AI Stuff
- The AI wars: Microsoft unveils Phi-3, a capable AI model that fits easily on a phonewww.notebookcheck.net The AI wars: Microsoft unveils Phi-3, a capable AI model that fits easily on a phone
In response to Meta's recent Llama-3 release, Microsoft has published findings on the latest iteration of its lightweight AI model. The technical report shows the Phi-3-mini outperforming LLMs such as GPT-3.5 despite being a fraction of their size.
- Stable Diffusion Turbo XL can generate AI images as fast as you can typearstechnica.com Stable Diffusion Turbo XL can generate AI images as fast as you can type
Even at home, SDXL Turbo can create detailed images with startling speed.
- UnitedHealth uses AI model with 90% error rate to deny care, lawsuit allegesarstechnica.com UnitedHealth uses AI model with 90% error rate to deny care, lawsuit alleges
For the largest health insurer in the US, AI's error rate is like a feature, not a bug.
- Teen boys use AI to make fake nudes of classmates, sparking police probearstechnica.com Teen boys use AI to make fake nudes of classmates, sparking police probe
Parents told the high school "believed" the deepfake nudes were deleted.
- Brave rolls out privacy-focused AI chatbot to lure you away from ChatGPTwww.androidauthority.com Brave rolls out privacy-focused AI chatbot to lure you away from ChatGPT
Brave announced its starting to roll out a privacy-focused, browser-native AI chatbot. It will have a free and premium version.
IMO the only private LLM is one being run locally. Thoughts?
- Biden AI executive order directs agencies to develop safety guidelines - The Vergewww.theverge.com Biden releases AI executive order directing agencies to develop safety guidelines
The executive order comes after a series of non-binding agreements with AI companies.
- Forbes now has its own AI search engine - The Vergewww.theverge.com Forbes now has its own AI search engine
Adelaide will help answer readers’ burning questions.
- University of Chicago researchers seek to “poison” AI art generators with Nightshadearstechnica.com 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.
- Google, Microsoft, Anthropic and Open AI name first Executive Director of Frontier Model Forumblog.google Anthropic, Google, Microsoft and OpenAI announce Executive Director of the Frontier Model Forum and over $10 million for a new AI Safety Fund
Google, Microsoft, OpenAI and Anthropic jointly share an update on the Frontier Model Forum.
- Harvard professor Lawrence Lessig on why AI and TikTok are causing a free speech crisis for the internet - The Vergewww.theverge.com Harvard professor Lawrence Lessig on why AI and social media are causing a free speech crisis for the internet
The internet policy legend has surprising thoughts on how to balance free speech and democracy.
- Chinese AI-related firm Zhipu says it has raised over $341.8 mln this yearwww.reuters.com Chinese AI-related firm Zhipu says it has raised over $341.8 mln this year
Chinese artificial intelligence-related firm Beijing Zhipu Huazhang Technology said on Friday it has raised over 2.5 billion yuan ($341.82 million) this year.
- IBM has made a new, highly efficient AI processorarstechnica.com IBM has made a new, highly efficient AI processor
A very unusual chip design provides a distinct approach to cutting AI's energy cost.
- Universal Music sues AI company Anthropic for distributing song lyrics - The Vergewww.theverge.com Universal Music sues AI company Anthropic for distributing song lyrics
Claude 2 can release song lyrics without permission, said music publishers.
- YouTube working on AI tool to help you sing like music artists9to5google.com YouTube is working on AI that lets you sing like famous music artists
A new tool from YouTube may allow creators to use AI to sound like their favorite music artists when singing.
- At TED AI 2023, experts debate whether we’ve created “the new electricity”arstechnica.com At TED AI 2023, experts debate whether we’ve created “the new electricity”
Is AI going to replace us all, or is it just humanity's newest tool?
- Rapper Pras’ lawyer used AI to defend him in criminal case—it did not go wellarstechnica.com 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.
- The world’s biggest AI models aren’t very transparent, Stanford study says - The Vergewww.theverge.com The world’s biggest AI models aren’t very transparent, Stanford study says
Stanford’s new Foundation Model Transparency Index found developers don’t disclose societal impact at all.
- A Chatbot Encouraged Him to Kill the Queen. It’s Just the Beginningwww.wired.com A Chatbot Encouraged Him to Kill the Queen. It’s Just the Beginning
Companies are designing AI to appear increasingly human. That can mislead users—or worse.
- Large Language Models up to 4x Faster on RTX With TensorRT-LLM for Windowsblogs.nvidia.com Large Language Models up to 4x Faster on RTX With TensorRT-LLM for Windows | NVIDIA Blog
Generative AI on PC is getting up to 4x faster via TensorRT-LLM for Windows, an open-source library that accelerates inference performance.
- Nvidia’s H800 AI chip for China is blocked by new export rules - The Vergewww.theverge.com Nvidia’s H800 AI chip for China is blocked by new export rules
Nvidia won’t be able to export its H800 chip.
- Clearview AI and the end of privacy, with author Kashmir Hill - The Vergewww.theverge.com Clearview AI and the end of privacy, with author Kashmir Hill
Journalist Kashmir Hill’s new book, Your Face Belongs to Us, is all about Clearview AI and the threat to privacy.
