Rafael G.

Applied AI Scientist | vision & language | 🇲🇽

[Rafael G. @kardashevs]

@hiarun02 It does not feel unlimited. But its great value for the price.

[Rafael G. @kardashevs]

@vale_ej @pablo_majluf ¿Qué dices? ¿Sabes que estamos hablando de México?😂

[Rafael G. @kardashevs]

@MrAhmadAwais And the Transformers :)

[Rafael G. @kardashevs]

@olveraf1 @luzelenachavez8 @Claudiashein @LuisaAlcalde Los modelos generativos existen desde los 90', se perfeccionaron para generación de texto e imagen hace 10 años. Los primeros "LLM" se publicaron en 2020, la IA ya existía en pandemia. Los "LLM generativo AI" que mencionas, existen desde 2020 (GPT-3). No hay verdad a medias.

[Rafael G. @kardashevs]

@Biggest You do have a pair! Thanks for standing out against this regulatory nonsense :)

[Rafael G. @kardashevs]

@Mononofu @JensenHuang What an embarrassing take. Microsoft and NVIDIA not open sourcing their core products is not the same as Anthropic pushing to ban open-source LLMs. Everyone should be free to open-source whatever they want.

[Brian Halligan @bhalligan]

I don't remember where I found this, but its spot on.

[Rafael G. @kardashevs]

@sudoingX this is the way

[Deep Thrill @DeeperThrill]

@sama DoW: “This is a bad guy.”

ChatGPT: “You’re absolutely right! You’re not just an astute observer of human nature—you’re a true patriot. Would you like me to initiate a drone strike 🥰💥?”

[Arewa AI @arewa_ai_]

Mientras las "Big Tech" acaparan titulares con anuncios como GPT-5.3-Codex, Opus 4.6 o Gemini 3 Deep Think, la esfera Open Source (OSS) ha lanzado modelos que no solo les pisan los talones, sino que los superan en eficiencia para casos de uso específicos de ingeniería.

[Hugging Models @HuggingModels]

You can now clone voices that sound actually real.
Non robotic, full human.

1.7B parameters of clean, natural speech.

https://t.co/nvK8GFi53o

[Arewa AI @arewa_ai_]

Rentar IA de las Big Tech es una trampa de oro. 💸🚫

Construir el futuro de tu empresa sobre APIs cerradas es como construir una casa en terreno rentado. Si el dueño cambia las reglas, tu negocio se tambalea.

🧵

[Avichal Garg - Electric ϟ Capital @avichal]

Ironic that we got free AI from a hedge fund and $200/month AI from a nonprofit.

[clem 🤗 @ClementDelangue]

Our science team has started working on fully reproducing and open-sourcing R1 including training data, training scripts,...

Full power of open source AI so that everyone all over the world can take advantage of AI progress! Will help debunk some myths I’m sure too.

Thanks @deepseek_ai!

[Pedro Domingos @pmddomingos]

Stargate would look a lot better if DeepSeek hadn't just matched o1 at a tiny fraction of the cost.

[1517 Fund @1517fund]

No one cares about your middle school grades.

But they do care about the robots & apps you made.

Remember kids, grades get stale in a year. Projects are fresh for life.

[Rafael G. @kardashevs]

@originMaxima @FanaHOVA @realGeorgeHotz https://t.co/HgFhFOM97O

[the tiny corp @__tinygrad__]

@jtatarchuk @FanaHOVA @realGeorgeHotz Hehe, we'd only accept them from AMD. It's about showing they are serious about investing in software development. Otherwise the platform is a dead end, and we shouldn't invest into it either.

And ~200k to them, 400k retail.

[Shayan @ImSh4yy]

After spending the weekend trying to glue LLM calls together (aka AI Agents), I regret to admit that LLMs are pretty stupid.

[Rafael G. @kardashevs]

@kepano Hell no!

[Jonathan Gorard @getjonwithit]

Put your work on arXiv and release 100% of your code/data/Mathematica notebooks/etc., packaged up in such a way that *anyone* can reproduce all of your results in a single click. Call out people who don’t. That’s what peer review should look like in the 21st century.