AI Coding Live News — Real-Time 𝕏/Twitter Updates & Breakthroughs

👀82

🔥80

Kimi launches 3D Make: K3.1 Pro drives Blender, CAD tools and Unreal, Unity and Godot to build models, scenes and games

Moonshot AI's official Kimi account announced 3D Make at 03:14 UTC on Oct 30, saying Kimi connects to tools such as Blender and CAD software to create models, assemble scenes and develop games, available now on kimi.ai and Kimi Work. The companion 3D guide says K3.1 Pro was specifically trained to drive 3D software through Python scripts, so specialized tools without MCP support can join the workflow, while meshes such as people and animals are handed to diffusion services like Tripo. Plugins that need locally installed software work only in Kimi Work.

Key Takeaways

Availability: 3D Make is live on kimi.ai and Kimi Work; plugins that need locally installed software are Kimi Work only
Two modeling routes: parametric models (furniture, mechanical parts, buildings) via Python scripts written by K3.1 Pro, and meshes via services such as Tripo, Meshy and Hyper3D
Engines: Kimi says it is adapted to Unreal Engine, Unity and Godot to import models, assemble scenes and build gameplay step by step
Showcase: the celestial palace in the K3.1 Pro launch video was built in Unreal Engine 5.8 from source models totaling 36 million triangles, rendered with Nanite and Lumen
Benchmarks: no 3D benchmarks were published; the guide admits some tasks need several rounds, and rigging and complex materials remain weak spots

Read full story

Linked Models & Agents:

⚡ Kimi K3.1 Pro 🤖 Kimi Work 🤖 Kimi Code 🌐 Moonshot AI

#Kimi#3D Make#Blender#CAD#Unreal Engine#Unity#Godot#Tripo#Agent

Anthropic@AnthropicAI·Collected 7d ago

👀78

🔥92

View on 𝕏 ↗

Anthropic files its S-1 with the SEC for a Nasdaq IPO under ticker ANT; H1 2026 revenue was $16.29B, up 1,300% year over year

On Oct 26 Anthropic announced on its official X account and website that it has publicly filed a Form S-1 registration statement with the U.S. SEC and plans to list its Class A common stock on the Nasdaq Global Select Market under the ticker "ANT". The share count and price range have not been set, and Goldman Sachs and Morgan Stanley lead the underwriters. The prospectus shows revenue of $16.286B for the first half of 2026 (versus $1.163B a year earlier) and $4.585B for full-year 2025. The H1 operating loss narrowed to $693M, but fair-value changes on convertible notes and similar instruments pushed the reported net loss to $80.7B. Customers paying over $100K a year grew from more than 950 to more than 11,900 in twelve months.

Key Takeaways

Listing: Class A common stock on the Nasdaq Global Select Market under ticker ANT; share count and price range not yet set; Goldman Sachs and Morgan Stanley lead the deal (announcement)
Revenue: $41M in 2023, $386M in 2024, $4.585B in 2025, and $16.286B in H1 2026, up 1,300% year over year (S-1 prospectus)
Profitability: H1 2026 operating loss of $693M (vs. $2.135B a year earlier); a $72.4B fair-value change on convertible notes, derivative liabilities and SAFEs drove an $80.7B net loss
Costs: H1 2026 compute and infrastructure spend was $8.945B, 55% of revenue, down from 201% a year earlier
Customers: those paying over $100K TTM grew from more than 950 (Sep 30, 2025) to more than 11,900 (Sep 30, 2026), with more than 1,900 paying over $1M
Governance: the Long-Term Benefit Trust elects a board majority through Class T stock; non-voting Class N shares are held mainly by strategic partners

Read full story

Linked Models & Agents:

⚡ Claude🤖 Claude Code 🌐 Anthropic

#IPO#S-1#Anthropic#Nasdaq#Earnings#Claude

Adina YakupAdinaYakup·Collected 14d ago

👀58

🔥82

View on 𝕏 ↗

VAST open-sources EdenLive, a real-time generative renderer: 5B and 14B weights released, about 24.7 FPS at 720P on a single RTX 5090

VAST, the company behind Tripo, open-sourced EdenLive on GitHub and Hugging Face. A game engine or coding agent runs the rules and world state and outputs a rough reference video; EdenLive renders it in real time into detailed frames, guided by a target first frame and a text prompt. VAST released 5B and 14B weights in bidirectional and few-step autoregressive variants, real-time streaming inference code, and a skill that turns a text prompt into a playable game prototype. Per VAST, the project page shows over 24 FPS, minute-long generation and 480P/720P output; with LightX2V the 5B model reaches about 24.7 FPS at 1280x704 on one RTX 5090.

Key Takeaways

The idea is to split state from appearance: an upstream program or engine keeps the rules and world state, and EdenLive only renders the reference video into detailed frames.
Released: 5B and 14B weights (bidirectional and few-step chunk-AR), real-time streaming inference code, the game-prototype-turbo skill and a frontend workbench.
Per VAST: over 24 FPS, minute-long generation, 480P/720P output; with LightX2V the 5B model reaches about 24.7 FPS at 1280x704 on one RTX 5090.
Built on Wan2.1-I2V-14B (480p) and Wan2.2-TI2V-5B (720p); the full weight tree is about 210 GB, and the verified setup is an 80 GB NVIDIA GPU with CUDA 13.
Licensing caveat: first-party code is Apache 2.0, but the workbench first-frame model FLUX.2 [klein] 9B is non-commercial, and the DINOv3 encoder has its own license.

