LogoThread Easy
  • 探索
  • 撰写 Thread
LogoThread Easy

您的一体化 Twitter 线程助手

© 2025 Thread Easy All Rights Reserved.

探索

最新在前,按卡片方式浏览线程

开启时会模糊预览图,关闭后正常显示

前两天接一个「3D 模型 AI 生成技术」的咨询,除了常规聊生成技术的原理和门槛、落地情况,Tripo、混元、字节、Meshy 等不同团队和模型情况...

咨询团队知道我有「3D打印/激光雕刻」这个行业的从业经历,问了我一个问题:3D 模型 AI 生成技术的未来,在不在 3D 打印行业?这个技术发展到什么程度,能促进 3D 打印机进一步普及?

这个问题我有点回答不好,直觉上 AI 生成技术应该能把想象力无限放大,但实际上拓竹 3D 打印机和 Tripo 的合作,没激起太大水花,3D 打印还是实用工具为主,偶有热门 IP 打印的短暂热度,貌似生成技术的创意性和想象力,没有让 3D 打印用户有太大兴趣,至于说促进 3D 打印机普及,就更谈不上了,毕竟他们还有太多打印问题没解决。

感觉 3D 模型生成技术,未来还是在实时生成世界模型上,这个方向对 3D 模型有更广泛和个性化的需求,而且也在快速发展中,不局限于游戏方向。

前两天接一个「3D 模型 AI 生成技术」的咨询,除了常规聊生成技术的原理和门槛、落地情况,Tripo、混元、字节、Meshy 等不同团队和模型情况... 咨询团队知道我有「3D打印/激光雕刻」这个行业的从业经历,问了我一个问题:3D 模型 AI 生成技术的未来,在不在 3D 打印行业?这个技术发展到什么程度,能促进 3D 打印机进一步普及? 这个问题我有点回答不好,直觉上 AI 生成技术应该能把想象力无限放大,但实际上拓竹 3D 打印机和 Tripo 的合作,没激起太大水花,3D 打印还是实用工具为主,偶有热门 IP 打印的短暂热度,貌似生成技术的创意性和想象力,没有让 3D 打印用户有太大兴趣,至于说促进 3D 打印机普及,就更谈不上了,毕竟他们还有太多打印问题没解决。 感觉 3D 模型生成技术,未来还是在实时生成世界模型上,这个方向对 3D 模型有更广泛和个性化的需求,而且也在快速发展中,不局限于游戏方向。

专注 - Context Engineering, AI(Coding)Agents. 分享 - AI papers, apps and OSS. ex Microsoft MVP 合作 - 私信/邮箱:shaomeng@outlook.com 📢 公众号/小红书: AI 启蒙小伙伴 🔗 信息卡提示词 🔽

avatar for meng shao
meng shao
Sun Nov 02 12:41:54
Knowledge is a person’s true beauty—an invisible, secure treasure. It is the teacher of teachers, the companion in foreign lands, and the supreme divine blessing. Among kings, the learned are honored; the unlearned, even if rich, are like animals.

Knowledge is a person’s true beauty—an invisible, secure treasure. It is the teacher of teachers, the companion in foreign lands, and the supreme divine blessing. Among kings, the learned are honored; the unlearned, even if rich, are like animals.

RL and efficient distributed pretraining • eXperiments lab • memes and training lores

avatar for tokenbender
tokenbender
Sun Nov 02 12:40:03
Growing up is knowing ‘candid shots’ are the most well planned.

Just like ‘no-makeup makeup’.

Growing up is knowing ‘candid shots’ are the most well planned. Just like ‘no-makeup makeup’.

AI @amazon. All views personal!

avatar for GDP
GDP
Sun Nov 02 12:39:42
RT @mariagalanis: @skominers 🫶

RT @mariagalanis: @skominers 🫶

Market Design/Entrepreneurship Professor @HarvardHBS & Faculty Affiliate @Harvard Economics; Research @a16zcrypto; Editor @restatjournal; Econ @Quora; … | #QED

avatar for Scott Kominers
Scott Kominers
Sun Nov 02 12:37:32
I really look forward to Anglos getting their «natural masters of the seas» LARP beaten the hell out of them. Can't build ships worth a damn but can preen about them. At this point they're down to pure magical thinking about Blue and Not-blue Water Navy. It's supremely annoying.

I really look forward to Anglos getting their «natural masters of the seas» LARP beaten the hell out of them. Can't build ships worth a damn but can preen about them. At this point they're down to pure magical thinking about Blue and Not-blue Water Navy. It's supremely annoying.

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Sun Nov 02 12:35:05
I think many people don’t fully realize why this CUDA-based implementation in LichtFeld Studio is such a big deal. I literally have control over the entire stack, including full memory management. As far as I know, there’s no other open-source solution ( not sure about commercial) that offers this level of control.

Sure, you can hack LibTorch’s memory management or replace the allocator, but you still won’t get full control. Maintaining something like that for a small project would be nearly impossible. And we haven’t even talked about the build times.

I’m really excited to see what this will enable and whether we can get large adoption of LichtFeld Studio once we can generate small binaries.

I think many people don’t fully realize why this CUDA-based implementation in LichtFeld Studio is such a big deal. I literally have control over the entire stack, including full memory management. As far as I know, there’s no other open-source solution ( not sure about commercial) that offers this level of control. Sure, you can hack LibTorch’s memory management or replace the allocator, but you still won’t get full control. Maintaining something like that for a small project would be nearly impossible. And we haven’t even talked about the build times. I’m really excited to see what this will enable and whether we can get large adoption of LichtFeld Studio once we can generate small binaries.

Your guide to radiance fields | Host of the podcast @ViewDependent | Founder and CEO of https://t.co/5MjtfpwEU3 | discord: https://t.co/lrl64WGvlD

avatar for MrNeRF
MrNeRF
Sun Nov 02 12:35:04
  • Previous
  • 1
  • More pages
  • 1352
  • 1353
  • 1354
  • More pages
  • 2137
  • Next