AGI

[Cadavre Exquis]

¿Qué podría salir mal si se utilizan LLMs como ChatGPT para crear guías sobre micología...?

'Life Or Death:' AI-Generated Mushroom Foraging Books Are All Over Amazon
Posted by msmash on Tuesday August 29, 2023 @12:00PM from the getting-out-of-hand dept.

samleecole writes:

A genre of AI-generated books on Amazon is scaring foragers and mycologists: cookbooks and identification guides for mushrooms aimed at beginners.

Amazon has an AI-generated books problem that's been documented by journalists for months. Many of these books are obviously gibberish designed to make money. But experts say that AI-generated foraging books, specifically, could actually kill people if they eat the wrong mushroom because a guidebook written by an AI prompt said it was safe.

The New York Mycological Society (NYMS) warned on social media that the proliferation of AI-generated foraging books could "mean life or death."

A quick scan of Amazon's mushroom and foraging books revealed a bunch of books likely written by ChatGPT, but are sold without any indication that they're AI-generated and are marketed as having been written by a human when they're probably not. 404 Media used GPT text detectors and AI image detection tools on some of the suspicious books, and found that they were very likely made with AI, with authors who may not even exist.

Saludos.

P.D.

Y sí, igual esta noticia debería ir en el hilo de Seguridad alimentaria o en el de A brave new world: La sociedad por venir  , pero me da que esto es solo el principio de lo que vamos a ver a raíz de la popularización de este tipo de tecnologías.

[Cadavre Exquis]

https://www.lavanguardia.com/tecnologia/20230830/9193334/google-meet-ofrece-ia-sustituya-usuarios-reuniones-virtuales.html

Saludos.

[Saturio]

https://www.lavanguardia.com/tecnologia/20230830/9193334/google-meet-ofrece-ia-sustituya-usuarios-reuniones-virtuales.html

Saludos.

Esto ya lo anticipé yo. Soy un genio.

Pero era un poco diferente. Mi anticipación era que nuestros smartphones conversarían por nosotros con los smartphones de otra gente. Saben quién es quién de cada uno de nuestros contactos y si son compañeros de trabajo o familiares o lo que sea. Se puede llegar a trazar los gustos y aficiones de cada una de las personas de la lista de tus contactos, incluso dónde están en cada momento y donde han estado en el pasado. Una IA incluso podría intentar predecir dónde van a estar en el futuro. De hecho se puede inferir si tienes una amante o si vas de putis. El smartphone de tu mujer puede montarle un pollo al tuyo y mientras tanto vosotros dos tan tranquilos cortando las cebollas en juliana en la cocina.

En realidad nosotros nos podríamos retirar de la vida. La IA podría postear por nosotros fotos de instagram y que otras inteligencias artificiales nos comentasen o nos diesen likes. Incluso nuestro bot se podría deprimir ante la constante exposición de positividad de los otros bots.

La idea de Google Meets me parece simplemente genial y todas las empresas deberían implementarla ya. Imaginemos una reunión de 6 personas en las que 4 son bots conversacionales de esos que cuando dices una burrada no te llevan la contraria para no molestar. El resultado sería fabuloso. Pasaríamos de reuniones de dos horas en las que una hora y cuarenta minutos se dedican al aire acondicionado a otro nivel.

Bueno, cuando Google lo implemente para ellos, luego ya miramos los demás si eso.

¡Ay qué difícil es vender!

[Cadavre Exquis]

He estado dudando entre si colgarlo aquí o en el hilo de El fin del trabajo, pero a pesar de estar centrado en el impacto que la IA tendrá en el mercado laboral, creo que encaja mejor en este hilo.

Marc Vidal | La mayor revolución laboral de la historia ya se ha iniciado

Saludos.

[Saturio]

He estado dudando entre si colgarlo aquí o en el hilo de El fin del trabajo, pero a pesar de estar centrado en el impacto que la IA tendrá en el mercado laboral, creo que encaja mejor en este hilo.

Marc Vidal | La mayor revolución laboral de la historia ya se ha iniciado

Saludos.

