Nobody cares about "AI" (Chatbot: I disagree.)

[Free]

Yeah, but it's kind of amusing if you look at it from the perspective of "Star Wars being real". Like, it's an entire galaxy of droids, and they all seem to have these weird personality quirks and hang-ups.

Not true. The gonk droids all tend to play it pretty straight.

 

[Noodles]

Not true. The gonk droids all tend to play it pretty straight.

Maybe. How do we know that "Gonk" isn't just droid speak for "Penis" and the thing isn't just walking around going "Penis.... penis.... penis..."

 

[Free]

Maybe. How do we know that "Gonk" isn't just droid speak for "Penis" and the thing isn't just walking around going "Penis.... penis.... penis..."

Constantly saying "Penis.... penis.... penis..." does not make one sassy. Though it can come off as gay.

 

[Noodles]

I guess that's true, I guess its more just being a dick than actually being sassy.

 

[Essence Lumin]

Maybe. How do we know that "Gonk" isn't just droid speak for "Penis" and the thing isn't just walking around going "Penis.... penis.... penis..."

Etsy has this.

 

[Free]

Etsy has this.

Troll dick pic?

 

[Free]

AI language models can exceed PNG and FLAC in lossless compression, says study

Is compression equivalent to general intelligence? DeepMind digs up more potential clues.

arstechnica.com

Effective compression is about finding patterns to make data smaller without losing information. When an algorithm or model can accurately guess the next piece of data in a sequence, it shows it's good at spotting these patterns. This links the idea of making good guesses—which is what large language models like GPT-4 do very well—to achieving good compression.

In an arXiv research paper titled "Language Modeling Is Compression," researchers detail their discovery that the DeepMind large language model (LLM) called Chinchilla 70B can perform lossless compression on image patches from the ImageNet image database to 43.4 percent of their original size, beating the PNG algorithm, which compressed the same data to 58.5 percent. For audio, Chinchilla compressed samples from the LibriSpeech audio data set to just 16.4 percent of their raw size, outdoing FLAC compression at 30.3 percent.

Over the past two decades, some computer scientists have proposed that the ability to compress data effectively is akin to a form of general intelligence. The idea is rooted in the notion that understanding the world often involves identifying patterns and making sense of complexity, which, as mentioned above, is similar to what good data compression does. By reducing a large set of data into a smaller, more manageable form while retaining its essential features, a compression algorithm demonstrates a form of understanding or representation of that data, proponents argue.

 

[Essence Lumin]

AI language models can exceed PNG and FLAC in lossless compression, says study

Is compression equivalent to general intelligence? DeepMind digs up more potential clues.

arstechnica.com

Interesting, but there are other factors. Speed of compression, decompression, amount of ram and cpu. An audio file format such as flac is not of much use if it takes longer to decode than the speed the song plays at.

 

[Bartholomew Gallacher]

Microsoft is looking into getting its own nuclear power plants to get masses of cheap energy for training their AI models. They are evaluating SMR designs for that task.

Microsoft is going nuclear to power its AI ambitions

Microsoft is hiring someone to lead its nuclear strategy.

www.theverge.com

 

[Free]

Microsoft is looking into getting its own nuclear power plants to get masses of cheap energy for training their AI models. They are evaluating SMR designs for that task.

Microsoft is going nuclear to power its AI ambitions

Microsoft is hiring someone to lead its nuclear strategy.

www.theverge.com

Apparently it'll be a while before they discover the "human battery" solution.

 

[Free]

There's always a way.

 

[Free]

Que sorprendida estoy...

4chan users manipulate AI tools to unleash torrent of racist images

AI tools are "supercharging" racists online, report says.

arstechnica.com

Despite leading AI companies' attempts to block users from turning AI image generators into engines of racist content, many 4chan users are still turning to these tools to "quickly flood the Internet with racist garbage," 404 Media reported.

404 Media uncovered one 4chan thread where users recommended various AI tools, including Stable Diffusion and DALL-E, but specifically linked to Bing AI's text-to-image generator (which is powered by DALL-E 3) as a "quick method." After finding the right tool—which could also be a more old-school photo-editing tool like Photoshop—users are instructed to add incendiary captions and share the images on social media to create a blitz of racist images online.

 

[Innula Zenovka]

AI was asked to create images of Black African docs treating white kids. How'd it go?

 

[Bartholomew Gallacher]

Ufo lovers in the US are now using ChatGPT to "decode" blackened files from the US government.

 

[Free]

Some people should not be lawyers.

