Filter
Exclude
Time range
-
Near
Replying to @Vryionic
See ur stuff on Tiktok!! But this is my first impression of you on X lmao. Kinda peak tbh
9
Replying to @Vryionic
Claude code hates him!!1 See how he optimized the optimizer with this one simple trick!
5
158
Can we not recommend AI Generated slop?
5
155
‼️ As someone graduating with a cybersecurity degree, do not run this shit. This is barely reviewed AI-generated garbage and changes settings on your PC I would not ever recommend. (including antivirus folder exclusions, registry edits that do nothing for vr performance, etc)
1
7
380
An the most important thing if u dont have a powerfull oc , CLOSE DISCORD.
2
608
- close background tasks, especially those whom have 100s of browser tabs open. Other than that, just setting the appropriate affinity (also core parking for those newer CPUs). Clearing cache, no background installations or downloads, usually does the trick.
4
207
To be fair, I personally haven't seen a single person say their performance has improved. Only people fighting about how it was made and that the video itself advertises (overly exaggerated) a jump from 10fps to 40fps. Most of the time performance issues go away when you just-
2
5
259
This would have been better if we could see the performance differences. I want to see it work before using it, but it seems no one wants to show that
2
8
1,157
VOS, The First VR Scanner for Your Computer! @Vryionic
14
45
465
19,215
Replying to @Vryionic
You should not have to apologise especially after you released the documentation Your tool is awesome even if it’s a tad odd at times (it recommended I buy 32Gb of ram when I have 128 xD) I love its easy to use features and in depth interface Twitter people are always hating :(
2
248
Replying to @Vryionic
its harsh seeing you receive so much backlash for a helpful tool. if someone doesn't trust it they could COMPILE IT THEMSELVES and view source beforehand. also thank you for documenting what's changed here vos.vryionic.com/docs/
1
3
839
Replying to @Vryionic
I do feel the hate was a bit forced. But your intentions did seem pure
4
647
I understand ya, and despite my objections to using AI for coding, I get where you're coming from SURELY you didn't foresee this blowing up the way it did haha
1
2
131
For coding it doesn't make a whole lot of sense to utilize it for anything where security is a concern since an algorithm isn't all that concerned about security because it's working in a vacuum, it has no idea the code needs to be airtight
1
33
Apr 29
So, after pokin at the code base, here's my two cents on this situation: TL;DR: I do not recommend the use of this. After looking through the code, I don’t think this is malicious, and I don’t think the author deserves a pile-on. But I do think this is a very good example of why people who now are using AI to program HAVE to be careful. This tool touches sensitive parts of Windows like Defender exclusions, registry keys, services, Hyper-V, process priority, and update behavior. A lot of those areas already have built-in Windows systems managing them, and many “optimizer” tweaks online are outdated, and misunderstood, specially by AI. AI can confidently stitch those ideas together into something that compiles and looks polished, but it cannot understand the tradeoffs for you. If you are building tools, any tool, you need to know exactly what everything does, when it helps, when it hurts, and how to safely undo it. First, I do not think this needs to become a pile-on. Vryionic clearly seems to care, and I think the better path here is education, slower development, and more review. That said, I do want to shift the focus slightly. I wanna talk about how I've been seeing an increase of devs and other peeps using AI to make stuff, and how well they actually understand what they're putting together. AI is incredibly powerful. It can generate good code, it can also give you code that looks right, compiles fine, and even works without you really knowing why. This is a MACHINE, quite literally a calculator of probabilities. It cannot stop and ask “should this be done?” or “what happens if this breaks?” And this seems like the perfect learning example. AI can stitch together complex behavior really quickly. That’s powerful. But it doesn’t *know* stuff. It is parroting the most probable answer to the question "Optimize my computer." It does not understand trade-offs, edge cases, or long-term effects, because it *can't*. If you’re building something like this, you should be able to explain every part of it. Every registry change, every service tweak, every PowerShell command, every updater behavior, every failure case. Not just “it works,” but why it works, and what happens when it doesn’t. For example, this is build with typescript. Now its a fine programming language, but TypeScript is a higher-level language optimized for developer productivity and web environments, is not something that you use in somewhere you want to optimize performance, specially when it is supposed to be kept running while in VR. This kind of low‑level tuning tool is better as a small, native utility than as a chunky Electron app produced by generic ‘build me a Windows optimizer’ prompts. But, TS is one of the most used languages in the world, meaning AI will default to it. It *is* the most probable answer when asking to build a project, and if you don't know what programming with this entails, you probably will accept this blindly. Another example: this program sets `SystemResponsiveness=0`. This manages MMCSS, which is the Windows Multimedia Class Scheduler Service. It is used to prioritize time-sensitive multimedia work such as audio and similar tasks. `SystemResponsiveness=0` quite literally does nothing, because it does not disable MMCSS or create a true “0% reserved” CPU mode. On modern Windows, values that low are normalized rather than used literally, so setting it to `0` just falls back to the minimum allowed behavior instead of unlocking extra CPU time or reducing latency in a meaningful way. A LOT of older tweak guides often describe it as a hidden optimization, but in practice it is mostly a misleading setting that usually changes little or nothing useful. (See the trend? Even if it does nothing, it is in the AI models training data, a lot. This means that it will be recommended, confidently too, even if it's wrong) A lot of what this tool is trying to “fix” is already handled by Windows in some form. Windows has Game Mode, power plans, MMCSS scheduling, graphics scheduling, Defender controls, USB power management, update controls, and built-in ways to tune performance. There are already projects like Winutil that centralize this, and do so in a really good manner. Obviously, I'm not saying that nobody should experiment, but with experimentation you need comprehension. Specially when you trust any ML model to do stuff for you. This program does a lot of stuff that I would recommend NOT to do, and in the current state, I do not recommend it's use.
