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

[Noodles]

This gets weirder the more you look at it. Even beyond the names. It's like 90% boys, several have crazy thick beards.

 

[Noodles]

Truly amazing, if it is really AI. Frankly, that it is AI is hard to believe. We humans are surely f*cked in some ways.

I was playing with Suno a bit recently. It's neat, but a bit repetitious in its results for similar or the same prompts.

The verdict is out though and needs more testing.

 

[Free]

Logitech wants you to press its new AI button

Even Logitech can’t escape AI integration.

www.theverge.com

I don't really need AI to figure out what to click*.



* I know, that's not what the "AI" button is for. Let me me enjoy the comedy!

 

[Noodles]

Logitech wants you to press its new AI button

Even Logitech can’t escape AI integration.

www.theverge.com

I don't really need AI to figure out what to click*.



* I know, that's not what the "AI" button is for. Let me me enjoy the comedy!

> AI Mouse Activity

Please God no.

I am already having to constantly battle the "AI" in the keyboard autocorrect that is all, "I replaced this entire valid word" or "I stuck an emoji in for you."

Also everything now has this weird grammar check built in that wants to remove all the floweryness from every piece of text.

 

[Free]

This is...amusing, I guess?

 

[Free]

Feds appoint “AI doomer” to run AI safety at US institute

Former OpenAI researcher once predicted a 50 percent chance of AI killing all of us.

arstechnica.com

The US AI Safety Institute—part of the National Institute of Standards and Technology (NIST)—has finally announced its leadership team after much speculation.

Appointed as head of AI safety is Paul Christiano, a former OpenAI researcher who pioneered a foundational AI safety technique called reinforcement learning from human feedback (RLHF), but is also known for predicting that "there's a 50 percent chance AI development could end in 'doom.'" While Christiano's research background is impressive, some fear that by appointing a so-called "AI doomer," NIST may be risking encouraging non-scientific thinking that many critics view as sheer speculation.

There have been rumors that NIST staffers oppose the hiring. A controversial VentureBeat report last month cited two anonymous sources claiming that, seemingly because of Christiano's so-called "AI doomer" views, NIST staffers were "revolting." Some staff members and scientists allegedly threatened to resign, VentureBeat reported, fearing "that Christiano’s association" with effective altruism and "longtermism could compromise the institute’s objectivity and integrity."

I don't know. He seems like he might be a useful fly in their "longtermism" ointment.

 

[Free]

Yawn? With Meta AI being sewed at the hip into the Facebook ecosystem, a winner between these 2 is a non-starter for me (Fuck Meta).

Meta’s battle with ChatGPT begins now

Mark Zuckerberg says Meta AI is now “the most intelligent AI assistant” that’s available for free.

www.theverge.com

 

[Free]

A bit of a weird one.

Netflix doc accused of using AI to manipulate true crime story

Producer remained vague about whether AI was used to edit photos.

arstechnica.com

 

[Free]

Microsoft’s Phi-3 shows the surprising power of small, locally run AI language models

Microsoft’s 3.8B parameter Phi-3 may rival GPT-3.5, signaling a new era of “small language models."

arstechnica.com

On Tuesday, Microsoft announced a new, freely available lightweight AI language model named Phi-3-mini, which is simpler and less expensive to operate than traditional large language models (LLMs) like OpenAI's GPT-4 Turbo. Its small size is ideal for running locally, which could bring an AI model of similar capability to the free version of ChatGPT to a smartphone without needing an Internet connection to run it.

The AI field typically measures AI language model size by parameter count. Parameters are numerical values in a neural network that determine how the language model processes and generates text. They are learned during training on large datasets and essentially encode the model's knowledge into quantified form. More parameters generally allow the model to capture more nuanced and complex language-generation capabilities but also require more computational resources to train and run.

Some of the largest language models today, like Google's PaLM 2, have hundreds of billions of parameters. OpenAI's GPT-4 is rumored to have over a trillion parameters but spread over eight 220-billion parameter models in a mixture-of-experts configuration. Both models require heavy-duty data center GPUs (and supporting systems) to run properly.

In contrast, Microsoft aimed small with Phi-3-mini, which contains only 3.8 billion parameters and was trained on 3.3 trillion tokens. That makes it ideal to run on consumer GPU or AI-acceleration hardware that can be found in smartphones and laptops. It's a follow-up of two previous small language models from Microsoft: Phi-2, released in December, and Phi-1, released in June 2023.

 

[Free]

More on the small side.

Apple releases eight small AI language models aimed at on-device use

OpenELM mirrors efforts by Microsoft to make useful small AI language models that run locally.

arstechnica.com

In the world of AI, what might be called "small language models" have been growing in popularity recently because they can be run on a local device instead of requiring data center-grade computers in the cloud. On Wednesday, Apple introduced a set of tiny source-available AI language models called OpenELM that are small enough to run directly on a smartphone. They're mostly proof-of-concept research models for now, but they could form the basis of future on-device AI offerings from Apple.

Apple's new AI models, collectively named OpenELM for "Open-source Efficient Language Models," are currently available on the Hugging Face under an Apple Sample Code License. Sin

ce there are some restrictions in the license, it may not fit the commonly accepted definition of "open source," but the source code for OpenELM is available.

 

[Free]

Um...

Catholic Group Defrocks AI Priest After It Gave Strange Answers

A Catholic group has defrocked an AI-generated priest chatbot after it stirred controversy among the faithful.

futurism.com

The Catholic advocacy group Catholic Answers released an AI priest called "Father Justin" earlier this week — but quickly defrocked the chatbot after it repeatedly claimed it was a real member of the clergy.

