I think a bandaid fix would be a website flag like movement items has, a better solution would be a proper tagging system per-trigger but the DX sounds difficult. maybe give world devs 2 layermasks, one for avi raycast and one for avi trigger raycast?
Still asking for that new character controller to be looked at again!
3 Likes
as a person thats had a stroke so cant physically speak, your restrictions on text chat, are you going to apply the same to voice users who go running around saying the hard R?(for example) or is this going to be an ablest 1 sided thing that only applies to the deaf/mute community only?
1 Like
If you need to play a specific frame precisely, you should copy the puppet float parameter to an int parameter and use the latter to drive a 1D BlendTree. And you only need to sync the int parameter.
1 Like
It’s not a good idea to inform the uninformed about things.
So Um… if that’s the case, im assuming it was a cake walk to get VRC running on android in general right?
1 Like
I hope work on player voice isn’t “feature complete” because there are still some notable feature requests related to world creation
- Allow udon to change what position a player’s voice is coming from
- Allow adding effects to player’s voice like reverb or high-pass
- Allow udon to know if a player is talking or not and how loud they are
- Allow udon to know a player’s voice latency
Many big creators in the canny comments saying how great these features would be
6 Likes
Forcing dunce avatars on bad players has finally become real
1 Like
I feel this is a big step in the right direction. Now users who want to follow the rules but just so happen to make a mistake, now have the opportunity to learn and grow.
Questions
Will the Safety Center display copyright strikes against your account, include an ability to submit a retraction request with the person who submitted the copyright removal request, and provide means to submit a counter notification from the Safety Center?
Will copyright strikes be handled and displayed separately from Community Guidelines strikes? The authority to issue copyright claims against an account is not necessarily the Trust & Safety team for interpretations of the Community Guidelines violations, and the interpretation of copyright infringement typically comes from the claimant and not VRChat. The intent I seek is to not have copyright strikes stack up with Community Guidelines strikes at the same time. The appeals (retraction requests, counter-notifications) also go to a different authority (the claimant) than the Trust & Safety team, because VRChat is not a direct party to a copyright dispute and may only be legally required to remove repeat infringers from its platform.
In general will warnings start expiring immediately (upon viewing a warning or upon issuance), or only after taking an optional policy/copyright training?
The screenshots of the new Safety Center do not display a way to appeal a warning. Will it be possible to appeal warnings from the Safety Center, or does this require submitting a web ticket to Trust & Safety?
Feedback
Please separate Community Guidelines strikes from copyright strikes on account status in the Safety Center.
Please make it a plausibility to appeal Community Guidelines warnings if the user thinks VRChat has made a mistake.
While warnings are currently rare on VRChat and used (mostly?) to deliver notifications about the Crisis Resources page to users reported in the Suicide and Self-Harm category against themselves (to the best of my awareness), warnings in general can become meaningless if there is no optional policy training involved to understand what mistake was made to stop repeating the same mistake.
If a user receives a warning, I think it should stay on the account for 90 days. But if the user violates the same policy again in that 90 days window, I think those warnings may not expire and may be met with additional strikes that expire after N days (preferably 90 days). This is meant to discourage gaming the system by bad actors, and to provide relief for accounts which stay in good standing.
I’d feel bad for a user in crisis who receives multiple warnings about the Crisis Resources page for their own suicidal behavior if it may lead to account bans/restrictions under this new Safety Center system.
Does this mean by any chance we’re getting the Public Ban world used for moderation again, like in year 2018-2019? 
I could see an use case where provocative behaviors/content which occurred in public may lead to a public instance restrictions on first violations/warnings, instead of an account suspension (or restriction) in all instance types as it currently is. The report API (as far as I know) doesn’t provide a means to tell which instance type a report came from, but it can be investigated manually.
When a post is removed by a moderator on the VRChat Ask forum, the submitter does not get any notification or warning to learn from this if the user is not banned outright. Appeals (or questions about post removal) to Trust & Safety team about it get closed without a reply, perhaps because they’re the wrong recipient. Did you know the VRChat Ask forum has its own Guidelines to follow, separate from and additional to the VRChat Community Guidelines to follow? I doubt the Safety Center helps or integrates into here, because the forum moderators are community volunteers on Ask.
I think it should also be (made) possible to appeal a VRChat Feedback ban from the Safety Center. Those without permission to send feedback may have their system_feedback_access user tag revoked in their profile, currently? (Unless I’m mistaken if this is related to an older world review/feedback system.)
3 Likes
I’d say it’d be better to instead just have a method via script to block Avatar VRCRaycasts
I assume it is not directly related, but it reminds me of the funny “Scaler Smoothing” thingies I have seen which supposedly makes avatar scaling look more smooth. Any chance this would be able to fix that too?
You can lose Avatar permissions forever? Was that always a thing?
I guess I’ll slap my two cents here. With my microphone audio I’ve found that I need to change the detection threshold to something ridiculous like 2% so my voice can be heard. I’m not sure why that’s how it works but it doesn’t offer much fidelity on that the cutoff point is.
Also with the frame standalone version of the game, I get why you uploaded the quest version, but it is a bit disappointing. My only request though is that the x86 build work through fex for those of us who want to take our chances. People would still need to go through menus to run that build, so it should be a bit of a deterrent to those who don’t know what they’re doing. I’m not sure what the challenges of doing that would be, but I’ve been told that the debug build runs if it doesn’t have eac running, and they’re able to launch worlds locally. Is EAC incompatible flat out with fex or is there a configuration you can use to get it to run through fex?
1 Like
I like the Safety Center, my main feedback is it needs a section to tell the player explicitly how and why they violated TOS, not just declare that they did.
Doesn’t need to be a big legalese block of text, just a line or two like “Your account/feature was suspended as we found you/your group did XYZ, which violates section X of the TOS”. Not telling the player how they tripped the TOS can sometimes just confuse them, especially when you’re threatening them with a further full account suspension for something they don’t necessarily know how they violated TOS the first time.
And if they’re just acting innocent about something they clearly actually did, it also gives whoever responds something to immediately point at and say “Nice try, no”.
2 Likes
Just as a small question, which is separate from what the AI form covers: Will VRChat ever disclose how it uses AI in development?
The way that some of the things have released recently in a half-baked state and now seeing these left border UI elements (wildly common with Claude), it makes me wonder if VRChat is using AI to genuinely improve work quality, or if it’s using it to just pump out more quantity, especially if it can ease PR.
I say this as someone who uses Claude regularly. I do use LLMs, but I try to be very cautious about anything I release to the public.
I guess in a code base that is closed source and heavily restricted, it just raises extra concerns about platform safety for me.
But it also kind of makes it feel funny, because with the efficiency of LLMs, it theoretically means that we should be able to get the old Udon 2 that was promised. :)
Given that the Linux version already works through Proton (and shockingly well at that), I am not going to be surprised when someone figures out how to get that version working just fine on the frame.
The real question will be if VRChat takes measures to prevent that from happening.
How soon will the beta SDK for avatars be available? I really want to try FT and Raycast.