Heyo, this is a long one so buckle up... Wanted to come out and say this as things have definitely spiraled out of control here. In regards to my software VOS that i just released to the public under open source, I never intended for any of this to be strewn this way. My intentions originally were to share a cool tool i made that has helped me and friends, with the public at large. Ive heard the criticisms, and seen your responses to it... I wanted to say Im sorry for how i marketed it, and for how the presentation and ambiguity came off for it. Im relatively new to the Twitter environment, and my other platforms i exist on know me for that style of editing and glamour, but i do realize now that the style i presented it in, without core information, comparisons, or just general care for how others would interpret it, definitely left so much room for distrust and fear without showing what it truly was. I appreciate those of you who came to defend the video i made, but the reality is it wasnt the right kind needed to build trust from nothing in an environment where people are getting hacked left and right. What the tool is, on the most basic level, is a collection of hundreds of fixes and step by step guides for multiple types of VR systems, all inside a dedicated package that can both show you whats wrong with your current build, how to fix it, and contains a live optimizer to help reduce bloat on the main cpu to help prevent overloading. Thats it. At its core, all fixes it does, including the auto fix section, could and can be done by someone if they did enough research or used enough guides. Of which there are plenty of on the internet, and some way more in depth than what is provided currently in VOS. It does not touch anythign sensitive or irreversible on your pc, and does not touch vrchat in any way, so its not against TOS or dangerous in any manner, as has been audited by coders in the comments in the main post. Though my code was sloppy. I did that research on my own. I scoured the web and compiled that data. Why? Because It was originally for me, because my vr setup kept lagging for streams and i wanted to figure out why. And i did, which i dont need to get into here. But as a prior access control specialist, I use tools in my workshop to help push something like that forward and develop faster because i was impatient. Outwardly, yes, "AI" was used to assist, but not in the way you think, and did not write it for me. I openly admit that an agent was used to assist in coding completions, but not as a creative engine where i just told it what to do. I did not prompt the AI to do things for me, it just essentially finished my "sentences" like autocorrect on your phone does. Regardless, using AI in any large open source capacity is unprofessional, as im not a corporate coding slave who has deadlines to meet. Taking the lazy way out of something i already knew how to do was wrong, and im fully aware of its impacts, of which i am actually sorry for... The code has been audited by new friends and critics alike, all of which have given me good advice on best practices and what to do to fix any potential security risks in it as time goes on. I hear all the feedback ive been getting, and am willing to hear more, if it does end up benefiting both my ability to work on software like this, and my ability to help others with something like this, since that was the whole goal. Since the project lives open source on github, I wont be removing it, but I will be urging all of you to remain careful of anything suspicious in any capacity. I realize how all this made both my skills and my intentions seem wrongly aligned, and i truly am sorry for that... Im working on a better demo for the future, showing exactly what it is, what it does, and how it works, so that no one is in the dark about what it is or why it matters (and wont kill your pc). On top of that, im re-coding parts of it myself with the help and advice from those who spoke out to give me pointers to make it more safe. At the same time, im applying for code signatures (meaning the files are verified by companies to ensure sfaety) before ever reaching the public, because the last thing i want is a tool like this being used in any wrong capacity. From this point forward, i will be removing the update function from the software until i get that signature, so that if you choose to update, you can grab it from the public website/github where all changes and updates are logged in the open for everyone to see, and you can feel safe knowing you arent downloading anything harmful. Ill be working in tandem with @VixenVRC and others to verify my code and keep people in the loop, because all i want to do is help others the way my stuff helped me and my friends. Again im sorry about how all this went down, you guys deserved better knowledge and more a more forthright video/marketing (if you even wanna call it that since its fee), ESPECIALLY with all the cybercrime going around. I hope I can build that trust with all of you over time, and if that isnt possible for some, thats fine. I still maintain that if you dont trust it, dont download it, even if its mine. If you have any questions, feel free to ask below. Ill answer.
