Cedric Born

We may disagree. I might even be wrong. But let's keep talking.

[Rahul @sairahul1]

Anthropic shipped 125 settings for Claude

The official docs cover 40.

One developer found the other 85.

His API bill dropped from $340 to $87.

Not by using a cheaper model.
Not by writing shorter prompts.

By moving one line in a config file to the right place.

> memory scoped per project → past clients never bleed into new work
> Extended Thinking on Light by default → 18–25% fewer Opus tokens in week one
> cache_control moved to the right line → the fix that turned a $340 bill into $87
> plugins and MCP servers toggled off when idle → saved 25–40K tokens per session
> per-project model override → Haiku for docs, Sonnet for infra, Opus only where it matters

Same model. Same prompts. Same work.

Most Claude users are running a $100/month tool at 30% of its actual capability.

Here are 25 features, workflows, and tricks that close that gap ↓

Bookmark this.

[Wildminder @wildmindai]

3D scene reconstructions by NVIDIA.
ArtiFixer - repairs artifacts and extends sparse views via Wan 2.1.
- high-fidelity inpainting in occluded regions
- gens hundreds of consistent frames in a single pass
- 3D Gaussian Splatting for navigable scene reconstruction

Makes the 3D environment look photorealistic and fully navigable for VR/AR. It basically turns a broken 3D model into a polished, professional scene.

https://t.co/weOQcfXleO

[Wildminder @wildmindai]

PEEK - efficient frame selection for video captioning.

- 5% overhead
- outperforms MaxInfo
- high-accuracy captioning with only 1–2 frames

https://t.co/CaMdrqqX3I

[Sadao Tokuyama @tokufxug]

Ultralyticsの水中映像から魚やサメを正確に切り出すAI「YOLO26-seg」

従来の四角い枠での検出とは異なり、個体の正確な形をピクセル単位で識別する「インスタンスセグメンテーション」を実行。

海洋研究や生態系監視に役立ち、Pythonを用いてモデルの訓練や予測が実行可能です。詳細はリプ欄。

[Will Eastcott @willeastcott]

Enter the Backrooms - in 15 minutes
🪄 Upload a movie still to @SpAItial_AI Echo-2
⏬ Download a 3D Gaussian splat PLY file
🥽 Upload to @PlayCanvas SuperSplat and enter VR
[1 / 3]

[Wildminder @wildmindai]

Lumos-Nexus. Generates high-quality vids that strictly follow complex, logic-heavy instructions.

uses a training-free inference strategy, hands off generation from a small semantic initiator to a large detail refiner.
- Omni-Video + Wan2.1+Qwen3-VL
- model-agnostic
- 1.2x latency overhead

https://t.co/heyejPNLpG

[Sadao Tokuyama @tokufxug]

360度パノラマ写真から高精度な3D空間を復元するAI「PaGeR」

ETHやGoogle等が開発。本モデルは360度パノラマ1枚から3D点群を生成します。

屋内・屋外で最高精度を記録し、特に従来のAIが破綻する屋外において他モデルと最も大きな性能差を示しました。

詳細はリプ欄。

[Will Eastcott @willeastcott]

Splats are getting larger and more detailed. The dataset for this scan of Koriyama Castle in Japan is half a gigabyte. But it loads nearly instantly. Imagine being able to visit perfect digital copies of any location in the world - this is where SuperSplat is heading.
[1 / 2]

[Sadao Tokuyama @tokufxug]

ゲームのキャラメイク感覚で3Dアバターを編集できる新技術「COSY」。

髪や顔を独立したパーツとして生成し、他を崩さず色や形を個別変更可能。

検証では類似手法を上回る画質精度（FID:4.42）を記録。NVIDIAらの共同研究。詳細はリプ欄。

[Rourke Heath @rourke_heath]

This AI workflow is gold for relighting videos 😲

Check out Beeble: https://t.co/G0fdNNuVCe

Beeble lets you replace lighting, backgrounds, and even clothing in existing footage using a node-based AI workflow, transforming any room into a cinematic set in just a few clicks. It’s a game-changer for creators, filmmakers, and video marketers who want studio-quality visuals without a studio budget.

Our mission is to create 100,000 new jobs for creatives, looking to learn and stay ahead in AI? Join GenHQ:
https://t.co/g0a6tIwUIv

#BeebleAI #AIRelighting #AIVideoEditing #CreativeAI #GenerativeAI

[Tulsi Soni @shedntcare_]

🚨 Someone just built the open-source tool the internet has wanted for YEARS.

