Paper : arxiv.org/abs/2602.11298 Model weights : huggingface.co/mistralai/Vox… Mistral AI studio live playground : console.mistral.ai/build/aud…
→ View original post on X — @guillaumelample, 2026-02-19 11:28 UTC
GENERATIVE AI
-
Voxtral Realtime: High-Performance Real-Time Transcription in Open Source
By
–
Voxtral Realtime paper is out! The model is released under the Apache 2 license, and achieves state-of-the-art transcription performance at sub-500ms latency. Mistral AI for Developers (@MistralDevs) Since launching Voxtral Realtime, the community response has been remarkable. Today, we share the technical report, launch the Realtime playground in Mistral Studio, and share the model in Hugging Face Transformers. 🧵 — https://nitter.net/MistralDevs/status/2024162999599005715#m [Translated from EN to English]
→ View original post on X — @guillaumelample, 2026-02-19 11:28 UTC
-
Framework for effective AI prompting
By
–
The difference between these prompts and random AI use: → You give it a role with real stakes
→ You force it to be honest, not polite
→ You ask for specific output, not general advice
→ You treat it like a senior expert, not a search bar That's the whole framework. -
How AI Is Rewiring Filmmaking Live Stream Today
By
–
Don't miss my latest #livestream at 12 noon today – "How AI Is Rewiring Filmmaking" Date: Today, 19th February Time: 12:00 GMT, 07:00 EST, 04:00 PST, 17:30 IST LinkedIn > https://
lnkd.in/gsq28F2
Twitter > https://
lnkd.in/ehrXiUW
Facebook > https://
lnkd.in/e_9hcfph
YouTube -
AI video enters cinema phase with Kling 3.0’s multi-shot narratives
By
–
My AI read 60,000 posts and analyzed them for discussion of Kling and found:
— Robert Scoble (@Scobleizer) 19 février 2026
1. AI video is entering the "cinema" phase — Moving from single clips to multi-shot narratives with audio. Kling 3.0 is the first to demonstrate this at scale.
2. China is competitive in generative… pic.twitter.com/Ot1OBM9Hz5My AI read 60,000 posts and analyzed them for discussion of Kling and found: 1. AI video is entering the "cinema" phase — Moving from single clips to multi-shot narratives with audio. Kling 3.0 is the first to demonstrate this at scale. 2. China is competitive in generative
-
Scobleizer’s Kling Report on AI Video Trends
By
–
The Kling Report. For those interested in AI video.
— Robert Scoble (@Scobleizer) 19 février 2026
I'm running a series of reports. Having AI read my lists here on X (I have about 50,000 people in AI industry on them) looking for interesting trends. I've been having a ton of fun playing with Kling, so ran a report on that… pic.twitter.com/F3hCYodtl4The Kling Report. For those interested in AI video. I'm running a series of reports. Having AI read my lists here on X (I have about 50,000 people in AI industry on them) looking for interesting trends. I've been having a ton of fun playing with Kling, so ran a report on that
-
Phoenix-4 Behavioral Model Advances Conversational AI Dynamics
By
–
What makes Phoenix-4 interesting is not real-time rendering at 40 FPS, but the behavioral model behind it.
— Antonio Grasso (@antgrasso) 19 février 2026
Full-duplex listening, explicit emotional control, and independently generated listening states bring conversational AI closer to real human dynamics.@tavus partner. https://t.co/Nd4PIPZFtMWhat makes Phoenix-4 interesting is not real-time rendering at 40 FPS, but the behavioral model behind it. Full-duplex listening, explicit emotional control, and independently generated listening states bring conversational AI closer to real human dynamics. @tavus partner.
-
Elon Musk: Ask Grok not yet on version 4.20, update soon
By
–
Ask Grok is not yet on the 4.20 version. Should happen within a day or two.
-
Identical agents with random expertises may converge via Grok
By
–
The agents are actually identical, the names are whimsically applied and their “expertises” are random at inference start. However, there could be a personality convergence and subsequent sticking due to Grok reading about them on 𝕏.
-
Lex interviews Steinacker: GPT 5.3 Codex vs Claude Opus comparison
By
–
这是前几天 Lex 访谈 @steipete 的片段，bro 真的端水大师。
— 艾略特 (@elliotchen100) 19 février 2026
从 1:38:52 开始讲 GPT 5.3 Codex vs Claude Opus 4.6。
这期是 Peter 去 OpenAI 官宣前的采访，但从语气看，他那时大概率已经决定去 OpenAI 了。
也正因为这样，他在对比 Codex 和 Opus… https://t.co/dNgIOTOhMI这是前几天 Lex 访谈 @steipete 的片段，bro 真的端水大师。 从 1:38:52 开始讲 GPT 5.3 Codex vs Claude Opus 4.6。
这期是 Peter 去 OpenAI 官宣前的采访，但从语气看，他那时大概率已经决定去 OpenAI 了。
也正因为这样，他在对比 Codex 和 Opus