PixVerse said it tried a complete parkour sequence on the platform with Kling 4.0; the movements from takeoff to landing were clean, momentum was maintained throughout the run, and the full prompt was included.
Matched fieldsTitle
Xuanjia Technology has completed its B1 funding round; about 30 days later, the B2 round was jointly invested by Shenye Capital, Fachuang Capital and Jiangbei New Area Science and Technology Investment Group, bringing cumulative funding to over 200 million yuan, with valuation expected to exceed 2 billion yuan. Founder Qin Lin says competition in AI video has shifted to industrialized mass-production capability, and the production cost of a 120-minute short drama can be compressed from 300,000 to 500,000 yuan to about 5,000 yuan. The company's AI business revenue is growing at about 50% per month, and the platform has over 100,000 online registered users.
Xuanjia Technology has completed its B1 funding round; about 30 days later, the B2 round was jointly invested by Shenye Capital, Fachuang Capital and Jiangbei New Area Science and Technology Investment Group, bringing cumulative funding to over 200 million yuan, with valuation expected to exceed 2 billion yuan. Founder Qin Lin says competition in AI video has shifted to industrialized mass-production capability, and the production cost of a 120-minute short drama can be compressed from 300,000 t
Matched fieldsTitle
The 2027 Tencent Video V Vision Conference was held in Shanghai, where more than 100 premium titles awaiting release were announced, along with a commitment to continue upholding a premium content strategy. Tencent Vice President Sun Zhonghuai stated that AI has made "producing" easier, but has not made "producing well" easier, and that the value of top-tier premium content and major IPs will become more prominent. The platform is exploring a "live-action + AI" hybrid production model and has launched WorkRally, a film and television industry-grade AI workbench.
The 2027 Tencent Video V Vision Conference was held in Shanghai, where more than 100 premium titles awaiting release were announced, along with a commitment to continue upholding a premium content strategy. Tencent Vice President Sun Zhonghuai stated that AI has made "producing" easier, but has not made "producing well" easier, and that the value of top-tier premium content and major IPs will become more prominent. The platform is exploring a "live-action + AI" hybrid production model and has la
Matched fieldsTitle
Matt Palmer demonstrated opening Chrome and real Slack on the Linux remote desktop of a Cursor Cloud Agent, having the Agent generate a Slack UI motion video with code, and then iterating and revising after checking easing and motion patterns against the real interface. He broke the process down into an inner loop (the Agent does the work, checks against the real thing, and fixes) and an outer loop (a human provides references, preferences, and permissions and accepts the output), and proposed the Do→Codify→Generalize ladder, packaging one success into a Skill and then generalizing it. He emphasized that Slack was not decompiled; he merely gave the Agent a reference and asked it to extract visual patterns by comparison.
Matt Palmer demonstrated opening Chrome and real Slack on the Linux remote desktop of a Cursor Cloud Agent, having the Agent generate a Slack UI motion video with code, and then iterating and revising after checking easing and motion patterns against the real interface. He broke the process down into an inner loop (the Agent does the work, checks against the real thing, and fixes) and an outer loop (a human provides references, preferences, and permissions and accepts the output), and proposed t
Matched fieldsTitle
Google Gemini has launched a personal avatar feature, allowing users to create and share avatars and appear together with friends in images and videos. After sharing, users can tag each other in prompts while maintaining consistency of characters and voices; users can also add or remove others' access permissions at any time.
Matched fieldsTitle
Sunra has released Sunra API, which lets users call all major image, video, audio and 3D models with a single key, billed at each model vendor's own API price, on a pay-as-you-go basis, with no subscription fee and no markup. Rohan Paul relays that Seedance 2.5 costs half as much on Sunra as on two other gateways, and that this unified entry point also makes it easy to run A/B tests on models from Google, OpenAI, ByteDance and Kling.
Matched fieldsTitle
Perplexity's official account showcased a 3:45 explainer video generated by Computer, explaining what changes Apple and Foxconn made at the underlying layer of the iPhone Duo folding screen to achieve mass production. Aravind Srinivas reposted it and commented that explainer videos have become his favorite way to learn.
