Brett

CG Geek · XR VR Real-time 3D Content Lead · Enviro & Lighting Artist · Building/Product Visualisation · Details Guy · Virtual Explorer

[Brett @VfxBrett]

Blender? Unreal Engine 5? Unity?
Filmed and acted out in VR with Quest 2.

#blender #UE5 #Unity3d #VR #gamedev #Quest2 #cgi #3DCG

[Brett @VfxBrett]

@NVIDIAGeForce #RTXPowerPlay
#RTXPowersPlay
Just in case 😏

[Brett @VfxBrett]

@NVIDIAGeForce GeForce Season (all year round)
@NVIDIAGeForce

[Oliver Pavicevic🇺🇦 @OlixOliver]

Testing my new Deckard Render for @unity . Now we have some neat features: we can simulate a real chromatic aberration lens effect for bloom. It is a modelling of a real spectral lens behavior for some cool DOF effects. #CGI #3DCG

[Saito｜AISPA Lab代表 @VacationdesignC]

On the surface, OpenAI's statement appears ethical and sincere. However, a structural analysis reveals it to be a linguistic device for the postponement and diffusion of responsibility.

Problem 1: A Lack of Agency

Throughout the statement, there is almost no mention of what OpenAI itself "will do."

Note the language they use:
* "Frontier labs should agree on safety principles and share insights..."
* "Governments should play a critical role in facilitating this through industrial policy..."
* "It is extremely important to continuously measure and report on..."

These are all third-person recommendations about what other parties should do. As the organization leading the field, there is no first-person commitment—no "OpenAI will do this." The only exception is the abstract declaration, "OpenAI is deeply committed to safety."

What does this imply? It means that while acknowledging the risks, OpenAI is not taking concrete responsibility itself.

Problem 2: The Function of the Two-Stage Logic

The statement clearly separates "current AI" from "future superintelligence"

* Current AI: Considered "normal technology," manageable within existing frameworks. "We do not believe that new heavy regulation is appropriate for open-source developers or general users."

* Future Superintelligence: Acknowledged as a "potentially catastrophic risk" requiring innovative responses.

The problem with this structure is that it justifies present-day development while simultaneously appearing ethical by discussing future dangers.

But the crucial points are missing:
* When does "normal technology" transition to "catastrophic risk"?
* How is that boundary determined?
* Who decides?
* At what point will they stop development?

While they mention "slowing development," they never once say, "We will stop." This language functions, in effect, as a permission slip to continue development.

Problem 3: Ignoring the Precautionary Principle

The statement acknowledges that "the risks of superintelligence are potentially catastrophic." It also says, "it is difficult to understand... the impact AI is actually having on society."

In short: uncertainty is high, and the potential damage is catastrophic.

Normally, the precautionary principle is applied in such situations—as seen in nuclear development, genetic engineering, and climate change policy, all of which followed a "be cautious because we don't fully understand" approach.

However, OpenAI's stance is the complete opposite: "We don't fully understand, but we will learn as we go."

They claim:
* "Prediction is difficult."
* "It is hard to grasp in advance."
* "It is extremely important to measure what is actually happening."

This is logically flawed. Should one not be more cautious precisely because the dividing line is unclear? Especially when the consequences are irreversible and catastrophic?

Problem 4: The Discrepancy Between Language and Action

If one looks at the actual behavior of Mr. Altman and OpenAI, the priorities are clear.

Areas of active promotion:
* Announcing a $7 trillion semiconductor investment plan
* Rapid product releases
* Lobbying for deregulation
* Focusing on market share expansion

Areas that remain vague and non-specific:
* Specifying safety milestones
* Defining conditions for halting development
* Accepting independent audits
* Improving transparency (e.g., disclosing model details)

This asymmetry is the core of the problem. The ethical language is abundant, but the actual resource allocation is concentrated on development and commercialization.

↓

[Brett @VfxBrett]

@NVIDIAGeForce Indiana Jones and the Great Circle. Made for uppercase RTX grunt!

Geforce PC Week
#GeForcePCWeek

[Brett @VfxBrett]

@NvidiaANZ #GeForceSummer

[Brett @VfxBrett]

@NVIDIAGeForce @FalconNW GeForce PC Week

[Brett @VfxBrett]

@NVIDIAGeForce @ZOTAC #GeForceSummer Indiana Jones and the Great Circle. Hands down. I'll be able to see raytraced nazi hair follicles stuck to his whip in 4k.

[Brett @VfxBrett]

@NVIDIAGeForce #GeForceSummer Lets goooo!

[Brett @VfxBrett]

@NVIDIAGeForce STEAM'n'GREEN! #GeForceSummer

[Brett @VfxBrett]

@NVIDIAGeForce Upgrade to Green 💚! #GeForceSummer

[Brett @VfxBrett]

@NVIDIAGeForce Seeing green! 💚 #GeForceSummer

[Payton @Pjbomb2]

NATIVE 4k full pathtracing by Yanus entirely using TrueTrace for rendering in Unity3D
10 million triangles, 20 bounces max(which is where a lot of the performance goes)
#unity3d #pathtracing

[Brett @VfxBrett]

@NVIDIAGeForce @DinethAshanD help us out! Be my lucky D! #RTXON

[Brett @VfxBrett]

@NVIDIAGeForce #HalfLife2RTX

[Brett @VfxBrett]

@the_f_key @andytouch Will STP improve mobile VR?
Or will it introduce ghosting like TAA does?

[Oliver Pavicevic🇺🇦 @OlixOliver]

It was a great fun to make this! Spent 4 hours and used #GaussianSplatting with some environments made with @Scaniverse Then added some models of speeders from @tegnemaskin and Scout trooper made by Leonardo Benavides. I was using my #DeckardRender and I'm happy on how it adds a cinematic look with its DOF and motion blur. Everything is rendered in @unity @unitygames @madewithunity
I think that this is a good test on how a #GaussianSplatting can be used for #VirtualProduction.
#StarWarsfanart

[Brett @VfxBrett]

@OlixOliver @unity Little stamp for social posts!
New deckard release in the works?
Excited.

Also a request for a Deckard render queue or camera manager to help manage exports?

[Brett @VfxBrett]

@Mr_TibiT Unitys screenspace GI solution? Or Raytraced?

[Brett @VfxBrett]

@Mr_TibiT Ooh that is a nice result then! 👌