2026-11-04

Real-time long-video generation requires continuous output with consistent appearance and natural motion. Existing chunk-wise autoregressive methods face challenges in both generation efficiency and long-video quality: serialized denoising and history-cache refresh require two forward passes even with one-step denoising, and longer rollouts can degrade quality and cause camera drift. We propose Wave Forcing (WF), which co-designs generation, training, and execution for continuous video output wi

🤖Wave Forcing enables efficient continuous long-video generation with consistent appearance and natural motion through co-designed generation, training, and execution, achieving high throughput and quality.

arXiv:2611.00247v1 Announce Type: new Abstract: Dense labels of the aorta in CT are slow to draw, and the public aortic CT datasets that could stand in for them each label different structures under their own conventions. Large language model (LLM) agents can now write code and read images. We asked whether such a model labels these scans better as a free-form agent that writes its own segmentation code, or inside a harness, where an automated pass labels the scan and the model reviews the resul

Coding agents solve repository issues through long sequences of exploration, editing, and validation. These trajectories are increasingly used as post-training data, so how they are collected shapes what the trained model learns. A common approach is to sample multiple trajectories independently for each issue and then select or curate them. However, as more trajectories are sampled independently, later trajectories become increasingly redundant, even though earlier trajectories could guide wher

🤖SWE-Trail improves trajectory collection for coding agents by using past rollouts to guide exploration without exposing history to the agent, yielding better-trained student models.

NVIDIA shows how 3 stacked optimizations speed up MoE inference on a single DGX Spark. On Ant Group inclusionAI's 124B Ling-3.0-flash, GSM8K decode throughput rose 42.5% with no detectable accuracy change. The post NVIDIA Shows 3 Ways to Speed Up MoE Inference on DGX Spark, Lifting Ling-3.0-flash Throughput 42.5% appeared first on MarkTechPost .

Soniox has released Soniox STT v6, the latest generation of its speech-to-text API. Soniox STT v6 is a speech-to-text model that reads what a voice agent says while hearing the caller. It uses that text to recognize short replies and to drop the agent’s own voice. The release ships as stt-rt-v6 for streaming and stt-async-v6 […] The post Soniox STT v6 Adds Agent Text Input and Echo Filtering for Voice Agents appeared first on MarkTechPost .

I saw HackerTube ( ) a few days ago, the YouTube videos HN keeps recommending, pulled from 19 years of comments. It made me want to flip through them instead of picking, so I built a TV for it. I have Algo fatigue. Each topic area is a channel (DEVTV, SOLDER, RETRO, ORBIT, ...), plus HN Prime (the top 400), a music channel, and Late Night on ch 99, which only broadcasts 11pm-5am your local time and shows color bars otherwise. Similar to HackerTube it was built with Claude Code in a few hours. Al

Hey HN, Ayan here from Finic. We’re launching Kairos, an open source alternative to Sift, Sardine, and other real-time fraud decisioning platforms. Kairos addresses two of the biggest problems facing traditional decisioning tools: a lack of labeled data to train on and dealing with fraudsters that are constantly changing their tactics. 1. Traditional fraud solutions use ML models such as XGBoost to detect fraud, but these models are dependent on quality labeled data which many businesses do not

In a new post today, Valve acknowledged that it's seen "a significant increase in submissions for store and page review in recent months" and says it's "running a little behind" by taking "up to 12 business days" for reviews instead of the typical 3-5 days. Indie developers have recently expressed frustrations over lengthy approval times […]

Article URL: Comments URL: Points: 40 # Comments: 4

I thought it would be fun to browse HN as if it were a newspaper. The site updates once an hour. Enjoy! Comments URL: Points: 451 # Comments: 144

My company is pretty paranoid about leaking IP to the frontier labs through prompts, so we've all been diligent about disabling data sharing in Codex/Claude Code/etc. But yesterday my coworkers and I noticed that the setting had been silently reenabled in Codex. None of us knowingly opted back in. You can check your settings here: Look for "Improve the model for everyone". Curious if this is an isolated issue or something more widespread. EDIT: looks like this has been noted before EDIT: apparen