AI
- Machine Learning Researcher Links OpenAI to Drug-Fueled Sex Partiesfuturism.com Machine Learning Researcher Links OpenAI to Drug-Fueled Sex Parties
An ML researcher is claiming to have knowledge of kinky drug-fueled orgies linked to OpenAI in Silicon Valley's storied hacker houses.
- Sam Altman Clearly Freaked Out by Reaction to News of OpenAI Silencing Former Employeesfuturism.com Sam Altman Clearly Freaked Out by Reaction to News of OpenAI Silencing Former Employees
OpenAI got a lot of criticism this week for departures and a restrictive nondisclosure agreement. Sam Altman is clearly freaked.
- ChatGPT can talk, but OpenAI employees sure can’twww.vox.com ChatGPT can talk, but OpenAI employees sure can’t
Why is OpenAI’s superalignment team imploding?
- GPT-4o’s Chinese token-training data is polluted by spam and porn websiteswww.technologyreview.com GPT-4o’s Chinese token-training data is polluted by spam and porn websites
The problem, which is likely due to inadequate data cleaning, could lead to hallucinations, poor performance, and misuse.
- Gentoo and NetBSD ban 'AI' code, but Debian doesn't – yetwww.theregister.com Gentoo and NetBSD ban 'AI' code, but Debian doesn't – yet
The problem isn't just that LLM-bot generated code is bad – it's where it came from
- Aleph Alpha enlists Cerebras waferscale supers to train AI for German militarywww.theregister.com Aleph Alpha enlists Cerebras supers for German military
Also demonstrates success accelerating molecular dynamics
- Together with AI Sweden, Fraunhofer IAIS gains large-scale computing capacities for training large language modelswww.iais.fraunhofer.de Breakthrough for generative AI research in Germany and Europe
Together with AI Sweden, Fraunhofer IAIS gains 8.8 million hours of computational capacity on the new high-performance computer MareNostrum 5 at the Barcelona Supercomputing Center: The goal of the EuroLingua-GPT project is to train large multilingual open-source language models.
The contingent approved via a EuroHPC “Extreme Scale Access” comprises 8.8 million GPU hours on H100 chips and has been available since May.
With the new computing capacities, small models in the range of 7 to 34 billion parameters and large models with up to 180 billion parameters can be trained from scratch.
The new EuroLingua models are based on a training dataset consisting of 45 European languages, dialects and codes, including the 24 official European languages. This gives a significant weight to European languages and values – multilingual large language models are still rare. Training will start at the end of May 2024 and the first joint models are expected to be published in the coming months.
Project leader Dr. Nicolas Flores-Herr, team leader Conversational AI at Fraunhofer IAIS says: “The goal of our collaboration with AI Sweden is to train a family of large language models from scratch that will be published open source.”
- Identifying (google)AI-generated images with SynthID - [Your Thoughts?]deepmind.google Identifying AI-generated images with SynthID
Today, in partnership with Google Cloud, we’re beta launching SynthID, a new tool for watermarking and identifying AI-generated images. It’s being released to a limited number of Vertex AI...
>Synth ID operates by embedding a digital watermark into AI-generated images, which serves as a hidden marker. This watermark is imperceptible to the human eye but can be detected using specialized tools. The tool is currently available exclusively to users of Google's AI image creator, Imogen.
https://youtu.be/9btDaOcfIMY?si=vK4bqDS3JOu5bl9H
https://deepmind.google/technologies/synthid/
https://www.toolify.ai/ai-news/googles-synth-id-a-gamechanging-tool-for-ai-image-identification-1243357
https://www.theverge.com/2023/8/29/23849107/synthid-google-deepmind-ai-image-detector
With SynthID, you should be able to edit a photo all you want without destroying the AI watermark. Image: Google
- OpenAI releases GPT-4o, a faster model that’s free for all ChatGPT userstechcrunch.com OpenAI debuts GPT-4o 'omni' model now powering ChatGPT | TechCrunch
During its "Spring Update" event, OpenAI debuted a number of upgrades and enhancements to its generative AI product portfolio.
- IBM open-sources its Granite AI models - and they mean businesswww.zdnet.com IBM open-sources its Granite AI models - and they mean business
Many companies claim to have open-sourced their LLMs, but IBM actually did it. Here's how.
