Henrique Massareli

[Henrique Massareli @onlyeva46j0e]

@imjustnewatai Consider that the average player should have a offboard GPU so probably it should be better than the mediapipe model.

The arena should reflect a real boxing arena.

There should be VFX for punches and knockout should be possible, when being hit in the face, it should move.

[Henrique Massareli @onlyeva46j0e]

@imjustnewatai Create the best boxing webcam game experience you could do, it should be in first person view, online multiplayer, browser based have realistic movements, especially the z axis (that is very tricky on some mocap models)
The character should be drawn in detailed cartoon like dota

[Henrique Massareli @onlyeva46j0e]

Is it me or get teeth are just very buggy?
It doesn't look real

[Tavus @tavus]

Introducing Griffin, the first model to pass the video Turing test.

48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.

It’s the first Human Interaction Model (HIM).

[Henrique Massareli @onlyeva46j0e]

@charles_maddock This changes a decision algorithm and replace to a Jev workflow

[Henrique Massareli @onlyeva46j0e]

@i549 @cline Why? Any ai nowadays would be capable of doing that

[Henrique Massareli @onlyeva46j0e]

@Badtheorylabs how is it not subagents?

[Henrique Massareli @onlyeva46j0e]

@iamspritz mano oq me impressionou é a voz, como fez pra criar essa entonação etc

também seria top se dissesse mais ou menos quais tool calls a IA faz pra criar essas músicas

[Henrique Massareli @onlyeva46j0e]

@OnlyTerp Day trading is a trap

[Hoosha @hooshaaii]

Attention is inherently noisy! 🧠

The "Differential Transformer" subtracts two attention maps to cancel out noise and focus purely on the signal.

A massive leap for long-context reasoning without hallucinations! 👇

https://t.co/lqnxKyEs2i

[Henrique Massareli @onlyeva46j0e]

@oC3B_ Ah mano vai cagar

[Henrique Massareli @onlyeva46j0e]

@oC3B_ é baseado em llm mesmo? funciona com contexto normal de texto como jev ou é mais limitado?

[Astropulse @RealAstropulse]

@RisingSunInt_ Actually pixel art is one of the hardest art forms for AI to get right- I've been working on it for over 4 years and we still aren't there completely.

Getting closer every day, but an experienced pixel artist can tell you a million things wrong with this horse animation.

[Henrique Massareli @onlyeva46j0e]

@VictorTaelin Ué mas vc tá rodando ia local? Pq conseguiu lotar tanto?

[Henrique Massareli @onlyeva46j0e]

@DonFutboI Cara fica muito claro como as linhas eram mais abertas na época haha, hoje em dia se vc receber a bola sem saber pra onde vai acabou...
É muiito mais compactado o bloco

[Supersonic Labs @supersonicai]

lol, and we mogged Jev on a i5 ideapad with 16gb of RAM:

[Henrique Massareli @onlyeva46j0e]

@fastrlife If you increase output.. at the end of the day you will still need caloric deficit to loose weight...

Also, caloric deficit increases lifespan according to several studies..

[Henrique Massareli @onlyeva46j0e]

So you mean that....
Good companies actually WANT to pay more for their employees?

[Alex Hormozi @AlexHormozi]

This was my most viewed short of the year.

[Henrique Massareli @onlyeva46j0e]

Unfortunately, grok 4.7 seems to have floped.
And mimo achieved the same being 1000x cheaper

[Artificial Analysis @ArtificialAnlys]

MiMo-V2.6-Pro debuts as the top open weights model on the Artificial Analysis Intelligence Index (46). At $0.13 per Intelligence Index task, it lands on the Intelligence vs. Cost per Task Pareto frontier

@Xiaomi has just released MiMo-V2.6-Pro, an open weights model with major advances in intelligence over its predecessor, MiMo-V2.5-Pro (Intelligence Index: 26). Despite the improvement, it retains the same attractive pricing at $0.435 per 1M input tokens (with a 99% cache-hit discount) and $0.87 per 1M output tokens. This makes MiMo-V2.6-Pro one of the most cost-efficient models to deploy.

MiMo-V2.6-Pro is an MoE model with 1.02T total parameters and 42B active parameters. Stay tuned for additional analysis of the model.

Check out MiMo-V2.6-Pro full benchmarking breakdown here: https://t.co/czBJhQKuWJ

[Henrique Massareli @onlyeva46j0e]

@muvverfcker @El_Aguilar_Azul @doctormorphh Hahaha