Matched fieldsTitle
Artificial Analysis launches the image-to-video evaluation benchmarks AA-Video-I2V v2.1 and AA-Video-I2V-Silent v2.1, ranking models at up to 1080p resolution across 10 use-case categories, 10 capabilities, and multiple visual styles. The two versions are based on more than 58,000 human preference votes across over 1,000 prompts and more than 44,000 votes across 500 prompts, respectively, with only 5% of prompts made public in the public arena and rotated monthly. In the first version of the leaderboard, Dreamina Seedance 2.5 ranks first overall, statistically tied with the Seedance 2.0 and MiniMax H3 series, at a price of $34.12/min; the open-weight MiniMax H3 (768p) ranks third, at a price of $4.80/min.
Artificial Analysis launches the image-to-video evaluation benchmarks AA-Video-I2V v2.1 and AA-Video-I2V-Silent v2.1, ranking models at up to 1080p resolution across 10 use-case categories, 10 capabilities, and multiple visual styles. The two versions are based on more than 58,000 human preference votes across over 1,000 prompts and more than 44,000 votes across 500 prompts, respectively, with only 5% of prompts made public in the public arena and rotated monthly. In the first version of the lea
Matched fieldsTitle
The author threw a practical, substantive article at Claude Opus 5.5, wrote just one prompt—"make me a video based on this article"—and completed the entire video. He believes Opus 5.5 has compressed a pipeline that previously required four or five specialized roles and several days into something one person can turn into a finished product with a few prompts in a few minutes, and asserts that this is the best era for ordinary people to do video self-media, as well as the last leverage window period the era has left for individuals.
The author threw a practical, substantive article at Claude Opus 5.5, wrote just one prompt—"make me a video based on this article"—and completed the entire video. He believes Opus 5.5 has compressed a pipeline that previously required four or five specialized roles and several days into something one person can turn into a finished product with a few prompts in a few minutes, and asserts that this is the best era for ordinary people to do video self-media, as well as the last leverage window pe
Matched fieldsTitle
PixVerse announced that Kling 4.0 is now available on its platform, capable of generating videos up to 30 seconds long in a single pass without stitching. This version can also maintain character consistency across multi-shot scenes and sustain stable motion during high-speed action and complex camera movements. The official complete prompt link was also provided.
Matched fieldsTitle
Rohan Paul commented on the Taku team craftsmanship video released by Dyna Robotics, saying that this video illustrates the company's atmosphere better than any recruiting page. In the quoted post, Dyna Robotics explained that the name Taku comes from the craftsman Takumi, and that behind it is a team that loves what they do and pursues excellence.
Matched fieldsTitle
PixVerse has announced that Kling 4.0 is now integrated into the platform, enabling the generation of videos up to 30 seconds long with smoother dynamic motion. Ultra members can enjoy 10 days of unlimited generation.
Matched fieldsTitle
Peter Yang demonstrated how a few prompts can get Claude to call HeyGen to generate multiple virtual characters and record talking clips, assembling them into a creator video wall. According to Jake Moran, he didn't even ask to be included in the frame—Claude decided on its own to add his avatar in the center position. The full tutorial is included in a 40-minute AI video editing program.
Matched fieldsTitle
Peter Yang and Jake Moran recorded a 40-minute episode demonstrating how to use Claude Opus together with HyperFrames to generate videos: creating a launch video from just a few images, replicating a creator's video style, and making precise edits using keywords. He emphasized that the key lies in how the prompts are written.
Matched fieldsTitle
In the latest SuperCLUE I2V (image-to-video) model overall ranking, MiniMax (Xiyu Technology) H3 took first place in China with an overall score of 71.46, and its application capability score of 75.82 was the highest on the entire list. Alibaba's Wan 3.0 ranked second with 66.63, and its basic capability score of 68.51 was slightly higher than that of MiniMax H3; ByteDance's Seedance 2.5 and Kuaishou's Kling V3 Pro tied for third with 64.63 and 64.06. The ranking was updated on October 23; overseas models are included only for reference and are not counted in the official ranking.