Mira Marc. Uber lleva funcionando desde 2010. 13 puñeteros años. Y en mi pueblo ni está ni se le espera. Jamás veré a Uber. Los resultados de esa compañía fueron de -9.200 millones de dólares el año pasado.
Cabify sí funciona. Pero sus tarifas son peores que las del taxi. Y también pierde dinero.

Me gustaría que Marc Vidal que es una de esas mentes que entiende, acepta y abraza el futuro, me explicase cual es la magia que hace que una empresa que ha perdido 24.000 millones en tres años (y todo lo que ha perdido en los otros 10 anteriores) siga funcionando. Igual eso explica mejor el mundo que las disrupciones tecnológicas.

Bien, ahora me hablas de los robotaxis que es Uber versión 2.0 y de la AI como disruptora. Por cierto, me encanta el verbo disromper puesto que el prefijo "dis"  significa contrariedad o contrario como esn disgusto (o su versión "des" en desvestir) o como en las clases de economía nos hablaban de los disbienes que son aquellas cosas de las que prefieres tener menos que más (y que por lo tanto pagas para que se las lleven).

Al grano, que me lío. La AI no es mala, dice Marc, salvo que te prepares. Ergo, <a href="http://salvo que consumas." target="_blank" class="new_win">http://salvo que consumas.</a>. Que de eso se trata todo, que carajo. ¡Qué lástima que no se pase mpt por aquí!. Porque creo que el meme de ¡Qué difícil es vender! le gustaría un montón (Como "de la petanca no se vive").

¡Qué difícil es vender, Marc!.

No tengo muy claro lo que pretendes vendernos. Parece que educación privada de esa adaptada de verdad a los cambiantes tiempos con tecnologías disruptoras. No nos dices en que consiste esa adaptación. Debe ser que las escuelas tienen que dejar de explicar lo que es un cabo o una península. O dejar de enseñar que los planetas orbitan en torno al sol o dejar de enseñar logaritmos o límites, o dejar de enseñar cosas anticuadas como qué es la rima.

Todos estos mensajes calan mucho en el pijerío mediopelista que asustado lleva a sus hijos a clase de programación. No tengo nada contra las clases de programación en sí. Programar es algo creativo e interesante. Una actividad genial para los chavales como pintar o hacer música. Lo que hago es mearme de la risa con la pequñísimo burguesía y su puto miedo a todo.

¡Qué difícil es vender, Marc!

Van a venir los okupas a quedarse con nuestras casas. Pongamos alarmas de Securitas Direct. Van a venir los disruptores a dejarnos sin empleo, consumamos algo para evitarlo.

¡Qué difícil es vender!

[sudden and sharp]

La inteligencia artificial al servicio de la sanidad madrileña: Una revolución en diagnóstico
https://noticias.ai/inteligencia-artificial-servicio-sanidad-madrilena-revolucion-diagnostico/

El mundo médico está a punto de experimentar una revolución en la Comunidad de Madrid. El Servicio Madrileño de Salud (SERMAS) da un paso pionero, siendo el primero en España, y globalmente, en la adopción de tecnologías avanzadas de Inteligencia Artificial (IA) Generativa para el diagnóstico clínico de enfermedades raras.

Originado de un acuerdo colaborativo entre la Consejería de Digitalización de la Comunidad de Madrid, la Fundación 29 y Microsoft, se ha desarrollado una innovadora solución: DxGPT. Esta aplicación web, que se prevé sea accesible desde finales de septiembre en los centros de Atención Primaria de la Comunidad de Madrid, opera sobre el modelo de lenguaje de IA de OpenAI, GPT-4, en Azure OpenAI Service de Microsoft, garantizando rigurosos estándares de seguridad y ética.

[Cadavre Exquis]

https://cincodias.elpais.com/companias/2023-09-22/las-tecnologicas-se-arman-para-evitar-los-abusos-de-la-inteligencia-artificial.html

Saludos.

[Cadavre Exquis]

https://cincodias.elpais.com/opinion/2023-09-22/ia-por-una-regulacion-razonable-y-un-uso-responsable.html

Saludos.

[Cadavre Exquis]

Dot CSV | Análisis de GPT-4V | ¡La VISIÓN ya ha llegado a ChatGPT!

