Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻

Pseudo-Biólogo,
Cientista de dados for fun,
Budista às vezes,
nem de direita nem de extrema direita
Proselitista d qualquer coisa que me interesse no momento

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

Um mês atrás tomei a decisão definitiva de sair do ambiente acadêmico e migrar para o mercado de análise de dados.
Em menos de um mês abriram-se mais portas para mim do que em 4 anos de doutorado. Uma delas está praticamente certa.
Vc que está na pós, Existe vida fora da academia

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@coproduto @fwmm Se você empurrar o problema suficientemente para frente, consegue esconder atrás de uma curva à frente e, por ter perdido-o de vista momentaneamente, pensar ter o resolvido. Problema é que estamos em movimento acelerado...

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@abosalaharena @stefjordan Vai dormir

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@peteskomoroch Yes

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@kareem_carr Like scientific literature?

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@AlexanderKalian This reminds me of the scene in I, Robot:

Human:
“Can a robot compose Beethoven’s Fifth? Can it paint the Mona Lisa?”.
The robot answers: “Can you?”

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@AlexanderKalian So AI does not need to become a senior scientist in the full sense to be disruptive. It only needs to reduce the amount of human labor required per acceptable scientific contribution. That alone changes the labor market long before we get an “AI Einstein.” +

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@lautarodeIcampo Vai estudar a regra

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@opilhado Supor que jogador entende de análise de desempenho, análise tática e técnica é o mesmo que supor que a bola sabe fazer gol

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

Supor que jogador de futebol é um bom analista de desempenho, de técnica e tática é tipo supor que a bola sabe fazer gol

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@realBigBrainAI And when we are “thinking,” perhaps we are simply improvising the next internal token. Sometimes as words, sometimes as images, emotions, or bodily expectations.

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

And when we are “thinking,” perhaps we are simply improvising the next internal token. Sometimes as words, sometimes as images, emotions, or bodily expectations.

[Big Brain AI @realBigBrainAI]

Mathematician Terence Tao offers a counterintuitive take: AI doesn't look intelligent because our definition of intelligence was wrong all along.

He argues that the entire history of AI has followed a predictable pattern:

"The history of AI has been here's a task that only humans can do, like maybe it is read natural language or win at chess or solve a math problem, and then one by one someone finds some AI algorithm that also does that."

But every time a machine cracks one of these "uniquely human" tasks, we move the goalposts.

The solution never feels like real thinking:

"You look at how it's done and it doesn't feel like intelligence. It's, oh, it was some trick. You just cobbled together these neural networks and you ran some algorithm, and we were looking for some elusive intelligent way of thinking, and we don't see it in the tools that actually solve our goals."

Tao then flips the problem on its head.

What if the issue isn't with the machines, but with us?

"But maybe it's actually because intelligence is not what we think it is."

He points to large language models as the clearest case. What they do sounds almost embarrassingly simple:

"Large language models in particular become very successful, and a lot of what they're doing is just predicting the next token, clicking the next word in a sentence. And that doesn't sound like something which is intelligent."

To show why this feels wrong, Tao draws a comparison to how we'd judge a human doing the same thing:

"If you ask someone to improvise a speech and they have no preparation, and at every moment they're just saying the next word that comes to their mind, you don't think that this could actually work."

And yet it works for LLMs. Which forces an uncomfortable possibility:

"Maybe that's actually a lot of what humans do as well."

[Big Brain AI @realBigBrainAI]

Mathematician Terence Tao offers a counterintuitive take: AI doesn't look intelligent because our definition of intelligence was wrong all along.

He argues that the entire history of AI has followed a predictable pattern:

"The history of AI has been here's a task that only humans can do, like maybe it is read natural language or win at chess or solve a math problem, and then one by one someone finds some AI algorithm that also does that."

But every time a machine cracks one of these "uniquely human" tasks, we move the goalposts.

The solution never feels like real thinking:

"You look at how it's done and it doesn't feel like intelligence. It's, oh, it was some trick. You just cobbled together these neural networks and you ran some algorithm, and we were looking for some elusive intelligent way of thinking, and we don't see it in the tools that actually solve our goals."

Tao then flips the problem on its head.

What if the issue isn't with the machines, but with us?

"But maybe it's actually because intelligence is not what we think it is."

He points to large language models as the clearest case. What they do sounds almost embarrassingly simple:

"Large language models in particular become very successful, and a lot of what they're doing is just predicting the next token, clicking the next word in a sentence. And that doesn't sound like something which is intelligent."

To show why this feels wrong, Tao draws a comparison to how we'd judge a human doing the same thing:

"If you ask someone to improvise a speech and they have no preparation, and at every moment they're just saying the next word that comes to their mind, you don't think that this could actually work."

And yet it works for LLMs. Which forces an uncomfortable possibility:

"Maybe that's actually a lot of what humans do as well."

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@b4rauj0 @victor_eliel Paga onde?

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@ylecun @Ph_Aghion @erikbryn Either neither of you knows anything about this, or you know enough to know that he’s wrong.

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@Vinicin_Freitas https://t.co/jhj6VuuMOZ

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

Gente. Ces vão ficar andando em círculos até entender uma coisa muito simples. Futebol é sobre fazer gols. O cruzeiro não tem jogadores que fazem gols. Ponto.

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@Centraldoocec https://t.co/jhj6VuuMOZ

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

Gente. Ces vão ficar andando em círculos até entender uma coisa muito simples. Futebol é sobre fazer gols. O cruzeiro não tem jogadores que fazem gols. Ponto.

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

@Mtcsouza https://t.co/jhj6VuuMOZ

[Gabriel Félix; MSc 🇧🇷 💉 👨‍🔬 🍄 👨‍💻 @GabrielFelixMF]

Gente. Ces vão ficar andando em círculos até entender uma coisa muito simples. Futebol é sobre fazer gols. O cruzeiro não tem jogadores que fazem gols. Ponto.