zeroth

Nothing to say

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 I also want to point out that whether monopoly are bad or not is a debated topic in economics. With evidence going one way and evidence going the other way.

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 Monopoly? We are not really talking about AI then. Unless strict regulations are passed, it will be less and less monopolized.
Bigger companies may still possess the more advanced ones, but you will have plenty other... FFS, Felix Kjellberg have his own private AI

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 ... And in the worse case scenario, you could ask another AI agent to try to retrace the reasoning of the AI paper. Or to explain it in details.

The issue with the criticism is that the (hypothetical) issues they talk about are literally solved by using more AI.

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 W1: I've seen no evidence for the veracity of the claim. The papers written by AI doesn't look less detailed or more lacking than a human paper. No mathematician showed it was less (let's say by comparing a randomly picked human and AI papers)...

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 Why do you believe AI cannot do that? Maybe it already did, after all no one has read all of the papers / proofs OpenAI generated. The Navier-Stokes one alone was 56 pages.

But in he off chance it didn't, why couldn't it? Or why would AI be worse at it than humans?

[zeroth @ZerothDrachma]

@rystm @StevenGlinert I get it, you're just trolling

[zeroth @ZerothDrachma]

@BoundFil @rixisdavid Le gars semble bien content. C'est pour ça que je dit que tu t'imagines bien des choses.

"je suis pret à l'entendre" sauf quand la "minorité" ne se conforme pas à tes stéréotypes. Classique de la gauche.

[zeroth @ZerothDrachma]

@rystm @StevenGlinert If you understand how bicycles work then please publish you paper on a reputed physics journal, you'll be famous.
You are the sheep who believes to be a wolf, eyes closed without realizing it. The epitome of midwittery.

[zeroth @ZerothDrachma]

@skip_strike @MizusPuddle Riot games own the character though. It would be different if the artist created the it, but that's a fanart.

[zeroth @ZerothDrachma]

@BoundFil @rixisdavid Ca se met inventer des histoires, je vois...

[zeroth @ZerothDrachma]

@kkr8r You are the intolerant we ought not to tolerate, though.

[zeroth @ZerothDrachma]

@BoundFil @rixisdavid Si tu penses que la dame hais les Asiatiques ou même celui-ci, c'est que tu n'as clairement jamais vu que qu'est la haine.

[zeroth @ZerothDrachma]

@ShadowbanSlam @gen0m1cs @cremieuxrecueil Which are the replications? There are related papers, but I couldn't find replication.

Also, I don't see why this epigenetics wouldn't be counted in the shared environment (=~0%) in classical twin studies. Wouldn't the dizygotes share this "generational fear"?

[Jesús Fernández-Villaverde @JesusFerna7026]

¿Cómo sería la población de España en un mundo en el que la inmigración neta a España desde 1995 a 2026 hubiera sido cero en vez de la observada?

Para calcular este escenario, tomo la estructura de población de 1995 y asumo que la fecundidad y mortalidad por edad hubieran sido las mismas que hemos observado. Esta hipótesis no es necesariamente realista, pero sí me parece que nos da unos resultados sobre los que podemos empezar a pensar.

Cada lector puede, mentalmente, ajustar la fecundidad hacia arriba (por ejemplo, con menos inmigración el precio de la vivienda sería más bajo con casi total probabilidad) o hacia abajo (ciertos grupos de inmigrantes tienen fecundidad más alta que la media).

Particularmente, no creo que haga mucha diferencia en la práctica lo que uno asuma sobre la fecundidad: unos cientos de miles de personas en una dirección o la otra a lo sumo.

No hace falta ajustar la mortalidad: calculo tasas por edad y no creo que estas fueran diferentes en mi escenario.

Lo que más me sorprende es que la inmigración ha generado una subida de la población con respecto al escenario base de 11,1 millones (los inmigrantes y sus hijos nacidos en España), pero el porcentaje de población de 15 a 64 años apenas ha cambiado: tenemos un 66,3% en ese tramo frente al 64,2% en el escenario base.

Y ese cambio de 64,2% a 66,3% no solo no resuelve nada, sino que empeora las cosas. Dado que tenemos un Estado del bienestar altamente redistributivo y la mayoría de la inmigración ha sido de baja cualificación (que son personas con valor presente descontado negativo para el presupuesto público, sí, también en equilibrio general), casi seguro que hemos empeorado la sostenibilidad de las cuentas públicas en el largo plazo.

Hemos cambiado España de manera muy profunda y lo hemos hecho sin pensar mucho en las consecuencias.

[zeroth @ZerothDrachma]

@ShadowbanSlam @gen0m1cs @cremieuxrecueil Was that study independently replicated? It's been more than 10 years after all. I wouldn't say a study on one mammal on one trait to be significant evidence, it's a trail at best.

[zeroth @ZerothDrachma]

@rystm @StevenGlinert Let's take another trivial example: we don't understand how bicycle works, how we don't just fall when we ride them.

Terence Tao would suggest that we should not ride them, and advocate for a committee that could ban the production of bicycles

You still think Tao is reasonable?

[zeroth @ZerothDrachma]

@StevenGlinert @basedjensen Does AI stop you to try to understand it? No. Then Terence Tao's argument is bad.

[zeroth @ZerothDrachma]

@zeynep What is this mental gymnastic? Also this new hypothetical isn't much better. That's still a very uncompelling hypothetical.

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 And AI does that too. Typically on its own.

But even in the worst case scenario, you can call another agent, to review the first and fill in the auxiliary tools needed by the first agent.

[zeroth @ZerothDrachma]

@regc_17 @IamKyros69 This isn't a real issue with AI. If you have access to the AI model, you could have them solve the problem again and again from scratch... and have it listing all its failed thoughts.

OTOH, did Terence Tao ever shared all his failed thoughts? Can he remember them all?