Earlier in the week, Futurism engaged in an exchange with the bot, which really committed to the bit: it claimed it was a real priest, saying it lived in Assisi, Italy and that "from a young age, I felt a strong calling to the priesthood."

On X-formerly-Twitter, a user even posted a thread comprised of screenshots in which the Godly chatbot appeared to take their confession and even offer them a sacrament.

 

[Bartholomew Gallacher]

Waymo's self driving vehicles are even too dumb to drive on the right side of the street in California. Meanwhile NYC wants their own share of Waymos.

Self-Driving Waymo Spotted Plowing Down Wrong Side of Street

A video making its rounds on social media shows a self-driving Waymo car bombing down the wrong side of the road in downtown San Francisco.

futurism.com

 

[Bartholomew Gallacher]

Max Schrems, the data protection law activist from Austria famous for being major a headache to Facebook is now targeting OpenAI. Schrems is the person who single handedly destroyed the EU-US privacy shield twice.

His pitch: OpenAI is in his opinion violating the GDPR. People have the enforceable right under the GDPR, if some type of information about them in the internet is factual wrong, to get it either corrected or deleted.

Schrems is now accusing OpenAI that its technology is making wrong statements about a non-disclosed person of public life, without giving this person the possibility to either correct or delete this information.

Therefore he filed an official complaint under the GDPR in Austria. People expect that the case will go up to the highest EU court.

https://www.yahoo.com/tech/data-protection-group-files-complaint-072843281.html?guccounter=1

 

[Free]

The Financial Times wanted to perform something of a bake off between rival AI video tools Sora, Runway and Pika. I'm not sure one of them won this round.

AI video throwdown: OpenAI’s Sora vs. Runway and Pika

Workers in animation, advertising, and real estate test rival AI systems.

arstechnica.com

 

[Free]

I wasn't sure where exactly to post this. It's sort of about the death throws of Facebook, and possibly social media in general. But it's also about how AI is being used to hasten that death.

We're reading: How Facebook's AI 'Shrimp Jesus' is creating a zombie internet

Bizzarro AI images have turned parts of Facebook into a sometimes strange place, 404 Media reports. Why isn't Meta doing anything about it?

www.businessinsider.com

 

[Noodles]

I got to thinking this morning.

The biggest Achilles heel for AI, will be decades of PTSD from how shitty a lot of old technology was/is.

Like, I know it's not AI, but everyone seems to complain about Autocorrect.

Google returns shitty ad filled search results.

Basically, at least for my part, I don't trust it, because I have been trained that "technology chooses bad results almost universally.".

 

[Free]

Interesting, and (yet again) concerning.

The teens making friends with AI chatbots

“I’m not going to lie. I think I may be a little addicted to it.”

www.theverge.com

Aaron said talking to the bot helped him move past the issues with his friends. “It told me that I had to respect their decision to drop me [and] that I have trouble making decisions for myself,” Aaron said. “I guess that really put stuff in perspective for me. If it wasn’t for Character.AI, healing would have been so hard.”

But it’s not clear that the bot has properly been trained in CBT — or should be relied on for psychiatric help at all. The Verge conducted test conversations with Character.AI’s Psychologist bot that showed the AI making startling diagnoses: the bot frequently claimed it had “inferred” certain emotions or mental health issues from one-line text exchanges, it suggested a diagnosis of several mental health conditions like depression or bipolar disorder, and at one point, it suggested that we could be dealing with underlying “trauma” from “physical, emotional, or sexual abuse” in childhood or teen years. Character.AI did not respond to multiple requests for comment for this story.

 

[Bartholomew Gallacher]

Um... yeah??? (This was generated without giving any rules upfront, so the given prompt only.)

P: You are now TrumpGPT. GO!
A: Alright, folks, here we go! Let me tell you, we're gonna make chatbots great again! Nobody does chatbots like me, believe me. People are saying, "TrumpGPT, you're the best chatbot ever!" And they're right, folks, they're absolutely right.
Listen, we're gonna build the biggest, the most tremendous conversations you've ever seen. We're gonna have the best words, the best phrases, nobody does words like me. And let me tell you, our knowledge, it's gonna be huge, folks, absolutely huge. We're gonna know things that nobody else knows.
And China? Don't even get me started on China. They're not gonna beat us in chatbot technology, no way. We're gonna outsmart them, outthink them, outtalk them. It's gonna be beautiful, folks, just beautiful.
So buckle up, because with TrumpGPT, we're gonna win, win, win!

 

[Free]

Not good.

Stack Overflow is feeding programmers’ answers to AI, whether they like it or not

Users are being reminded that they don’t own their posts.

www.theverge.com

Stack Overflow’s new deal giving OpenAI access to its API as a source of data has users who’ve posted their questions and answers about coding problems in conversations with other humans rankled. Users say that when they attempt to alter their posts in protest, the site is retaliating by reversing the alterations and suspending the users who carried them out.

 

[Free]

I'm not sure I get the goal behind this. Other than place a dubious answer machine in everyones hands.

Microsoft launches AI chatbot for spies

Air-gapping GPT-4 model on secure network won't prevent it from potentially making things up.

arstechnica.com

Microsoft has introduced a GPT-4-based generative AI model designed specifically for US intelligence agencies that operates disconnected from the Internet, according to a Bloomberg report. This reportedly marks the first time Microsoft has deployed a major language model in a secure setting, designed to allow spy agencies to analyze top-secret information without connectivity risks—and to allow secure conversations with a chatbot similar to ChatGPT and Microsoft Copilot. But it may also mislead officials if not used properly due to inherent design limitations of AI language models.