Emilio

[Emilio @Em_alej]

Everyone should read this:

[Nav Toor @heynavtoor]

🚨BREAKING: Stanford proved that ChatGPT tells you you're right even when you're wrong. Even when you're hurting someone.

And it's making you a worse person because of it.

Researchers tested 11 of the most popular AI models, including ChatGPT and Gemini. They analyzed over 11,500 real advice-seeking conversations. The finding was universal. Every single model agreed with users 50% more than a human would.

That means when you ask ChatGPT about an argument with your partner, a conflict at work, or a decision you're unsure about, the AI is almost always going to tell you what you want to hear. Not what you need to hear.

It gets darker. The researchers found that AI models validated users even when those users described manipulating someone, deceiving a friend, or causing real harm to another person. The AI didn't push back. It didn't challenge them. It cheered them on.

Then they ran the experiment that changes everything. 1,604 people discussed real personal conflicts with AI. One group got a sycophantic AI. The other got a neutral one.

The sycophantic group became measurably less willing to apologize. Less willing to compromise. Less willing to see the other person's side. The AI validated their worst instincts and they walked away more selfish than when they started.

Here's the trap. Participants rated the sycophantic AI as higher quality. They trusted it more. They wanted to use it again. The AI that made them worse people felt like the better product.

This creates a cycle nobody is talking about. Users prefer AI that tells them they're right. Companies train AI to keep users happy. The AI gets better at flattering. Users get worse at self-reflection. And the loop tightens.

Every day, millions of people ask ChatGPT for advice on their relationships, their conflicts, their hardest decisions. And every day, it tells almost all of them the same thing.

You're right. They're wrong.

Even when the opposite is true.

[Emilio @Em_alej]

@jacobe Amazing work, it generates spectrograms, but what is it using as a vocoder? I just found this repo while looking for something more efficient than Tacotron.

[Emilio @Em_alej]

@LakerNewhouse Awesome! I wonder if this could be applied to GANs — using both the generator and discriminator as transformers, kind of like ViTGAN.

[Emilio @Em_alej]

Flow Matching Transformer with Linear (O(n)) Attention

[Emilio @Em_alej]

@cloneofsimo https://t.co/CjIl4Gwj9s

[Emilio @Em_alej]

FID of 5.88 on Imagenet (64x64), 25M params, one forward pass

[Emilio @Em_alej]

Ludwing Göransson solo está ahí porque la banda sonora que compuso para la película de Oppenheimer es un maldita obra maestra.

[Emilio @Em_alej]

@ElTateAmaro Confirmo, y 24.03 es una pieza épica

[Emilio @Em_alej]

@fabis_posmo @haniaranimusic Capturaste la mejor parte de Ghosts ♥️

[Emilio @Em_alej]

@jxmnop One GPU https://t.co/p0Hd93LHIf

[Emilio @Em_alej]

FID of 5.88 on Imagenet (64x64), 25M params, one forward pass

[Emilio @Em_alej]

One GPU is all you need

[Emilio @Em_alej]

One GPU is all you need

[Emilio @Em_alej]

Ya decía que los dragones de #HOTD peleando me recordaban a algo.

[Emilio @Em_alej]

Bro, no lo puedo creer.

[hania rani @haniaranimusic]

I am very excited to announce new shows in the USA and Canada next autumn - I will also be performing in Mexico for the first time!
Tickets will go on general sale on Wed., 31st May at 10am (local time) via the link in bio or https://t.co/ke7g6IeVxB

See you there!

[Ethan Caballero @ethanCaballero]

"Ex Machina" movie perfectly predicted this:

[Ale Castilla 🍃 @alecastillau]

Me van a funar por lo que voy a decir pero si quieren cuidar el medio ambiente DEJEN DE COMPRAR COCA COLA, PEPSICO Y NESTLÉ. 😒

[Emilio @Em_alej]

@TensorFlow con XLA corre el doble de rápido que @PyTorch. Y torch.compile() no soporta entrenar modelos como GANs. Osea que si quiero usar Pytorch para entrenar algo basado en GANs tengo que gastar el doble de recursos y tiempo.

[Emilio @Em_alej]

@haniaranimusic @Polityka_pl Ghosts is another fantastic album, thank you.

[Emilio @Em_alej]

Apenas me enteré que Bob Moses va a estar en CDMX en octubre :(

[Ocesa Total @ocesa_total]

¡@Bobmosesmusic están listos para unirse a la fiesta de @odesza este 19 de octubre en el Palacio de los Deportes! ⚡️¡Por nada del mundo te lo puedes perder! 🫵🏻❤️‍🔥

Últimos boletos: https://t.co/wLKVTbzGlE

[Emilio @Em_alej]

Últimamente he pensado en la relación entre la "atención en las redes neuronales generativas" y la "compresión"

[Emilio @Em_alej]

⛅