- AI Chatbots Can Guess Your Personal Information From What You Typewww.wired.com AI Chatbots Can Guess Your Personal Information From What You Type
The AI models behind chatbots like ChatGPT can accurately guess a user's personal information from innocuous chats. Researchers say the troubling ability could be used by scammers or to target ads.
- Microsoft-affiliated research finds flaws in GPT-4techcrunch.com Microsoft-affiliated research finds flaws in GTP-4 | TechCrunch
A Microsoft-affiliated study has found flaws in GPT-4 that enable malicious actors to prompt the model to generate biased, toxic text.
- AI helps decipher first text of “unreadable” ancient Herculaneum scrollarstechnica.com AI helps decipher first text of “unreadable” ancient Herculaneum scroll
Computer science student Luke Farritor won $40,000 from the Vesuvius Challenge.
- Exclusive: US tackles loopholes in curbs on AI chip exports to Chinawww.reuters.com Exclusive: US tackles loopholes in curbs on AI chip exports to China
The steps are intended to prevent American chipmakers from selling products to China that circumvent government restrictions as part of the administration's actions to block more AI chip exports.
- The U.S. is updating rules to prevent American chipmakers from selling AI chips to China that skirt existing restrictions.
- These updated rules will expand the scope of restricted chips and demand reporting for certain shipments.
- The move aims to prevent U.S. technology from strengthening China's military but risks complicating U.S.-China diplomatic efforts.
- Consumer chips are exempt, but the new rules aim to close loopholes and evolve with technology advancements.
- The Linux Foundation Launches New Event: AI.dev: Open Source GenAI & ML Summitwww.prnewswire.com The Linux Foundation Launches New Event: AI.dev: Open Source GenAI & ML Summit
/PRNewswire/ -- The Linux Foundation, the nonprofit organization enabling mass innovation through open source, and LF AI + Data Foundation, the organization...
- Google Bard readies ‘Memory’ to adapt to important details about you9to5google.com Google Bard readies ‘Memory’ to adapt to important details about you
Each of us has preferences or requirements that need to be respected, and soon Google’s Bard will be able to...
cross-posted from !google@lemdro.id
> - Google's Bard is developing a "Memory" feature that allows the chatbot to remember and adapt to user-specific details and preferences. > - Currently, each conversation with Bard starts afresh, requiring users to restate particulars like dietary restrictions. > - The Memory feature will improve personalized interactions, such as recommending recipes based on a user's noted vegetarianism or travel suggestions considering the number of kids a user has. > - Users will be able to manage their stored preferences via a Memory page and can delete incorrect or unwanted memories. > - A toggle will allow users to easily disable the Memory function for privacy concerns or to have non-personalized conversations.
- Amazon to invest as much as $4 billion in AI startup Anthropictechcrunch.com Amazon to invest as much as $4 billion in AI startup Anthropic | TechCrunch
Amazon has agreed to invest up to $4 billion in the AI startup Anthropic, the two firms said, as the e-commerce group steps up its rivalry against In Amazon, Anthropic finds a deeply-pocketed strategic investor that can also provide compute power and help sell the offerings to cloud customers.
- Open Interpreter: An Interesting AI Tool to Locally Run ChatGPT-Like Code Interpreterbeebom.com Open Interpreter: An Interesting AI Tool to Locally Run ChatGPT-Like Code Interpreter
Open Interpreter is an open-source implementation of ChatGPT Code Interpreter that lets you achieve many tasks locally with ease.
- Alibaba launches its ChatGPT-like AI model for public use amid loosening restrictions in Chinacointelegraph.com Alibaba launches its ChatGPT-like AI model for public use amid loosening restrictions in China
Alibaba’s proprietary large language model, Tongyi Qianwen, will be available for public and enterprise access starting Sept. 13.
- Why Japan is building its own version of ChatGPTwww.nature.com Why Japan is building its own version of ChatGPT
Some Japanese researchers feel that AI systems trained on foreign languages cannot grasp the intricacies of Japanese language and culture.
- Join us for OpenAI’s first developer conference on November 6 in San Francisco - OpenAI Blogopenai.com Join us for OpenAI’s first developer conference on November 6 in San Francisco
Developer registration for in-person attendance will open in the coming weeks and developers everywhere will be able to livestream the keynote.
- Microsoft to defend customers on AI copyright challenges | Reuterswww.reuters.com Microsoft to defend customers on AI copyright challenges
Microsoft will pay legal damages on behalf of customers using its artificial intelligence (AI) products if they are sued for copyright infringement for the output generated by such systems, the company said on Thursday.
cross-posted from !technology@beehaw.org
- Douglas Lenat, Who Tried to Make A.I. More Human, Dies at 72 - NY Timeswww.nytimes.com Douglas Lenat, Who Tried to Make A.I. More Human, Dies at 72
He spent decades working on artificial intelligence, striving to create computers that could replicate common sense.
cross-posted from: https://infosec.pub/post/2221265 >NYT gift article expires in 30 days. > > https://ghostarchive.org/archive/FAewq
- Introducing Code Llama, a state-of-the-art large language model for codingai.meta.com Introducing Code Llama, a state-of-the-art large language model for coding
Code Llama, which is built on top of Llama 2, is free for research and commercial use.
- Microsoft’s AI recommends the Ottawa Food Bank as a tourist destinationwww.theverge.com Microsoft’s AI recommends the Ottawa Food Bank as a tourist destination
The company let go of human journalists to replace them with AI in 2020.
