Asterisk's new chan_websocket module simplifies media exchange

📕 Building a Realtime Chat Application with Nuxt and Socket.IO 👉🏻 Socket.IO is a powerful tool for building applications that demand instant communication. https://lnkd.in/eXfHbKpT

For any of those interested in open source video generation. Best read up. Wan 2.5 preview is behind closed doors for now, and perhaps will not be released, but my sources indicate that Wan 2.5 will eventually be released and open source. Given the communities ability to quantize models, once Wan 2.5 is released, it will quickly get quantized and be able to run on fairly modest consumer GPUs. https://lnkd.in/g3g_37Kk

Ever wondered how large-scale systems slash GC pauses from seconds to milliseconds—and what lessons you can deploy in your own stack? Twitch’s journey with Go’s garbage collector is an eye-opener. Recommend as must read for all #performanceengineering guys. Key points: 1️⃣ GC Pause Reduction Multi-seconds → <1ms (100x faster) Achieved via GC tuning & evolution (Go 1.2 → Go 1.7) 2️⃣ Concurrent & Incremental GC Go 1.5 introduced mostly-concurrent GC 10x latency reduction → chat survived massive loads smoothly 3️⃣ Root Cause Analysis Perf profiling uncovered GC bottlenecks: Stack shrinking Finalizer scanning Fixes shipped in Go 1.6 & Go 1.7 4️⃣ NUMA & Memory Locality NUMA page faults = hidden latency culprit Pinning/tuning cut pauses from 70ms → 10–15ms Outcome: GC improvements for the whole Go ecosystem

“the #web is lacking in compassion, but this is ... a design issue.” The algorithms that make social media so compelling have harmful consequences and should be regulated, argues computer scientist Tim Berners-Lee, who invented the World Wide Web, https://lnkd.in/d4NV4_Zv

VoXtream is now open-sourced! VoXtream is a full-stream zero-shot TTS model for real-time use that begins speaking from the first word. 𝗞𝗲𝘆 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀: - 𝗦𝘁𝗿𝗲𝗮𝗺𝗶𝗻𝗴: Support a full-stream scenario, where the full sentence is not known in advance. The model takes the text stream coming word-by-word as input and outputs an audio stream in 80ms chunks. - 𝗦𝗽𝗲𝗲𝗱: Works 𝟱𝘅 times faster than real-time and achieves 𝟭𝟬𝟮 𝗺𝘀 first packet latency on GPU. - 𝗤𝘂𝗮𝗹𝗶𝘁𝘆 𝗮𝗻𝗱 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆: With only 9k hours of training data, it matches or surpasses the quality and intelligibility of larger models or models trained on large datasets. The work was done under the guidance and supervision of Gustav Eje Henter and Gabriel Skantze. 🔗 Paper: https://lnkd.in/d5u_hfcr 🔗 Demo: https://lnkd.in/d7zj2mgc 🔗 Code: https://lnkd.in/dYv5ceib 🔗 Model: https://lnkd.in/dmkq--8K #tts #texttospeech #streaming

New Blog: Based on his talk at Black Hat, Steve Salinas discusses integrating Windows AI/ML APIs into Cobalt Strike’s workflows and presents proof-of-concept implementations for AI-augmented post-exploitation capabilities in Cobalt Strike. https://ow.ly/qsKO50WWTW3

How to lie about latency. There are lots of claims of ever lower latencies online, especially from upcoming consumer technologies like WebRTC and Media-over-QUIC. Find out in our blog post about how their cherry-picking of latency measurements means they are not comparable to professional video transport products. https://lnkd.in/gmwkWT9q

Your team shouldn’t be firefighting message duplication, ordering bugs, or global routing logic just to build chat. Group chat at scale means more than connecting sockets. It’s about uptime, latency, and keeping conversation flow consistent - everywhere. Ably Chat handles message delivery, presence, and realtime behaviors like reactions or typing indicators, all with global-scale guarantees. Built on a platform with 5+ years of 100% uptime and billions of monthly messages, it helps product teams focus on features, not edge cases. Don’t write infra your users will never see - but will definitely feel when it breaks. Find out more - link in comments below 👇

What if the future of the internet runs through a single toll booth? In my conversation with Stewart III on Stewart Squared, we explored Cloudflare’s role as the “network administrator” of the web, why 80% of AI queries already touch their system, and what it means if AI crawlers have to pay, like an app store for information. The takeaway? The business model of the internet is cracking, and what replaces it could reshape how knowledge, money, and power flow online. Watch the whole episode on Stewart Squared