The Dawn of LMMs: Preliminary Explorations with GPT-4V(ision)
Zhengyuan Yang, Linjie Li, Kevin Lin, Jianfeng Wang, Chung-Ching Lin, Zicheng Liu, Lijuan Wang

Large multimodal models (LMMs) extend large language models (LLMs) with multi-sensory skills, such as visual understanding, to achieve stronger generic intelligence. In this paper, we analyze the latest model, GPT-4V(ision), to deepen the understanding of LMMs. The analysis focuses on the intriguing tasks that GPT-4V can perform, containing test samples to probe the quality and genericity of GPT-4V's capabilities, its supported inputs and working modes, and the effective ways to prompt the model. In our approach to exploring GPT-4V, we curate and organize a collection of carefully designed qualitative samples spanning a variety of domains and tasks. Observations from these samples demonstrate that GPT-4V's unprecedented ability in processing arbitrarily interleaved multimodal inputs and the genericity of its capabilities together make GPT-4V a powerful multimodal generalist system. Furthermore, GPT-4V's unique capability of understanding visual markers drawn on input images can give rise to new human-computer interaction methods such as visual referring prompting. We conclude the report with in-depth discussions on the emerging application scenarios and the future research directions for GPT-4V-based systems. We hope that this preliminary exploration will inspire future research on the next-generation multimodal task formulation, new ways to exploit and enhance LMMs to solve real-world problems, and gaining better understanding of multimodal foundation models. Finally, we acknowledge that the model under our study is solely the product of OpenAI's innovative work, and they should be fully credited for its development. Please see the GPT-4V contributions paper for the authorship and credit attribution: this https URL

https://llava-vl.github.io/

Saludos.

[puede ser]

https://www.elconfidencial.com/motor/nueva-movilidad/2023-11-09/taxi-autonomo-cruise-california-gm-bolt-licencia-atropello_3770844/

General Motors deberá revisar 950 coches
General Motors y Cruise retiran sus taxis autónomos tras el atropello de un peatón
La compañía Cruise, especializada en taxis de conducción autónoma, retirará temporalmente 950 coches de las calles de San Francisco para actualizar su software y mejorar la respuesta ante situaciones como la del pasado 2 de octubre
(...)
Además, el incidente se produce cuando Cruise ya se enfrenta a la investigación por dos incidentes anteriores, en los que taxis robotizados de la compañía no habrían cedido el paso a los peatones en zonas donde sí deberían haberlo hecho
(...)

No sé si esto va bien aquí o estaba mejor en el hilo de los inventos 

[Saturio]

https://www.elconfidencial.com/motor/nueva-movilidad/2023-11-09/taxi-autonomo-cruise-california-gm-bolt-licencia-atropello_3770844/

General Motors deberá revisar 950 coches
General Motors y Cruise retiran sus taxis autónomos tras el atropello de un peatón
La compañía Cruise, especializada en taxis de conducción autónoma, retirará temporalmente 950 coches de las calles de San Francisco para actualizar su software y mejorar la respuesta ante situaciones como la del pasado 2 de octubre
(...)
Además, el incidente se produce cuando Cruise ya se enfrenta a la investigación por dos incidentes anteriores, en los que taxis robotizados de la compañía no habrían cedido el paso a los peatones en zonas donde sí deberían haberlo hecho
(...)

No sé si esto va bien aquí o estaba mejor en el hilo de los inventos 

A la fuerza ahorcan, al menos para Cruise:

California DMV suspends permits for Cruise driverless robotaxis

https://www.cbsnews.com/sanfrancisco/news/cruise-llc-driverless-robotaxis-california-dmv-suspends-permits/

[Cadavre Exquis]

Former President Obama Warns 'Disruptive' AI May Require Rethinking Jobs and the Economy
Posted by EditorDavid on Sunday November 12, 2023 @07:24AM from the yes-we-rethink dept.

This week the Verge's podcast Decoder interviewed former U.S. president Barack Obama for a discussion on "AI, free speech, and the future of the internet."

Obama warns that future copyright questions are just part of a larger issue. "If AI turns out to be as pervasive and as powerful as it's proponents expect — and I have to say the more I look into it, I think it is going to be that disruptive — we are going to have to think about not just intellectual property; we are going to have to think about jobs and the economy differently."