In the latest SuperCLUE I2V (image-to-video) model overall ranking, MiniMax (Xiyu Technology) H3 took first place in China with an overall score of 71.46, and its application capability score of 75.82 was the highest on the entire list. Alibaba's Wan 3.0 ranked second with 66.63, and its basic capability score of 68.51 was slightly higher than that of MiniMax H3; ByteDance's Seedance 2.5 and Kuaishou's Kling V3 Pro tied for third with 64.63 and 64.06. The ranking was updated on October 23; overs
Matched fieldsTitle
The guest lecture video by Nathan Lambert for Harvard CS 2881R "AI Security" has been released. He surveys the state of open-source models and argues that Chinese open-source models lead comprehensively across three dimensions—performance, downloads, and paper citations—while US open-source models lag China by another 3 to 4 months. He judges that distillation is not the key issue, that the total compute of the top five US labs averages 16 times that of the top five Chinese labs, and points out that banning open source is unfeasible in practice and would only suppress economic activity.
The guest lecture video by Nathan Lambert for Harvard CS 2881R "AI Security" has been released. He surveys the state of open-source models and argues that Chinese open-source models lead comprehensively across three dimensions—performance, downloads, and paper citations—while US open-source models lag China by another 3 to 4 months. He judges that distillation is not the key issue, that the total compute of the top five US labs averages 16 times that of the top five Chinese labs, and points out
Matched fieldsTitle
Vaibhav Srivastav reposted OpenAI's post introducing GPT-6 Astra, saying that the demo video has about 140 million views and that he still finds it hard to believe. The original OpenAI post said that anything users can do on a computer, Astra can do for them, and quickly.
Matched fieldsTitle
Sebastian Raschka has released 8 programming videos on building a reasoning model from scratch, totaling 11 hours, using Python and PyTorch to implement evaluation, inference-time scaling, reinforcement learning, and distillation starting from a pretrained Qwen3 0.6B model. The course first builds a math answer verifier and evaluates on MATH-500, then implements temperature sampling, top-p filtering, and self-consistency, followed by RLVR training with a simplified version of GRPO, and finally distills a small model using answers generated by DeepSeek R1. The content accompanies his book "Build a Reasoning Model (From Scratch)"; the early text generation and evaluation examples can run on a CPU, while later training experiments are recommended to use a GPU.
Sebastian Raschka has released 8 programming videos on building a reasoning model from scratch, totaling 11 hours, using Python and PyTorch to implement evaluation, inference-time scaling, reinforcement learning, and distillation starting from a pretrained Qwen3 0.6B model. The course first builds a math answer verifier and evaluates on MATH-500, then implements temperature sampling, top-p filtering, and self-consistency, followed by RLVR training with a simplified version of GRPO, and finally d
Matched fieldsTitle
AI video generators produce temporally coherent frames from text, images, video, or control signals, while striving to maintain motion and scene consistency. The pipeline can be broken down into five steps: encoding prompts and reference materials, initializing a noisy spatiotemporal representation, jointly predicting motion and appearance, denoising or transporting the latent, and decoding, upsampling, and checking temporal consistency. The article emphasizes that the key difference from independent frame-by-frame image generation lies in the causal and consistency constraints across frames.
AI video generators produce temporally coherent frames from text, images, video, or control signals, while striving to maintain motion and scene consistency. The pipeline can be broken down into five steps: encoding prompts and reference materials, initializing a noisy spatiotemporal representation, jointly predicting motion and appearance, denoising or transporting the latent, and decoding, upsampling, and checking temporal consistency. The article emphasizes that the key difference from indepe
Matched fieldsTitle
Tencent Hunyuan releases the open-source free-form interactive video world model Hy-Nomas, which supports interaction using text prompts and camera trajectories, and features long-horizon generation with world memory. The model can be adapted to specialized domains such as robotics via lightweight LoRA, and has been optimized for consumer-grade GPU inference. The official release also provides website, GitHub, Hugging Face, and demo entry points.
Matched fieldsTitle