduced a very different sort of stress because the meter was running.\n\nEvery minute I spent debugging was literally costing money.\n\nI had also read enough about cloud GPU rentals to know that the machine you think you are renting is not always quite what you get. Sometimes the environment is more constrained than expected. Sometimes software listed as available is not actually there.\n\nThe instance I chose was supposed to have what I needed installed.\n\nIt did not.\n\nSo I ended up fixing the environment while paying for the privilege.\n\nEventually I ran the same code I had tested on my laptop.\n\nThe laptop result was around 155 milliseconds.\n\nThe RTX 4090 result was 2.4 milliseconds.\n\nAbout 64 times faster.\n\nI stared at the number for a second and then said something out loud to an empty room.\n\nI do not remember exactly what.\n\nProbably not anything sophisticated.\n\nBut that was the first time the performance side of the system felt real rather than theoretical.\n\nGiving the whole thing to a machine that knew nothing about me\n\nThe test I cared about most came later.\n\nI took a second laptop that had only ever run Windows and tried to install JAZZ using nothing except the public repository and the instructions.\n\nNo shortcuts.\n\nNo files copied from my main machine.\n\nNo “I know where this config really lives”.\n\nJust the repo.\n\nIt failed.\n\nI fixed the first problem and started again.\n\nIt failed again.\n\nThen again.\n\nSeven failures in total.\n\nNone of them was dramatic.\n\nThat almost made it worse.\n\nOne dependency was missing.\n\nOne path was hardcoded to something that happened to exist on my machine.\n\nA stale lock file had been left behind after unreliable Wi-Fi.\n\nThere were small assumptions scattered everywhere.\n\nOn my own laptop, I had stopped seeing them.\n\nA clean machine saw every one.\n\nThat was uncomfortable, but useful.\n\nEvery time I fixed something, I went back to the beginning and ran the process again.\n\nI did not want a patch that worked once.\n\nI wanted to know whether someone starting from nothing could actually reproduce the system.\n\nOn the final run, the second laptop booted into JAZZ.\n\nThat was probably the moment I started trusting the project.\n\nNot when it worked on my machine.\n\nWhen it worked on one that had no history with it.\n\nThe bug that turned out to be me\n\nMy favourite failure was the shutdown button.\n\nIt stopped working.\n\nNaturally, I assumed I had broken the code.\n\nSo I spent a good part of the day looking in the obvious places.\n\nNothing made sense.\n\nThe implementation looked fine.\n\nEventually I discovered that the problem was not the shutdown code at all.\n\nIt was how I had been testing the machine remotely.\n\nI had accidentally attached a process to the wrong login session.\n\nThe symptom appeared in the desktop, but the cause was several layers away from it.\n\nThe normal local desktop had never had the problem.\n\nI had created the bug myself, then spent hours trying to find it in the software.\n\nThere is a useful lesson in that, although I did not particularly enjoy learning it at the time.\n\nSometimes the thing you are debugging is your own assumption.\n\nI could have quietly fixed it and moved on.\n\nInstead I wrote it into the build notes.\n\nIt seemed worth remembering.\n\nWhat building it actually felt like\n\nI am not sure there is a neat conclusion to any of this.\n\nBuilding JAZZ did not feel like a smooth sequence of good decisions.\n\nIt felt more like this:\n\nA black screen.\n\nA filesystem check I was nervous to run.\n\nA rented GPU charging me while I fixed something that was supposed to be pre-installed.\n\nA second laptop exposing seven assumptions I had forgotten I was making.\n\nAnd one shutdown bug that was entirely my own fault.\n\nThere were good moments too.\n\nSeeing the machine boot after I thought I had broken it.\n\nSeeing 2.4 milliseconds on the GPU test.\n\nWatching a completely separate laptop install the system from the public instructions.\n\nThose moments felt good precisely because the failures before them were real.\n\nJAZZ is public now, but it is still early.\n\nI expect people will find things I missed.\n\nI hope they do.\n\nIf you install it and something breaks, I would genuinely like to hear about it:\n\nThe repository is here:\n\n[https://github.com/holycowprojects/jazz](https://github.com/holycowprojects/jazz)\n\nThe bugs, fixes and commits are all there too.\n\nI would rather leave that history visible than pretend the project arrived fully formed.", "url": "https://wpnews.pro/news/the-machine-that-finally-wake-up", "canonical_source": "https://dev.to/akashholycowprojects/the-machine-that-finally-wake-up-en2", "published_at": "2026-09-18 07:41:42+00:00", "updated_at": "2026-09-18 07:53:03.413648+00:00", "lang": "en", "topics": ["ai-tools", "developer-tools", "ai-infrastructure"], "entities": ["JAZZ", "Linux", "Windows", "RTX 4090", "Nvidia"], "alternates": {"html": "https://wpnews.pro/news/the-machine-that-finally-wake-up", "markdown": "https://wpnews.pro/news/the-machine-that-finally-wake-up.md", "text": "https://wpnews.pro/news/the-machine-that-finally-wake-up.txt", "jsonld": "https://wpnews.pro/news/the-machine-that-finally-wake-up.jsonld"}}