Specific issues may include the length of the work week and the fact that health insurance coverage is currently tied to employment — but it goes far beyond that:

The broader question is going to be what happens when 10% of existing jobs now definitively can be done by some large language model or other variant of AI? And are we going to have to reexamine how we educate our kids and what jobs are going to be available...?

The truth of the matter is that during my presidency, there was I think a little bit of naivete, where people would say, you know, "The answer to lifting people out of poverty and making sure they have high enough wages is we're going to retrain them and we're going to educate them, and they should all become coders, because that's the future." Well, if AI's coding better than all but the very best coders? If ChatGPT can generate a research memo better than the third-, fourth-year associate — maybe not the partner, who's got a particular expertise or judgment? — now what are you telling young people coming up?

While Obama believes in the transformative potential of AI, "we have to be maybe a little more intentional about how our democracies interact with what is primarily being generated out of the private sector. What rules of the road are we setting up, and how can we make sure that we maximize the good and maybe minimize some of the bad?"

AI's impact will be a global problem, Obama believes, which may require "cross-border frameworks and standards and norms". (He expressed a hope that governments can educate the public on the idea that AI is "a tool, not a buddy".) During the 44-minute interview Obama predicted AI will ultimately force a "much more robust" public conversation about rules needed for social media — and that at least some of that pressure could come from how consumers interact with companies. (Obama also argues there will still be a market for products that don't just show you what you want to see.)

"One of Obama's worries is that the government needs insight and expertise to properly regulate AI," writes the Verge's editor-in-chief in an article about the interview, "and you'll hear him make a pitch for why people with that expertise should take a tour of duty in the government to make sure we get these things right."

You'll hear me get excited about a case called Red Lion Broadcasting v. FCC, a 1969 Supreme Court decision that said the government could impose something called the Fairness Doctrine on radio and television broadcasters because the public owns the airwaves and can thus impose requirements on how they're used. There's no similar framework for cable TV or the internet, which don't use public airwaves, and that makes them much harder, if not impossible, to regulate. Obama says he disagrees with the idea that social networks are something called "common carriers" that have to distribute all information equally.

Obama also applauded last month's newly-issued Executive Order from the White House, a hundred-page document which Obama calls important as "the beginning of building out a framework."

We don't know all the problems that are going to arise out of this. We don't know all the promising potential of AI, but we're starting to put together the foundations for what we hope will be a smart framework for dealing with it... In talking to the companies themselves, they will acknowledge that their safety protocols and their testing regimens may not be where they need to be yet. I think it's entirely appropriate for us to plant a flag and say, "All right, frontier companies, you need to disclose what your safety protocols are to make sure that we don't have rogue programs going off and hacking into our financial system," for example. Tell us what tests you're using. Make sure that we have some independent verification that right now this stuff is working.

But that framework can't be a fixed framework. These models are developing so quickly that oversight and any regulatory framework is going to have to be flexible, and it's going to have to be nimble.

Saludos.

[Cadavre Exquis]

https://www.expansion.com/tecnologia/2023/11/11/654ea014468aebe54d8b458b.html

[pollo]

Former President Obama Warns 'Disruptive' AI May Require Rethinking Jobs and the Economy
Posted by EditorDavid on Sunday November 12, 2023 @07:24AM from the yes-we-rethink dept.

This week the Verge's podcast Decoder interviewed former U.S. president Barack Obama for a discussion on "AI, free speech, and the future of the internet."

Obama warns that future copyright questions are just part of a larger issue. "If AI turns out to be as pervasive and as powerful as it's proponents expect — and I have to say the more I look into it, I think it is going to be that disruptive — we are going to have to think about not just intellectual property; we are going to have to think about jobs and the economy differently."

Specific issues may include the length of the work week and the fact that health insurance coverage is currently tied to employment — but it goes far beyond that:

The broader question is going to be what happens when 10% of existing jobs now definitively can be done by some large language model or other variant of AI? And are we going to have to reexamine how we educate our kids and what jobs are going to be available...?

