The care VRChat is putting into this discussion is encouraging, especially its respect for creators who use AI and those who don't. Not everyone will agree, but this feels like a step in the right direction toward an open, productive community.
A few weeks ago, we asked the community to submit feedback on how they felt about AI in VRChat.
Here's our takeaways and next steps:
ask.vrchat.com/t/where-were-…
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
DeepSeek-V4.1-Flash-1017 is genuinely a blast to work with. It surprisingly excels at the vast majority of tasks I've been giving it, not to mention it's super duper fast. 🐳
Huh? Take a look at the charts, lol.
Kimi K3.1 Proが登場。
最新の3兆パラメータ級オープンウェイトモデル。
🔹100万トークンのコンテキスト/オープンウェイト
🔹 Unreal、Unity、Godotでのゲーム開発を一貫して支援
🔹 BlenderやCADでのパラメトリック3Dモデリングに強み
🔹 金融業務にも強み:OfficeQA Pro v2(64.4)、SpreadsheetBench v2(36.3)で最高水準の性能を達成。DCF/LBOモデリングにも一貫して対応
Kimi K3.1 Proは、Kimi.ai、Kimi Work、Kimi Code、Kimi APIで提供を開始しました。 モデルウェイトは2026年11月4日までに公開予定です。
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
Can’t wait to see Kimi K3.1 Pro on KeyGen Bench. @JustMicrock
keygen.micr.dev/
Travis Knight's recent comments about AI got me thinking about its place in creative work. I wrote about how these tools can expand what artists can accomplish, and why human judgment, craft, and control continue to be just as important. 🛠️
That's a notable shift from June, when he told TechRadar that "artificial intelligence can be a tool, or it could be a crutch," and that it isn't a technology to fear. I respect Travis as a director, and the craftsmanship behind Laika's films, but I think his earlier take was the right one. I find the categorical rejection of AI as a tool disheartening, especially considering Laika itself collaborated with Intel to develop machine-learning tools that automate tedious rotoscoping and facial seam cleanup, freeing artists to focus on the creative aspects of their work.
Now, that's not the same as modern generative AI, but it illustrates that the meaningful distinction isn't whether AI is involved, but how it's used and what role human creativity and judgment continue to play.
There's an enormous difference between using AI to generate an entire film with minimal human involvement and using it to help an artist explore ideas, refine concepts, or accomplish something that might otherwise be prohibitively difficult. Look at what Zack London (@Gossip_Goblin) and his team are accomplishing. They're using generative AI to create visually ambitious, narratively rich science-fiction films that would ordinarily require vastly larger production budgets. According to London, their upcoming feature, Gods Don't Give Gifts, was self-financed and employed 30 people over nine months, including writers, voice actors, composers, and post-production artists. The score contains no AI music at all.
These aren't people typing one simple prompt and letting ChatGPT make all the creative decisions. They're building original worlds, writing stories, directing performances, and exercising artistic judgment throughout production. AI provides the means to realize creative ambitions that might otherwise remain financially out of reach.
And this extends well beyond filmmaking. AI has extraordinary potential to accelerate discoveries in science, medicine, and mathematics. Dismissing the technology as a whole because of particular applications ignores that potential.
I also don't think autonomy disqualifies something from being a tool. A technology doesn't stop being a tool simply because it can execute increasingly complex tasks without constant human intervention. It only changes how we need to think about oversight and responsibility.
There can be legitimate arguments to be made about consent, training data, intellectual property, job displacement, and artistic integrity, but I don't think the answer is to draw a sweeping line around an entire technological field and declare it illegitimate.
Yes, it's not an LLM. But Embark's audio team has said the commentators are "AI text-to-speech" built from the actors' voices. That's neural voice cloning, which is generative AI. Obviously, nobody recorded Scotty saying your gamertag, and screen readers using generic synthetic voices is not the same as cloning a performer to generate lines they never recorded. pcgamesn.com/the-finals/ai-v…
Is every popular UEFN game just a reskinned version of something that was trending on another platform, like, a year ago?
New SOTA just dropped.
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work.
At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated.
We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.
Overall, I'd say Krea 3 is a big step up on fine detail in illustration. It does lose some of the loose, painterly charm Krea 2 has, and leans polished even when the prompt asks for the opposite.
But, it's still a really great model so far. I'm excited to dig into the editing tools next and keep testing!






























