"We Need to Talk About How Good A.I. Is Getting"

[funkervogt]

This NYT article is paywalled, so here's a way to see it for free: https://archive.ph/CbYzu#selection-313.0-313.46

The first half of the article recaps the last five years of progress in machine learning, and goes over things we've been discussing at length on this forum (GPT-3, DALL-E 2, etc.). Here are the bits that were interesting to me:

Ajeya Cotra, a senior analyst with Open Philanthropy who studies A.I. risk, estimated two years ago that there was a 15 percent chance of “transformational A.I.” — which she and others have defined as A.I. that is good enough to usher in large-scale economic and societal changes, such as eliminating most white-collar knowledge jobs — emerging by 2036.

But in a recent post, Ms. Cotra raised that to a 35 percent chance, citing the rapid improvement of systems like GPT-3.

“A.I. systems can go from adorable and useless toys to very powerful products in a surprisingly short period of time,” Ms. Cotra told me. “People should take more seriously that A.I. could change things soon, and that could be really scary.”

Even if the skeptics are right, and A.I. doesn’t achieve human-level sentience for many years, it’s easy to see how systems like GPT-3, LaMDA and DALL-E 2 could become a powerful force in society. In a few years, the vast majority of the photos, videos and text we encounter on the internet could be A.I.-generated. Our online interactions could become stranger and more fraught, as we struggle to figure out which of our conversational partners are human and which are convincing bots. And tech-savvy propagandists could use the technology to churn out targeted misinformation on a vast scale, distorting the political process in ways we won’t see coming.

Third, the news media needs to do a better job of explaining A.I. progress to nonexperts. Too often, journalists — and I admit I’ve been a guilty party here — rely on outdated sci-fi shorthand to translate what’s happening in A.I. to a general audience. We sometimes compare large language models to Skynet and HAL 9000, and flatten promising machine learning breakthroughs to panicky “The robots are coming!” headlines that we think will resonate with readers. Occasionally, we betray our ignorance by illustrating articles about software-based A.I. models with photos of hardware-based factory robots — an error that is as inexplicable as slapping a photo of a BMW on a story about bicycles.

[Nanotechandmorefuture]

We are going to find a world where nothing is what it seems and to be somewhat reasonable some will be half cyborg at least. Right on track for that omnipresent AI like in Cyberpunk.

[peekpok]

I feel like there's going to be a sort of "machine-learning-optimized-everything" S-curve in every industry in the next 5-20 years. I don't know how far away AGI is, but narrow AI is becoming damn good and nobody can deny that. Any sort of business where reliably predicting the behavior of a complex system ahead of time makes the business more efficient will benefit. Logistics, manufacturing, farming, distribution, those types of industries. But we will also see leaps in product design, such as using ANI to optimize designs. This is already happening with computer chips, but will eventually be used for everything. Even simple stuff like reducing the total material needed to build a bridge, or whatever.

[caltrek]

"World's First Robot Lawyer" To Make History Defending A Client In Court
by Jack Dunhill
January 9, 2023

Introduction:

(IFL Science) An Artificial Intelligence (AI) lawyer is gearing up to make history as the first AI to defend a case in court. Dubbed the world’s first “robot lawyer”, the bot will give prompts to the defendant on how to best argue their case against a speeding ticket when the hearing takes place next month.

Created by startup DoNotPay, the chatbot will be on a mobile phone and will deliver prompts through headphones to the defendant, who has agreed to only say what the AI tells it to. The landmark case is designed as a test for DoNotPay’s services and they will pay for any fines incurred if the AI fails to win the case for the defendant.

While it may sound like the future of court hearings, there’s a catch – mobile phones and headphones are generally not allowed in courtrooms. However, according to the firm’s founder Joshua Browder, the hearing will take place in a location that will consider the headphones a “hearing aid”, thus allowing the AI.

“It’s technically within the rules, but I don’t think it’s in the spirit of the rules,” says Browder in a statement to New Scientist.
DoNotPay launched in 2015 as a chatbot startup aiming to help consumers get out of paying fines and late fees, and has since moved to AI following rapid advancements in technology that have occurred recently in the form of OpenAI’s GPT-3.

Read more here: https://www.iflscience.com/-world-s-fi ... urt-66986

[funkervogt]

I started this thread just five months ago. Since then, ChatGPT has been unveiled and has expanded its customer base faster than any other invention in history.

https://www.reuters.com/technology/chat ... 023-02-01/

I've also used it and find it very impressive. A ton of other tech companies are now in a race to make even better algorithms than ChatGPT, and they will surely succeed in the near future.