The truth of the matter is that during my presidency, there was I think a little bit of naivete, where people would say, you know, "The answer to lifting people out of poverty and making sure they have high enough wages is we're going to retrain them and we're going to educate them, and they should all become coders, because that's the future." Well, if AI's coding better than all but the very best coders? If ChatGPT can generate a research memo better than the third-, fourth-year associate — maybe not the partner, who's got a particular expertise or judgment? — now what are you telling young people coming up?

While Obama believes in the transformative potential of AI, "we have to be maybe a little more intentional about how our democracies interact with what is primarily being generated out of the private sector. What rules of the road are we setting up, and how can we make sure that we maximize the good and maybe minimize some of the bad?"

AI's impact will be a global problem, Obama believes, which may require "cross-border frameworks and standards and norms". (He expressed a hope that governments can educate the public on the idea that AI is "a tool, not a buddy".) During the 44-minute interview Obama predicted AI will ultimately force a "much more robust" public conversation about rules needed for social media — and that at least some of that pressure could come from how consumers interact with companies. (Obama also argues there will still be a market for products that don't just show you what you want to see.)

"One of Obama's worries is that the government needs insight and expertise to properly regulate AI," writes the Verge's editor-in-chief in an article about the interview, "and you'll hear him make a pitch for why people with that expertise should take a tour of duty in the government to make sure we get these things right."

You'll hear me get excited about a case called Red Lion Broadcasting v. FCC, a 1969 Supreme Court decision that said the government could impose something called the Fairness Doctrine on radio and television broadcasters because the public owns the airwaves and can thus impose requirements on how they're used. There's no similar framework for cable TV or the internet, which don't use public airwaves, and that makes them much harder, if not impossible, to regulate. Obama says he disagrees with the idea that social networks are something called "common carriers" that have to distribute all information equally.

Obama also applauded last month's newly-issued Executive Order from the White House, a hundred-page document which Obama calls important as "the beginning of building out a framework."

We don't know all the problems that are going to arise out of this. We don't know all the promising potential of AI, but we're starting to put together the foundations for what we hope will be a smart framework for dealing with it... In talking to the companies themselves, they will acknowledge that their safety protocols and their testing regimens may not be where they need to be yet. I think it's entirely appropriate for us to plant a flag and say, "All right, frontier companies, you need to disclose what your safety protocols are to make sure that we don't have rogue programs going off and hacking into our financial system," for example. Tell us what tests you're using. Make sure that we have some independent verification that right now this stuff is working.

But that framework can't be a fixed framework. These models are developing so quickly that oversight and any regulatory framework is going to have to be flexible, and it's going to have to be nimble.

Saludos.

Se sobreestima mucho la capacidad de la IA y se subestima mucho, pero mucho, la complejidad de muchas tareas y asuntos nada triviales como la propia definición de las mismas. Está claro que uno de estos inventos puede ahorrar tiempo en determinados casos, pero hay otras cosas en las que se está cayendo constantemente en el efecto Dunning-Kruger y se subestima la complejidad oculta de muchos trabajos. Cuando vemos a un gran maestro tocar un instrumento siempre parece fácil.

Con esto va a haber muchos chascos. Los que tengan un poco de cabeza y no se dejen arrastrar por toda la flipadura, podrán sacar partido de ello. Los que se pasen y apuesten todo al rojo (despidamos a media plantilla, y los cambiamos por un PC), van a verse tirados y es posible que hundan sus negocios de una forma estúpida.

Y los que contraten exclusivamente novatos o jóvenes recién egresados y pretendan sacar algo adelante con una IA y gente inexperta, van a llevarse la mayor hostia de todas, porque eso será una bomba de relojería.

[uno]

jajajajajaja

Han pedido a ChatGPT el número del Gordo del Sorteo de Lotería de Navidad 2023, y se ha agotado en minutos en Elche

https://www.genbeta.com/actualidad/han-pedido-a-chatgpt-numero-gordo-sorteo-loteria-navidad-2023-se-ha-agotado-minutos-elche

Esto demuestra una vez más la confianza que la gente deposita en ChatGPT, y en este caso en GPT-4. Y es que un número que se ha obtenido a partir de la serie histórica y que en el fondo tiene una probabilidad similar al resto de números al no seguir el sorteo un patrón estadístico que pueda anticiparnos cuál va a salir en la siguiente tirada del gran bombo.