1
1
2
405
Replying to @Vryionic
Great post. Gotta say the whole paragraph about AI is stupid. Apologizing for using AI to help you on what was initially a personal project and then saying it's bad cus you're not a corpo feels so fucking backwards. One day the world will be normal.
2
5
1,444
Replying to @Vryionic
Good that you recongise the issue here. From what I've seen, people are being overly critical of the tool. I honestly think this wouldn't have happened to this extent if the presentation video was different. You live and learn.
5
631
Replying to @Vryionic
Hope to see this one day in the hands of many, the biggest issue in the community is ownership and professionalism, you've done both here and I'm proud to be your friend and professional peer in the Industry, I look forward to seeing where your project goes!
7
1,296
Go give it a good read, Thank You @Vryionic for talking with me privately and others to show your willing to improve and keep things secure in the community, I look forward to it
Heyo, this is a long one so buckle up... Wanted to come out and say this as things have definitely spiraled out of control here. In regards to my software VOS that i just released to the public under open source, I never intended for any of this to be strewn this way. My intentions originally were to share a cool tool i made that has helped me and friends, with the public at large. Ive heard the criticisms, and seen your responses to it... I wanted to say Im sorry for how i marketed it, and for how the presentation and ambiguity came off for it. Im relatively new to the Twitter environment, and my other platforms i exist on know me for that style of editing and glamour, but i do realize now that the style i presented it in, without core information, comparisons, or just general care for how others would interpret it, definitely left so much room for distrust and fear without showing what it truly was. I appreciate those of you who came to defend the video i made, but the reality is it wasnt the right kind needed to build trust from nothing in an environment where people are getting hacked left and right. What the tool is, on the most basic level, is a collection of hundreds of fixes and step by step guides for multiple types of VR systems, all inside a dedicated package that can both show you whats wrong with your current build, how to fix it, and contains a live optimizer to help reduce bloat on the main cpu to help prevent overloading. Thats it. At its core, all fixes it does, including the auto fix section, could and can be done by someone if they did enough research or used enough guides. Of which there are plenty of on the internet, and some way more in depth than what is provided currently in VOS. It does not touch anythign sensitive or irreversible on your pc, and does not touch vrchat in any way, so its not against TOS or dangerous in any manner, as has been audited by coders in the comments in the main post. Though my code was sloppy. I did that research on my own. I scoured the web and compiled that data. Why? Because It was originally for me, because my vr setup kept lagging for streams and i wanted to figure out why. And i did, which i dont need to get into here. But as a prior access control specialist, I use tools in my workshop to help push something like that forward and develop faster because i was impatient. Outwardly, yes, "AI" was used to assist, but not in the way you think, and did not write it for me. I openly admit that an agent was used to assist in coding completions, but not as a creative engine where i just told it what to do. I did not prompt the AI to do things for me, it just essentially finished my "sentences" like autocorrect on your phone does. Regardless, using AI in any large open source capacity is unprofessional, as im not a corporate coding slave who has deadlines to meet. Taking the lazy way out of something i already knew how to do was wrong, and im fully aware of its impacts, of which i am actually sorry for... The code has been audited by new friends and critics alike, all of which have given me good advice on best practices and what to do to fix any potential security risks in it as time goes on. I hear all the feedback ive been getting, and am willing to hear more, if it does end up benefiting both my ability to work on software like this, and my ability to help others with something like this, since that was the whole goal. Since the project lives open source on github, I wont be removing it, but I will be urging all of you to remain careful of anything suspicious in any capacity. I realize how all this made both my skills and my intentions seem wrongly aligned, and i truly am sorry for that... Im working on a better demo for the future, showing exactly what it is, what it does, and how it works, so that no one is in the dark about what it is or why it matters (and wont kill your pc). On top of that, im re-coding parts of it myself with the help and advice from those who spoke out to give me pointers to make it more safe. At the same time, im applying for code signatures (meaning the files are verified by companies to ensure sfaety) before ever reaching the public, because the last thing i want is a tool like this being used in any wrong capacity. From this point forward, i will be removing the update function from the software until i get that signature, so that if you choose to update, you can grab it from the public website/github where all changes and updates are logged in the open for everyone to see, and you can feel safe knowing you arent downloading anything harmful. Ill be working in tandem with @VixenVRC and others to verify my code and keep people in the loop, because all i want to do is help others the way my stuff helped me and my friends. Again im sorry about how all this went down, you guys deserved better knowledge and more a more forthright video/marketing (if you even wanna call it that since its fee), ESPECIALLY with all the cybercrime going around. I hope I can build that trust with all of you over time, and if that isnt possible for some, thats fine. I still maintain that if you dont trust it, dont download it, even if its mine. If you have any questions, feel free to ask below. Ill answer.
3
271