- And it begins. OpenAI mulls NSFW AI model outputwww.theregister.com OpenAI model behavior guidance contemplates NSFW AI
That's a new twist on open, then
- Leaked Deck Reveals How OpenAI Is Pitching Publisher Partnershipswww.adweek.com Leaked Deck Reveals OpenAI's Pitch on Publisher Partnerships
OpenAI's Preferred Publisher Program offers media companies licensing deals.
- Axon plans to launch an AI based police report writing tool - thoughts?
From axons website: > The average officer spends 10 to 15 hours per week on report writing. Given a typical 40 hour workweek, that means up to 40% of time is spent on just paperwork. It’s time to rewrite report writing with Draft One. > > For agencies that struggle with low staffing levels and overwhelmed officers, Draft One force multiplies every officer by drafting report narratives in seconds, saving an hour or more per shift. Less time on reports ensures more time responding to calls for service, reducing crime and supporting communities. > > Join Axon Founder and CEO Rick Smith, Principal Product Manager Noah Spitzer-Williams, and some of our early access customers to learn how Draft One leverages the power of AI to draft police report narratives based on body-worn camera audio.
What do you guys think about this?
- Brain-Inspired Computer Approaches Brain-Like Sizespectrum.ieee.org Brain-Inspired Computer Approaches Brain-Like Size
Gestalt design yields flexibility and neuromorphic computational scale
- Burnout Is Pushing Workers to Use AI—Even if Their Boss Doesn’t Knowwww.wired.com Burnout Is Pushing Workers to Use AI—Even if Their Boss Doesn’t Know
People are using AI at work whether their bosses want them to or not, new data shows. But it’s desperation, not innovation, driving the change.
- What service should i use for AI backend that isn't openai
I have written myself a web fronetend using langchain that uses a json based agent to do advanced tool use (websearch, calculations, document summarisation/information extraction, etc) and chain of thought reasoning. At the moment im using openai gpt-4-turbo for the main agent and gpt-3.5-turbo for some of the summation tools but openai models are limited in capabilities (wont talk about certain topic to the point of disregarding required output json formatting). What service can use that will run a preferably foss model as powerfull as gpt-4 that is pay per use with a langchain integration?
here's the project https://github.com/muntedcrocodile/Sydney
EDIT: Added link to project
- What is RHEL AI? A guide to the open source way for doing AIwww.redhat.com What is RHEL AI? A guide to the open source way for doing AI
Dive into the product details of th developer preview of Red Hat Enterprise Linux AI, an open source generative AI product targeted at Large Language Model (LLM) developers and domain experts.
- Will AIs do drugs?
being able to alter our consciousness with drugs is something all life forms with a nervous system do- there must be an evolutionary reason for it. How will AI acheive this?
- StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generationstorydiffusion.github.io StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation
StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation
- Metas new plan. Paris Hilton bot, biological stupidity recreated with artificial intelligence. DISRUPTION!uk.pcmag.com Meta's Celebrity AI Chatbots Finally Launch for US Users
US users can now chat with the celeb chatbots on WhatsApp, Messenger, and Instagram.
- Microsoft bans US police departments from using enterprise AI tool for facial recognitiontechcrunch.com Microsoft bans US police departments from using enterprise AI tool for facial recognition | TechCrunch
Microsoft has changed its terms of service to ban certain police departments from using Azure OpenAI Service.
- Microsoft's FOMO after seeing Google AI drove investment in OpenAIwww.theregister.com Microsoft's FOMO on Google AI drove funding of OpenAI
Historical email in antitrust case shows execs 'very, very worried' about capability gaps
- Google Search results polluted by buggy AI-written code frustrate coderswww.theregister.com Developers seethe as Google surfaces buggy AI-written code
Pulumi claims it has culled bad infrastructure-as-code samples
- Are studios starting to produce AI films?
I just finished watching this German produced Netflix film and it was completely awful. If you are looking forward to this film, go watch it first if you want and skip my little review below.
spoiler
If you don't care and want to just take my word for it ... it is a waste of four one hour episodes, a total of four hours of content.
Everything inside the film and its marketing is an AI like production with all the sci-fi alien contact cliches and movie tropes from the past 20 years. A mysterious woman in space having an epiphany, hearing voices; a father with his daughter, a daughter with a disability, a father with a truckers hat, construction worker garb running in a cornfield, mysterious no-name SWAT team/soldiers, a nuclear explosion, helicopters, jet fighters, astronauts, the ISS, scientists, the desert, cityscapes, car crashes, car chase, secret agents, a plane crash, .... input all this into an AI program and ask it for a script and you'll end up with 'The Signal'
There is so little soul, sense or common logic in the film that to me it has all the hallmarks of a script and writing that was all generated by AI tools. If you look at the writers who put this together it is a group of four young writers with almost no prior experience in major film, yet they were handed the keys to a multi-million dollar production.
