January 2024 Summaries

The text discusses how Mux was able to offer auto-generated captions on all videos uploaded for free by leveraging AI models like Whisper and OpenCLIP, which were previously considered expensive. They discovered GGML, an open source tensor library that optimizes CPU utilization, making it cost-effective for deployment. By using smaller model sizes, they achieved acceptable results while keeping costs low. Running inference on CPUs was found to be cheaper than using GPUs or third-party APIs. Mux's existing infrastructure for transcode jobs allowed them to efficiently schedule and run Whisper jobs alongside video transcodes, resulting in free auto captions for all their customers.

In this post, the author discusses their excitement for various developments at Mux in 2024. They highlight shipping more features faster than ever before and being closer to fulfilling their original vision for the product. The author is particularly excited about pricing-related features that allow developers to create high-quality video infrastructure without compromising on quality. Additionally, they mention client-side tooling improvements, including open-source contributions, as well as advanced video features like DRM and SRT. Longer data retention options are also mentioned as a valuable addition for customers. The author concludes by encouraging developers to keep an eye on Mux's upcoming developments through their new Beta page.

Spotify is revolutionizing the podcast industry by building a best-in-class platform and transforming how creators create and monetize their work. Since launching video podcasts (vodcasts) in 2020, they have seen a rise in engagement and now serve more than 100,000 video podcasts globally. Video offers an additional opportunity for creators to connect deeper with their audience, as visual information is processed faster in the brain than text. Spotify has partnered with Mux, a video infrastructure and QoE analytics tool, to improve viewer Quality of Experience (QoE) by making data useful for decision-making and learning from incidents. The integration of Mux Data has allowed Spotify to resolve incidents faster and gain insights into user behavior, content formats, and delivery improvements.

We've added support for 21 new languages to our VOD auto-generated captions feature, allowing us to increase accessibility for content created all over the world. We've also introduced a new API that enables retroactive caption addition, which can be used up to 7 days after content creation. Additionally, the VOD auto-generated captions feature is now Generally Available, marking it as stable and ready for scale.

Mux Video has introduced Automatic Cold Storage, a cost management feature that programmatically transitions assets to different storage levels based on their view count. The new feature offers varying storage discounts depending on the asset's encoding tier and resolution. While assets in Cold storage may experience reduced Quality of Experience (QoE), Mux Data allows users to monitor this impact. Automatic Cold Storage is currently available in beta for pay-as-you-go customers, with a 30-day and 90-day window for transitioning assets to different tiers. The company seeks feedback on the feature's performance and potential improvements.

The text discusses how various social media platforms have shaped consumer expectations regarding video creation experiences. It introduces Mux, a platform that simplifies video uploads and playback, and Shotstack, a cloud-based video editing platform used by developers and marketers to enhance, personalize, and automate video production. The text then provides an example of how these two platforms can be combined to create a complete video workflow for a fictitious mobile fitness app called Stride. It outlines the application components, including the mobile app, server, Mux, and Shotstack, and describes the process of uploading footage to Mux from the app, generating the video using Shotstack, hosting the video with Mux, and notifying the user. The text also highlights the power of Shotstack's template-based approach to video editing in creating unique videos for every user.