AI is definitely advancing quickly. I didn't think I'd see a time when major improvements like this would be happening every few months.

[ººº]

I started this thread just five months ago. Since then, ChatGPT has been unveiled and has expanded its customer base faster than any other invention in history.

https://www.reuters.com/technology/chat ... 023-02-01/

I've also used it and find it very impressive. A ton of other tech companies are now in a race to make even better algorithms than ChatGPT, and they will surely succeed in the near future.

AI is definitely advancing quickly. I didn't think I'd see a time when major improvements like this would be happening every few months.

The next Steve Jobs/Bill gates is in the making (or very soon to be).

[raklian]

The next Steve Jobs/Bill gates is in the making (or very soon to be).

Who says it has to be a human?

Imagine a predecessor AI designing and coding the next generation AI.

I suppose the next human Steve Jobs will be the one to start this freight train of self-improvement.

[funkervogt]

That's an interesting thought. What if the first AGI has no direct human inventor? A narrow AI just built a better narrow AI, which then repeated that process, until after some number of iterations an AGI arose, with no human input? It's code would be complex that human computer scientists that set the first narrow AI out on the task wouldn't comprehend the final product.

[caltrek]

...

...

The History Channel at its finest. I wish all of its programming met the high standard set by this type of show.

[caltrek]

Can Pigeons Match Wits With Artificial Intelligence?
February 7, 2023

Introduction:

(EurekAlert) Can a pigeon match wits with artificial intelligence? At a very basic level, yes.

In a new study, psychologists at the University of Iowa examined the workings of the pigeon brain and how the “brute force” of the bird’s learning shares similarities with artificial intelligence.

The researchers gave the pigeons complex categorization tests that high-level thinking, such as using logic or reasoning, would not aid in solving. Instead, the pigeons, by virtue of exhaustive trial and error, eventually were able to memorize enough scenarios in the test to reach nearly 70% accuracy.

The researchers equate the pigeons’ repetitive, trial-and-error approach to artificial intelligence. Computers employ the same basic methodology, the researchers contend, being “taught” how to identify patterns and objects easily recognized by humans. Granted, computers, because of their enormous memory and storage power—and growing ever more powerful in those domains—far surpass anything the pigeon brain can conjure.

Still, the basic process of making associations—considered a lower-level thinking technique—is the same between the test-taking pigeons and the latest AI advances.

Read more here: https://www.eurekalert.org/news-releases/978286

[caltrek]

Carnegie Mellon University's AI-Powered FRIDA Robot Collaborates With Humans to Create Art
February 7, 2023

Introduction:

(EurekaAlert) Carnegie Mellon University's Robotics Institute has a new artist-in-residence.

FRIDA, a robotic arm with a paintbrush taped to it, uses artificial intelligence to collaborate with humans on works of art. Ask FRIDA to paint a picture, and it gets to work putting brush to canvas.

"There's this one painting of a frog ballerina that I think turned out really nicely," said Peter Schaldenbrand, a School of Computer Science Ph.D. student in the Robotics Institute working with FRIDA and exploring AI and creativity. "It is really silly and fun, and I think the surprise of what FRIDA generated based on my input was really fun to see."

FRIDA, named after Frida Kahlo, stands for Framework and Robotics Initiative for Developing Arts. The project is led by Schaldenbrand with RI faculty members Jean Oh and Jim McCann, and has attracted students and researchers across CMU.

Users can direct FRIDA by inputting a text description, submitting other works of art to inspire its style, or uploading a photograph and asking it to paint a representation of it. The team is experimenting with other inputs as well, including audio. They played ABBA's "Dancing Queen" and asked FRIDA to paint it.

Read more here: https://www.eurekalert.org/news-releases/978829

[caltrek]

How AI Could Take Over Elections – and Undermine Democracy
by Archon Fung and Lawrence Lessing
June 2, 2023

Introduction:

(The Conversation) Could organizations use artificial intelligence language models such as ChatGPT to induce voters to behave in specific ways?

Sen. Josh Hawley asked OpenAI CEO Sam Altman this question in a May 16, 2023, U.S. Senate hearing on artificial intelligence. Altman replied that he was indeed concerned that some people might use language models to manipulate, persuade and engage in one-on-one interactions with voters.

Altman did not elaborate, but he might have had something like this scenario in mind. Imagine that soon, political technologists develop a machine called Clogger – a political campaign in a black box. Clogger relentlessly pursues just one objective: to maximize the chances that its candidate – the campaign that buys the services of Clogger Inc. – prevails in an election.