It’s called ReClip — and it downloads videos from almost any website.
100% free. Self-hosted. No ads. No tracking. No nonsense.

Paste any link from:

→ YouTube
→ TikTok
→ Instagram
→ Twitter/X
→ and 1000+ more sites

Choose MP4 or MP3, pick your resolution, and download instantly.

That’s it.

No popups.
No rate limits.
No forced signups.
No “premium” paywall halfway through.

Just your machine, your files, your privacy.

What makes it even better:

↳ Batch download multiple links at once
↳ Clean browser-based UI
↳ Lightweight setup anyone can self-host
↳ Built with simple HTML
↳ MIT licensed
↳ Open source forever

And the craziest part?

It went from 0 → 1.4K GitHub stars in just 9 days.
Already crossed 239 forks.

That kind of momentum says everything.

For the last decade, people have been stuck using sketchy downloader websites packed with ads, trackers, and fake buttons.

ReClip fixes that.

A clean, private alternative built for people who care about ownership, speed, and freedom.

No data harvesting.
No subscriptions.
No upsells.

Just download what you need—on your own terms.

Open source wins again. 👇

[Pierrick Chevallier | IA @CharaspowerAI]

Everyone talks about video generation models…
But the real insane combo right now is Seedance 2 + Astra (Hyperion 2) in @topazlabs

Generate cinematic footage then instantly hit it with HDR, upscale and pro post-production in browser.
This is next level.The real pro difference is happening more and more in post-production.

Team Generation or Team Post-Prod?

[A.I.Warper @AIWarper]

So close.... Google Omni

[3DxDEV @3DxDEV7]

Check out this city generator addon — it’s a really cool tool for city building and urban design work.

Link In Comments👇

#Blender #CityBuilding #3DArt #UrbanDesign #Gaming

[DeepTechTR 🇹🇷 @DeepTechTR]

🚨 NVIDIA’dan sessiz sedasız gelen bir 3D devrimi: **Lyra 2.0**

Tek bir statik görüntüden devasa, kalıcı ve tamamen keşfedilebilir 3D dünyalar yaratıyor.

En büyük sorunu çözdü: Temporal drifting. Yani AI’nin ürettiği videolarda ve 3D modellerde zamanla oluşan o meşhur “bozulma, kayma, erime” problemi. Lyra 2.0, her karede katı 3D geometriyi koruyor ve kendi hatalarını kendi düzeltiyor.

Bir fotoğrafı alıp içine girebileceğiniz, gezebileceğiniz, keşfedebileceğiniz canlı bir dünyaya dönüştürmek artık gerçek.

Bu sadece teknik bir ilerleme değil… Yaratıcılığın, tasarımın ve hikâye anlatımının sınırlarını yeniden çiziyor.

[Wildminder @wildmindai]

Another cool stuff from NVIDIA.
LocateAnything - high-speed visual search engine. You provide a text prompt and it instantly pinpoints that object's exact location in an image.

- 10x speedup for dense object detection
- Qwen2.5-3B + Moon-ViT
- Fast/Slow/Hybrid modes
- trained on 138M samples for UI, docs, generic grounding.

https://t.co/bEvD6pRKaR

[Riley Brown @rileybrown]

Bro, what????

[Wildminder @wildmindai]

SUPIR upscaler is outdated. ASASR- turns blurry, low-quality photos into sharp, high-res images. Prevents the fake hallucinated details.

- improves OCR
- high segmentation accuracy
- based on FLUX.1 dev

this looks sweet

https://t.co/kx4A4N1wDc

[huangserva @servasyy_ai]

微软发布了一个 4B模型，可以在 3 秒内把任意图像转换为 3D 资产。

它使用一种名为 O‑Voxel 的新几何格式，并且可以在 CUDA 上用不到 100 毫秒转换为带纹理的网格模型。

输出为 GLB 文件，包含完整的 PBR 材质，可直接用于 Blender、Unity 和 Unreal。

完全开源👇

[MrNeRF @janusch_patas]

RT-Splatting: Joint Reflection-Transmission Modeling with Gaussian Splatting

Contributions:
• We introduce a unified surface-volume Gaussian scene representation for jointly modeling sharp specular reflections and clear transmission in real-world scenes containing thin semi-transparent surfaces.

• We propose Specular-Aware Gradient Gating to suppress misleading gradients from complex specular regions, substantially reducing floaters in the transmission branch.

• Extensive experiments demonstrate that RT-Splatting significantly outperforms prior methods while maintaining real-time rendering and enabling flexible scene editing.