Arte Suave @fuzflow

Build with AI. Fight in gi. Fly worldwide. Rio de Janeiro, Brazil Joined March 2010

Tweets

1K
Followers

289
Following

848
Likes

19K

David Ondrej @DavidOndrej1

2 weeks ago

99% of slop *does not* come from AI it comes from Humans, who have bad taste current AI models can build anything but it's up to YOU to know what to build if you suck at Product, no agent will save you

35 19 303 11K 26

View Details

Arte Suave @fuzflow

a month ago

@MarioNawfal Yes. It is. Thank you for the attention on this matter.

0 0 0 309 0

View Details

Arte Suave @fuzflow

a month ago

@georgesttock Claude

0 0 0 12 0

View Details

Spencer Shulem @Spshulem

a month ago

Me and Fable 5 for the 3 days I had her

214 2K 24K 1.4M 3K

View Details

Arte Suave @fuzflow

a month ago

@mrbrandowner Playbook

0 0 0 6 0

View Details

Arte Suave @fuzflow

a month ago

Rio's out here "innovating" by straight-up stealing AI Models now. Robbery 2.0:same favela energy, just without the gun in your face! 🇧🇷

Nex @NexEcosystem

a month ago

The Rio 3.5 model broke the internet this week. The plot twist? It’s essentially our open-source model, Nex N2 Pro, wearing a different hat. 🤯 We analyzed the weights, and the recipe is exact: Rio 3.5 ≈ 0.6 * Nex N2 Pro + 0.4 * Qwen 3.5 It even literally introduces itself

221 536 5K 907K 1K

0 0 0 22 0

View Details

Arte Suave @fuzflow

2 months ago

Nike did it again. 💯

0 0 0 8 0

View Details

Arte Suave @fuzflow

2 months ago

@tye1337 Start

0 0 0 6 0

View Details

Reads with Ravi @readswithravi

2 months ago

Marcus Aurelius wrote this over 1800 years ago: “Do not disturb yourself by imagining your whole life at once.”

157 10K 61K 1.3M 8K

View Details

José Vitor ⬛️🟨⬜️ @JosVitorCaires1

2 months ago

O melhor edit que eu já vi. Dava pra ser propaganda oficial

118 998 5K 102K 884

View Details

Dmitrii Kovanikov @ChShersh

2 months ago

For my non-tech friends, this is like Ronaldo joining Manchester City

Andrej Karpathy @karpathy

2 months ago

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

8K 11K 150K 27.8M 14K

254 1K 19K 757K 853

View Details

Arte Suave @fuzflow

2 months ago

@karpathy Somehow reading this made me sad. Hope it’s for the best though. You are a huge inspiration.

0 0 0 3 0

View Details

Dr. Clown, PhD @DrClownPhD

2 months ago

Rocky IV, but instead of Sylvester Stallone, it’s Elliot Page.

2K 8K 102K 8.0M 7K

View Details

Dr. Clown, PhD @DrClownPhD

2 months ago

Troy, but instead of Brad Pitt, it’s Elliot Page.

6K 21K 215K 17.2M 19K

View Details

Elon Musk @elonmusk

3 months ago

Perhaps a restoration of dignity is in order

17K 24K 305K 78.2M 8K

View Details

GREG ISENBERG @gregisenberg

3 months ago

What business models/ideas work now because AI agents can actually do stuff? a thread of a few i think are REALLY interesting:

88 83 801 95K 2K

View Details

Ofir Press @OfirPress

3 months ago

1) Our team at Meta has a tough new coding benchmark challenging models to code entire programs including ffmpeg and the PHP compiler from scratch. 2) Top accuracy is 0% 3) We will be making the benchmark harder.

John Yang @jyangballin

3 months ago

How much of SQLite, FFmpeg, PHP compiler can LMs code from scratch? Given just an executable and no starter code or internet access. Introducing ProgramBench: 200 rigorous, whole-repo generation tasks where models design, build, and ship a working program end to end. 🧵

107 246 2K 741K 663

36 62 1K 137K 252

View Details

Alexander Whedon @alex_whedon

3 months ago

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.