While platforms like Facebook, Twitter and YouTube use forms of AI to get users to spend more time on their sites, Clogger’s AI would have a different objective: to change people’s voting behavior.

How Clogger would work

As a political scientist and a legal scholar who study the intersection of technology and democracy, we believe that something like Clogger could use automation to dramatically increase the scale and potentially the effectiveness of behavior manipulation and microtargeting techniques that political campaigns have used since the early 2000s. Just as advertisers use your browsing and social media history to individually target commercial and political ads now, Clogger would pay attention to you – and hundreds of millions of other voters – individually.

Read more here: https://theconversation.com/how-ai-cou ... y-206051

[caltrek]

ChatGPT Shows ‘Impressive’ Accuracy in Clinical Decision Making
August 2, 2023

Introduction:

(Eurekalert) A new study led by investigators from Mass General Brigham has found that ChatGPT was about 72 percent accurate in overall clinical decision making, from coming up with possible diagnoses to making final diagnoses and care management decisions. The large-language model (LLM) artificial intelligence chatbot performed equally well in both primary care and emergency settings across all medical specialties. The research team’s results are published in the Journal of Medical Internet Research.

“Our paper comprehensively assesses decision support via ChatGPT from the very beginning of working with a patient through the entire care scenario, from differential diagnosis all the way through testing, diagnosis, and management,” said corresponding author Marc Succi, MD, associate chair of innovation and commercialization and strategic innovation leader at Mass General Brigham and executive director of the MESH Incubator. “No real benchmarks exists, but we estimate this performance to be at the level of someone who has just graduated from medical school, such as an intern or resident. This tells us that LLMs in general have the potential to be an augmenting tool for the practice of medicine and support clinical decision making with impressive accuracy.”

Changes in artificial intelligence technology are occurring at a fast pace and transforming many industries, including health care. But the capacity of LLMs to assist in the full scope of clinical care has not yet been studied. In this comprehensive, cross-specialty study of how LLMs could be used in clinical advisement and decision making, Succi and his team tested the hypothesis that ChatGPT would be able to work through an entire clinical encounter with a patient and recommend a diagnostic workup, decide the clinical management course, and ultimately make the final diagnosis.

Read more here: https://www.eurekalert.org/news-releases/999029

[caltrek]

Exascale Revolution: Supercomputers Unleash a New Era in Biophysics Discovery
August 22, 2023

Introduction:

(Eurekalert) In a recently published article featured on the cover of the Biophysical Journal, Dr. Rafael Bernardi, assistant professor of biophysics at the Department of Physics at Auburn University, and Dr. Marcelo Melo, a postdoctoral researcher in Dr. Bernardi's group, shed light on the transformative capabilities of the next generation of supercomputers in reshaping the landscape of biophysics.

The researchers at Auburn delve into the harmonious fusion of computational modeling and experimental biophysics, providing a perspective for a future in which discoveries are made with unparalleled precision. Rather than being mere observers, today's biophysicists, with the aid of advanced high-performance computing (HPC), are now trailblazers who can challenge longstanding biological assumptions, illuminate intricate details, and even create new proteins or design novel molecular circuits.

One of the most important aspects discussed on their perspective article is the new ability of computational biophysicists to simulate complex biological processes that range from subatomic process to whole-cell models, with extraordinary detail. As Dr. Bernardi articulates, "The new exascale computers allow computational biophysicists to go beyond what can done experimentally and simulate biological processes with a much higher level of detail. For instance, we can now understand how pathogenic bacteria bind to humans during infection at an atomistic level, generating data for AI models and opening new roads of exploration."

Historically, fields such as physics and chemistry have relied heavily on theoretical models to guide experiments. Today, biology stands at a similar crossroads, with novel software and specialized hardware becoming pivotal in deciphering experimental data and proposing innovative models. The inaugural public exascale supercomputer, Frontier, which was deployed by the Oak Ridge National Laboratory in late 2021, coupled with the rapid proliferation of artificial intelligence tools tailored for biophysics, exemplifies the profound strides being made to seamlessly bridge simulation with actual observation.

The momentum gained by computational biophysics signifies a monumental shift. As biophysical research progresses, the seamless integration of experimental and computational efforts is expected to redefine the frontiers of knowledge, laying the groundwork for unprecedented discoveries that could reshape our understanding of the biological world.