I like watching new films and I regularly take my chances at watching something new and different. This one was interesting but the more I watched it, the more I became suspicious that it was either written by a terrible dyslectic writer with little life experience ... or it was mashed together with an AI text generator.
One of the biggest giveaways that it most likely was an AI generated script was the corny voice over wrap up at the end of the film. I felt like I had just watched a four hour version of those Youtube auto generated AI fake film previews that are popular right now.
- Forget the AI doom and hype, let's make computers usefulwww.theregister.com Forget the AI doom and hype, let's make computers useful
Machine learning has its place, just not in ways that suits today's hypesters
- Llama 3 Establishes Meta as the Leader in “Open” AIspectrum.ieee.org Llama 3 Establishes Meta as the Leader in “Open” AI
Meta’s new AI model was trained on seven times as much data as its predecessor
- Reid Hoffman meets his AI twin - Full
cross-posted from: https://lemmy.dbzer0.com/post/19085113
> This guy built his own HAL 9000.
- Cops Using AI That Automatically Generates Police Reports From Body Cam Footagefuturism.com Cops Using AI That Automatically Generates Police Reports From Body Cam Footage
Taser maker Axon has announced a new product called "Draft One," an AI that can generate police reports from body cam audio.
- Ex-Amazon exec claims she was asked to ignore copyright law in race to AIwww.theregister.com Ex-Amazon AI exec claims she was asked to ignore IP law
High-flying AI scientist claims unfair dismissal following pregnancy leave
- Meta’s Llama 3 AI Is Smart, But Who Is Going To Profit From It?www.nextplatform.com Meta’s Llama 3 AI Is Smart, But Who Is Going To Profit From It?
With large language models, bigger is better (and faster) but better is also better. And one of the key insights that the Meta AI research team had with
- Netflix True Crime Producer Responds to AI Allegations in "What Jennifer Did" Documentaryfuturism.com Netflix True Crime Producer Responds to AI Allegations in "What Jennifer Did" Documentary
An executive producer of the Netflix documentary "What Jennifer Did" has weighed in on the purported use of AI.
- Elon Musk’s Grok keeps making up fake news based on X users’ jokesarstechnica.com Elon Musk’s Grok keeps making up fake news based on X users’ jokes
X likely hopes to avoid liability with disclaimer that Grok "can make mistakes."
- Too many modelstechcrunch.com Too many models | TechCrunch
How many AI models is too many? It depends on how you look at it, but 10 a week is probably a bit much. That's roughly how many we've seen roll out in the
- Meta's Llama 3 will force OpenAI and other AI giants to up their gamewww.itpro.com Meta's Llama 3 will force OpenAI and other AI giants to up their game
The new model pushes open source as a serious contender in the AI space, and proprietary models might soon find themselves playing catch-up
- Microsoft’s VASA-1 can deepfake a person with one photo and one audio trackarstechnica.com Microsoft’s VASA-1 can deepfake a person with one photo and one audio track
YouTube videos of 6K celebrities helped train AI model to animate photos in real time.
cross-posted from: https://sh.itjust.works/post/18066953
> On Tuesday, Microsoft Research Asia unveiled VASA-1, an AI model that can create a synchronized animated video of a person talking or singing from a single photo and an existing audio track. In the future, it could power virtual avatars that render locally and don't require video feeds—or allow anyone with similar tools to take a photo of a person found online and make them appear to say whatever they want.
- Old Movies Are Being Enhanced With AI Tools and Not Everyone is Happypetapixel.com Old Movies Are Being Enhanced With AI Tools and Not Everyone is Happy
What do you think? Better or worse.
- AI Detects Mysterious Detail Hidden in Famous Raphael Masterpiecewww.sciencealert.com AI Detects Mysterious Detail Hidden in Famous Raphael Masterpiece
Artificial intelligence (AI) can be trained to see details in images that escape the human eye.
- AI startup Stability lays off 10% of staff after controversial CEO's exit: Read the full memowww.cnbc.com AI startup Stability lays off 10% of staff after controversial CEO's exit: Read the full memo
Stability AI laid off several employees to "right-size" the business after a period of unsustainable growth, according to an internal memo obtained by CNBC.