Read full story

Linked Models & Agents:

⚡ EdenLive⚡ Wan2.1⚡ Wan2.2⚡ FLUX.2 klein🤖 Claude Code 🌐 VAST 🌐 Tripo 🌐 LightX2V

#EdenLive#VAST#Tripo

@PaddlePaddle@PaddlePaddle·Collected 16d ago

👀51

🔥94

Baidu PaddlePaddle Open-Sources FlashMask-3: Hopper 3-Role Warp Specialization and Mask-Aware Distributed Attention with up to 3.71x CP Speedup

Baidu PaddlePaddle introduces FlashMask-3 (arXiv:2610.17901), combining a 3-role warp-specialized Hopper kernel with zero-communication mask localization, HCC-IPO sparsity load balancing, and mask-aware communication overlap. It boosts single-H100 TFLOPs by 45-141% over FlashMask, accelerates Context Parallelism by up to 3.71x over Megatron-Core, and speeds up 128-H100 118B-A14B MoE attention by up to 44.9%.

Key Takeaways

3-role warp-specialized Hopper kernel boosts single-H100 TFLOP/s across 12 masks (4K-128K) by 45-141% over FlashMask and 34-71% over PyTorch FlexAttention
Zero-communication mask localization and HCC-IPO sparsity load balancing achieve 1.80x-3.71x higher throughput than Megatron-Core CP ring attention at CP degrees 4-32
Accelerates attention fwd+bwd by 24.5-44.9% and raises end-to-end 128-H100 training throughput by 7.3-11.0% on a 118B-A14B MoE model at 128K context

Read full story

Linked Models & Agents:

#FlashMask-3#PaddlePaddle#FlashAttention-3#Context-Parallelism#Hopper-GPU#CUDA

Hugging Face Daily Papers@HuggingFace·Collected 19d ago

👀92

🔥97

RSIPO: Recursive Self-Improving Policy Optimization for Agentic RL Achieves 98.6% Success on ALFWorld via Skill-Policy Co-Evolution

To improve credit assignment in long-horizon agentic tasks, combining reinforcement learning with verifiable rewards (RLVR) and on-policy self-distillation (OPSD) has emerged as a leading paradigm. However, existing pipelines neglect the dynamic evolution of agentic skills and the reliability of privileged teacher signals. Researchers from CUHK and Tencent introduced RSIPO, enabling skills and policies to co-evolve recursively. RSIPO extracts skills directly from trajectory rollouts, optimizes skill generation via replay, and employs Empirical Bayes shrinkage to calibrate student trust in teacher logits. Operating on Qwen2.5-7B, RSIPO achieves 98.6% success on ALFWorld, outperforming standard GRPO by 17.4 percentage points.

Key Takeaways

Co-evolution of skills and policies: the agent continuously extracts skills from executed trajectories, breaking the rigidity of static skill registries
Empirical Bayes shrinkage calibration: dynamically modulates student trust in teacher supervision, shielding the policy from noisy privileged guidance
State-of-the-art 98.6% on ALFWorld: delivers a 17.4 percentage-point improvement over standard GRPO on Qwen2.5-7B, excelling across embodied and search tasks

Read full story

Linked Models & Agents:

⚡ Qwen2.5-7B⚡ Qwen-family⚡ GRPO-Base🤖 Embodied Agents 🤖 Search Agents 🤖 Self-Evolving Agents 🌐 Chinese University of Hong Kong 🌐 Tencent

#RSIPO#Agentic RL#Empirical Bayes#Skill Co-Evolution#ALFWorld#GRPO

TechCrunch@TechCrunch·Collected 58d ago

👀39

🔥95

View on 𝕏 ↗

TechCrunch/Bloomberg: Crusoe reportedly raises $3B at ~$30B valuation

TechCrunch, citing Bloomberg, reports AI data-center/cloud builder Crusoe — which counts Meta, Microsoft, and OpenAI among customers — raised a new ~$3B round at about a $30B valuation. Atreides Management and Valor Equity Partners are co-leading, with Abu Dhabi's Mubadala Capital participating. Bloomberg also reported a ~$13B five-year cloud contract supplying Jane Street with GPUs and AI infrastructure. The raise comes ~10 months after a $1.38B round at a ~$10B valuation last October; Axios previously said Crusoe met Goldman Sachs and Morgan Stanley about a near-term IPO.

Key Takeaways

Reported ~$3B raise at ~$30B; Atreides and Valor co-lead, Mubadala Capital in
Also: ~$13B five-year Jane Street GPU/AI cloud contract
~10 months after ~$1.38B / ~$10B; near-term IPO talks reported

Read full story

Linked Models & Agents:

🌐 Crusoe 🌐 OpenAI 🌐 Meta 🌐 Microsoft 🌐 Jane Street 🌐 Mubadala

#Crusoe#Fundraising#AI infrastructure#Neocloud#Jane Street#data center#IPO