Read more here: https://www.eurekalert.org/news-releases/999144

[weatheriscool]

ChatGPT Shows ‘Impressive’ Accuracy in Clinical Decision Making
August 2, 2023

Introduction:

(Eurekalert) A new study led by investigators from Mass General Brigham has found that ChatGPT was about 72 percent accurate in overall clinical decision making, from coming up with possible diagnoses to making final diagnoses and care management decisions. The large-language model (LLM) artificial intelligence chatbot performed equally well in both primary care and emergency settings across all medical specialties. The research team’s results are published in the Journal of Medical Internet Research.

“Our paper comprehensively assesses decision support via ChatGPT from the very beginning of working with a patient through the entire care scenario, from differential diagnosis all the way through testing, diagnosis, and management,” said corresponding author Marc Succi, MD, associate chair of innovation and commercialization and strategic innovation leader at Mass General Brigham and executive director of the MESH Incubator. “No real benchmarks exists, but we estimate this performance to be at the level of someone who has just graduated from medical school, such as an intern or resident. This tells us that LLMs in general have the potential to be an augmenting tool for the practice of medicine and support clinical decision making with impressive accuracy.”

Changes in artificial intelligence technology are occurring at a fast pace and transforming many industries, including health care. But the capacity of LLMs to assist in the full scope of clinical care has not yet been studied. In this comprehensive, cross-specialty study of how LLMs could be used in clinical advisement and decision making, Succi and his team tested the hypothesis that ChatGPT would be able to work through an entire clinical encounter with a patient and recommend a diagnostic workup, decide the clinical management course, and ultimately make the final diagnosis.

Read more here: https://www.eurekalert.org/news-releases/999029

It isn't nearly good enough because I tried to use it to find news on womb transfers into biological males and it completely fucked up. It is kind of inaccurate sometimes. Got my hope up!

[Powers]

It isn't nearly good enough because I tried to use it to find news on womb transfers into biological males and it completely fucked up. It is kind of inaccurate sometimes. Got my hope up!

Are you using the free version?

[weatheriscool]

It isn't nearly good enough because I tried to use it to find news on womb transfers into biological males and it completely fucked up. It is kind of inaccurate sometimes. Got my hope up!

Are you using the free version?

Yes,

[Powers]

It isn't nearly good enough because I tried to use it to find news on womb transfers into biological males and it completely fucked up. It is kind of inaccurate sometimes. Got my hope up!

Are you using the free version?

Yes,

I should use it more by the way.

[caltrek]

The Mind of the Machine
November 1, 2023

Introduction:

(Eurekalert) In a startlingly short time span, artificial intelligence has evolved from an academic undertaking into a practical tool. Visual models like DALL·E can create images in any style an individual might fancy, while large language models (LLMs) like Chat GPT can generate essays, write computer code and suggest travel itineraries. When prompted, they can even correct their own mistakes.

As AI models become ever more sophisticated and ubiquitous, it’s crucial to understand just what these entities are, what they can do and how they think. These models are becoming very similar to humans, and yet they are so very different from us. This unique combination makes AI intriguing to contemplate.

For instance, large AI models are trained on immense amounts of information. But it isn’t clear to what extent they understand this data as a coherent system of knowledge.

Read more here: https://www.eurekalert.org/news-releases/1006652

[caltrek]

A New Kind of AI Copy Can Fully Replicate Famous People. The Law Is Powerless.
by Mohar Chatterjee
December 30, 2023

Introduction:

(Politico) Martin Seligman, the influential American psychologist, found himself pondering his legacy at a dinner party in San Francisco one late February evening. The guest list was shorter than it used to be: Seligman is 81, and six of his colleagues had died in the early Covid years. His thinking had already left a profound mark on the field of positive psychology, but the closer he came to his own death, the more compelled he felt to help his work survive.

The next morning he received an unexpected email from an old graduate student, Yukun Zhao. His message was as simple as it was astonishing: Zhao’s team had created a “virtual Seligman.”

Zhao wasn’t just bragging. Over two months, by feeding every word Seligman had ever written into cutting-edge AI software, he and his team had built an eerily accurate version of Seligman himself — a talking chatbot whose answers drew deeply from Seligman’s ideas, whose prose sounded like a folksier version of Seligman’s own speech, and whose wisdom anyone could access.

Impressed, Seligman circulated the chatbot to his closest friends and family to check whether the AI actually dispensed advice as well as he did. “I gave it to my wife and she was blown away by it,” Seligman said.

The bot, cheerfully nicknamed “Ask Martin,” had been built by researchers based in Beijing and Wuhan — originally without Seligman’s permission, or even awareness.

Read more here: https://www.politico.com/news/magazine ... 00132682