BriannaLovey

heh. If I had the money, I would start building computers for people affected by this, preinstall a stable user-friendly OS like OpenSUSE on them along with all the current TPVs, then give them away. But until then I will have to help them in other ways.

It will be wise to tell them to treat all future releases of Firestorm as alpha/beta-stage software, and to only upgrade when bugs are fixed. For other viewers, this may or may not be applicable.

This is like measuring the rate of obesity in Uganda and saying that the same obesity rate applies to the USA Clarification: What I mean is that you can't draw conclusions about one population by using data from another population.

My point is that they already do. If you look in the viewer source code, you will see that all hardware info is collected by the viewer and transmitted to LL. They just have no incentive to make the data public. The reasons they gather such data are very different from the reason that the average user would want to see the data too.

Her viewer is used by hardly anyone anyways. I was more speaking of those who develop Firestorm or Alchemy. Maybe I will suggest it in their support groups.

Wasn't talking to you, don't respond to me again.

What you can do, is: 1) Join all viewer support groups 2) Find anybody having performance issues or any other bugs 3) Tell them to switch to a different viewer, depending on both their UI preferences and the limits of their hardware. I have been doing this for many people I see struggling, and most I have contacted have had positive results. I usually recommend them either Alchemy (if they are attached to how Firestorm's UI looks,) Cool VL Viewer (if they absolutely need best performance possible,) or Genesis (if the first two suggestions are inadequate.)

Third-party viewer developers need to start gathering anonymized hardware data about their users, and see how much of a priority should be spent on improving performance. We all know LL isn't going to reveal anything about this (because it would reveal crucial information to competitors about the demographic here at best, or reveal to potential buyers/investors that they have made a grave mistake at worst,) and no amount of begging on our part will make this happen.

You can always use a viewer that allows you to turn it off 😄

The other hope is that AI will allow people to develop efficient software more easily, so it may even help reverse the current trend!

Doesn't copilot run on somebody else's computer in a cloud far, far away? I don't think that would impose any extra performance requirements on the user.

Hardware upgrades are needed for "regular apps" because of software bloat, which was ultimately caused by the pressure to lower the labor cost of software development across the industry by paying . This can only go so far though, as moore's law collides head-first into the laws of physics. Also, it depends on the operating system. Most Linux-based operating systems will run just fine on 15-year-old hardware, while the latest Windows will absolutely struggle.

Kitely (an OpenSim grid) used to do something similar, where regions would be shut down whenever nobody was using them in order to conserve costs.

Yeah, that would be an absolute disaster in the long run. I see the incentives that would push them in that direction, though.

The implementation of PBR in unity is more efficient than the implementation of PBR in the SL viewer, so this is not particularly true. It makes me wonder if they plan to eventually make the desktop viewer based on Unity as well in order to eliminate the cost associated with developing a custom rendering engine. The only disadvantage here is that Unity is a proprietary dependency, and LL's fate would be at the whims of whatever pricing model Unity has in the future.

Thanks! I love how reliable you are 😄

heh. I mean I don't need to know it now, but you have given me enough info for me to figure it out later once needed.

I can still see some reflections on the dress and on that metal art piece on the wall, unless those are baked on. If they aren't, is it the case that one would need to modify the rendering engine itself to remove reflections?

Hypothetically, if you wanted to just completely disable reflections from happening at all in the engine, what would have to happen?

Its almost as if making a relatively complicated feature like this for a platform with unpredictable, user-generated content is a bad idea or something. There is a reason that other platforms with user-generated stuff use simpler lighting models. A day hasn't gone by where I am not thankful for Henri for providing us with a way to not use this.

The only solution that doesn't involve throwing more hardware at it is to disable reflections entirely. Something that I suspect a lot of people will be doing anyways.

Nah, NVIDIA has much better things to do.

I mean Jensen Huang, CEO and partial owner of NVIDIA, the company that has a near-monopoly on AI hardware.

Honestly I see this happening. If AI gets good enough to mimic user activity and cheap enough to run without shelling out hundreds of thousands of dollars to the Leather Jacket Man, it will absolutely be used to fill in the void for the lack of users. This is already done on certain multiplayer games where user behavior is easy to replicate without using AI models.

30 years from now, we will all have 1 terabyte of RAM in our computers, and it will be just enough to open 3 chrome tabs