Daniel Morales

[marcus @marcusyul]

🚨ESTO ACABA DE MATAR A ELEVENLABS

74.000 personas ya lo han guardado en GitHub y casi nadie habla de ello.

Se llama VoiceStudio y es un reemplazo 100% gratis que corre en tu propio ordenador.

Te permite:

→ Clonar tu propia voz
→ Doblar vídeos a 646 idiomas
→ Crear audiolibros y dictar
→ Elegir entre 17 motores

Sin pago por carácter.
Sin límites de uso.

Todo en tu máquina.

🔖 Guárdalo.

[Yapayzekaserisi @YapayZekaSerisi]

The Lord of the Rings: MC Gandalf x DJ Frodo...

Made with @higgsfield

[New Mentalities @NewMentalities]

You've been marinating in potential long enough. Time to cook.

[Daniel Morales @danieldaniel]

@testingcatalog They finally noticed that the idea of talking to multiple bots to streamline a singular goal is not efficient, nice.

[Daniel Morales @danieldaniel]

@0xInk_ MMORPG with those graphics in ThreeJS and you’re billionaire.

[Aquib @heystykid]

POV: you switched to a low reasoning effort model

[Daniel Morales @danieldaniel]

@JC_builds Thanks JC! Really appreciated, this contribution will be very useful to everyone’s projects.

[雨夹雪❄️ @mizorewww]

介绍比Jev快50倍，在你设备上跑的laya-mlx!

只在你的设备上占用最高1G内存

Laya是一个开源的类似于Jev的,基于文本输出概率的分类系统

我将其移植到MLX,并且做了一些性能优化!

视频中就是这个模型在我的本地M3Max上玩贪吃蛇

这个模型能够以每秒决策60次的速度玩贪吃蛇！

https://t.co/31KGUiNunb

[Daniel Morales @danieldaniel]

You can now use Siri for any simple-medium question instead of Grok or Claude Anonymous. To replicate the “Memory” file, create an Apple Notes and write everything.

[Daniel Morales @danieldaniel]

Fusion energy and AGI go hand in hand.

[Jay Dwivedi @jaydwivedi_]

Here’s how I design apps (Step by Step) ⬇️

- Idea
- User Flow
- Wireframe
- First Draft
- Iterations
- Final Design

[Daniel Morales @danieldaniel]

In less than two years, we’re going to be able to customise LLMs in a modular way to reduce the amount of unnecessary parameters, data, languages, and knowledge that will never be used. That will be part of fitting intelligence into small devices to make it accessible.

[José Siles | AI | Data @josesilesdata]

🚨SE ACABÓ NO PODER EJECUTAR MODELOS PESADOS DE IA EN TU MÁQUINA!

kimi-k3-in-c es un motor puro en C99 de apenas 176 KB que transmite los expertos de Kimi K3 directo desde el disco

Sin cargarlos nunca en memoria!

- Sin GPU
- Sin CUDA
- Sin frameworks

Ejecuta un modelo de 2,78 BILLONES de parámetros en una CPU con solo 8,24 GB de RAM 🤯

100 % código abierto.

El futuro de la IA acaba de llegar... y cabe en tu portátil.

Te dejo el link en los comentarios 👇🏼

[Nicolas Krassas @Dinosn]

I tried a Local AI model (Qwen 3.6 27b) for security research and it works surprisingly well. https://t.co/09HOfCr6vH