Sebastián Gramajo

Building @Buholegaltech | legal AI to eliminate bureaucracy in Bolivia 🇧🇴
| Estudiante de Derecho y Autodidactica en sistemas e inteligencia artificial

[Sebastián Gramajo @anaiamake]

Adiós 2025.
-
HOLA 2026:
- Lanzamiento oficial de @BuhoLegaltech
- Videos diarios en REDES
- Gym peak
- 4to año de derecho
- agencia contable junto a mi madre
"Ejecución, mejora, estudio e iterar lo mas rápido que se pueda."
........
a mis 21 años.

[Adrian Diaz Marro 李安 @AdrianDiazMarro]

Así es... y aún así el esfuerzo y el sacrificio chino fue monetizado gracias a los occidentales que critica. Con lo que, indirectamente China se enriqueció gracias a venderle productos a los saqueadores, es decir, el dinero que obtuvo China es el mismo "pote de oro" que fue previamente saqueado. Y el sureste asiático crecerá gracias a los chinos que crecieron gracias a los occidentales que crecieron gracias a saquear al resto. Estamos todos conectados. Los argumentos historicistas no sirven para nada.

[teebee @flashytrans]

yeah bro the ai bubble is gonna burst any minute bro haha it'll all be back to normal any week now bro haha

[Aryan Sawhney @aryansawhney17]

Everytime, I think AI models are expensive, I look back at this screenshot and thank capitalism

[Ole Lehmann @itsolelehmann]

it must feel so insane being in university and seeing these models progress so fast while people teach you the concepts of a world that stops existing

[PredictableLiar @PredictableLiar]

Va a ser una masacre la que la que va a sufrir la economía que conocimos durante décadas.

Va a ser adaptarse o morir.

Mientras tanto, más de la mitad de la población acá se pregunta cuál es el stock de mercadolibre o globant...

[Sebastián Gramajo @anaiamake]

@darel023 @LottielabHQ BURP

[Sebastián Gramajo @anaiamake]

Wu Wei

[Sebastián Gramajo @anaiamake]

información, construcción y flujos singulares

[Riley Brown @rileybrown]

Hate to admit it… we’ve hit a wall… I’ve seen no noticeable improvements to the models in the last 6-9 hours.

[How To Prompt @HowToPrompt__]

Google has found something worst than AI hallucinations.

they call it “Metacognitive Failure.”

when an LLM hallucinates a fact, it's a data error. bad memory retrieval, wrong weights, fine. you can fact-check it.

metacognitive failure is a structural psychological defect in the model's architecture.

according to the research, current frontier models exhibit massive gaps in their internal self-monitoring:

the supreme confidence trap: they routinely hallucinate with maximum, bulletproof confidence. they sound just as sure of themselves when they are completely wrong as they do when they are 100% right.

the boundary blindness: they have zero internal mechanism to recognize their own knowledge boundaries. they blindly step off cliffs because they can't sense the edge.

the calibration mismatch: there is a complete disconnect between what the model actually "knows" internally and what it expresses in its output.

the fix: reinforcement learning with metacognitive feedback (rlmf)

the researchers operationalized a new technique to force models to face reality. instead of just rewarding the model for getting the right answer, rlmf grades the model on how accurately it evaluates its own performance and uncertainty.

by aligning a model's expressed confidence with its actual intrinsic uncertainty, they managed to massively improve model calibration without dropping raw accuracy.

as we shift into an agentic era where ai systems are running code, managing infrastructure, and making automated decisions without humans in the loop, a model that doesn't know what it doesn't know is dangerous.

[Sebastián Gramajo @anaiamake]

@cy_burns ESTUDIO

[Daniel Jeffries @Dan_Jeffries1]

Human written code was mostly drek unless you were a genius like John Carmack.

Most human writing is garbage unless you're a master like Hemingway or Amy Tan.

Human driving is a nightmare that kills 1.2M and injures 50M more a year.

AI is an improvement for 99% of the pop.

[🕊️ @lichthauch]

Most people who get good grades are just good at memorization, not actually smart, they just have a neat stack of flashcards for a brain and they can sort through them quickly. pure intelligence is about accessing the subconscious directly. its like a dark cavern underneath your thoughts and when you shine the right light down there, surprising shit comes out. connections that you wouldnt make on your own, insights that feel like they came from somewhere else, because they did. the mind is a lot bigger and deeper than most people use, they just skate on the top layer. but there are fucking cathedrals down there man, temples buried in the mud, and all kinds of wild animals that you can catch and harness if youre not scared to climb down the ladder into the dark

[Sebastián Gramajo @anaiamake]

@Mooncore777 @teortaxesTex Exactamente, la gente en occidente está muy sesgada por las emociones, buscan respuesta cortas, rápidas de digerir y mejor si van alineadas con sus sesgos previos, buscan respuesta plausibles, no interés que tengan o no sentido o sean o no profundas, con que se sientan así cumple

[David Shapiro (L/0) @DaveShapi]

This is why economists were dead wrong about AI.

Since they aren't technologists, they didn't accurately price generative AI. They looked at it and saw a chatbot.

They had no idea how this technology can scale.

There are still economists who don't believe that generative AI is a General Purpose Technology.

This is one of the biggest reasons that the future is mispriced by the establishment.

[Adrian Diaz Marro 李安 @AdrianDiazMarro]

El hombre puede creer en lo imposible, pero jamás creerá en lo improbable.

[Sebastián Gramajo @anaiamake]

@peperia1354 @calotonterias muchas de esas personas delgadas de tribus que mencionas en realidad tienen una buena masa muscular, ya que tienden a hacer mucho más ejecicio físico y a tener dietas aunque no siempre muy calóricas, ricas en proteínas y carbos, los sedentarios jamás crean músculos y son grasosos

[Chris Lakin @chrislakin]

what your risk-aversion costs you

[Sebastián Gramajo @anaiamake]

@Duckman_72 @IN_Vic_TusX @mattscreates Recomiendas esa estrategia para conseguir una cuenta de douyin o Weibo? O conoces alguna otra estrategia siendo extranjero fuera de China?