Rapper Pras’ lawyer used AI to defend him in criminal case—it did not go well

Pras seeks new trial, says lawyer had "undisclosed financial stake" in AI tool.

arstechnica.com

After being convicted of federal crimes related to a foreign influence campaign, rapper Prakazrel "Pras" Michel says he should get a new trial because his lawyer "used an experimental AI program to write his closing argument." In a motion for a new trial filed Monday, the Fugees rapper's new lawyers say Michel's previous representation used "an experimental AI program in which they had a financial stake to write the closing argument, resulting in a frivolous and ineffectual closing argument."

Michel was represented at trial by defense counsel David Kenner, who is accused of failing to provide a cogent defense and misattributing two songs to the Fugees. The allegations about Kenner's use of AI are reminiscent of a previous incident in which a lawyer admitted using ChatGPT to help write court filings that cited six nonexistent cases invented by the artificial intelligence tool.

According to the motion for a new trial, "Kenner failed to familiarize himself with the charged statutes, causing him to overlook critical weaknesses in the Government's case," and he "did not understand the facts or allegations."

Kenner "outsourced trial preparations to inexperienced contract attorneys who worked for an e-discovery vendor, Business Intelligence Associates, Inc. (BIA)," the motion said. His trial team included BIA co-founder Alon Israely, "a non-practicing attorney with no white collar or even litigation experience." The motion said that Kenner generated his closing argument "using a proprietary prototype AI program in which he and Alon Israely appear to have had an undisclosed financial stake."

My actual favorite part:

"Far from hiding this fact, Kenner boasted about it after Michel was convicted, stating; 'The system turned hours or days of legal work into seconds,'" the motion said. "The AI company touted it as the first use of 'generative AI in a federal trial.' It showed.

 

[Free]

University of Chicago researchers seek to “poison” AI art generators with Nightshade

Altered images could destroy AI model training efforts that scrape art without consent.

arstechnica.com

On Friday, a team of researchers at the University of Chicago released a research paper outlining "Nightshade," a data poisoning technique aimed at disrupting the training process for AI models, reports MIT Technology Review and VentureBeat. The goal is to help visual artists and publishers protect their work from being used to train generative AI image synthesis models, such as Midjourney, DALL-E 3, and Stable Diffusion.

The open source "poison pill" tool (as the University of Chicago's press department calls it) alters images in ways invisible to the human eye that can corrupt an AI model's training process. Many image synthesis models, with notable exceptions of those from Adobe and Getty Images, largely use data sets of images scraped from the web without artist permission, which includes copyrighted material.

 

[Free]

Hammer drop in the U.S.

Biden issues sweeping executive order that touches AI risk, deepfakes, privacy

Order details US admin's approach to AI safety, media authenticity, job loss, and more.

arstechnica.com

On Monday, President Joe Biden issued an executive order on AI that outlines the federal government's first comprehensive regulations on generative AI systems. The order includes testing mandates for advanced AI models to ensure they can't be used for creating weapons, suggestions for watermarking AI-generated media, and provisions addressing privacy and job displacement.

Using his authority to set terms for government contracts, Biden aims to influence AI standards by stipulating that federal agencies must only enter into contracts with companies that comply with the government's newly outlined AI regulations. This approach utilizes the federal government's purchasing power to drive compliance with the newly set standards.

 

[Innula Zenovka]

 

[Free]

This happening was a certainty.

Teen boys use AI to make fake nudes of classmates, sparking police probe

Parents told the high school "believed" the deepfake nudes were deleted.

arstechnica.com

This October, boys at Westfield High School in New Jersey started acting "weird," the Wall Street Journal reported. It took four days before the school found out that the boys had been using AI image generators to create and share fake nude photos of female classmates. Now, police are investigating the incident, but they're apparently working in the dark, because they currently have no access to the images to help them trace the source.

According to an email that the WSJ reviewed from Westfield High School principal Mary Asfendis, the school "believed" that the images had been deleted and were no longer in circulation among students.

It remains unclear how many students were harmed. A Westfield Public Schools spokesperson cited student confidentiality when declining to tell the WSJ the total number of students involved or how many students, if any, had been disciplined. The school had not confirmed whether faculty had reviewed the images, seemingly only notifying the female students allegedly targeted when they were identified by boys claiming to have seen the images.

It's also unclear if what the boys did was illegal. There is currently no federal law restricting the creation of faked sexual images of real people, the WSJ reported, and in June, child safety experts reported that there was seemingly no way to stop thousands of realistic but fake AI child sex images from being shared online.

 

[Noodles]

It mentions about how there isn't a law against making fake nudes of people, I also wonder since theybare students if it would even count as CP, legally. Its probably something that may have to go to court to be decided eventually, and maybe in this case everyone was over 18.