pinboarded

12669 bookmarks
Custom sorting
How I Use "AI"
How I Use "AI"
Most of the people online I find who talk about LLM utility are either wildly optimistic, and claim all jobs will be automated within three years, or wildly pessimistic, and say they have contributed nothing and never will. So in this post, I just want to try and ground the conversation. I'm not going to make any arguments about what the future holds. I just want to provide a list of 50 conversations that I (a programmer and research scientist studying machine learning) have had with different large language models to meaningfully improve my ability to perform research and help me work on random coding side projects.
·nicholas.carlini.com·
How I Use "AI"
Open Voice
Open Voice
We introduce OpenVoice, a versatile instant voice cloning approach that requires only a short audio clip from the reference speaker to replicate their voice and generate speech in multiple languages. OpenVoice enables granular control over voice styles, including emotion, accent, rhythm, pauses, and intonation, in addition to replicating the tone color of the reference speaker. OpenVoice also achieves zero-shot cross-lingual voice cloning for languages not included in the massive-speaker training set. OpenVoice is also computationally efficient, costing tens of times less than commercially available APIs that offer even inferior performance. The technical report and source code can be found at https://arxiv.org/pdf/2312.01479.pdf and https://github.com/myshell-ai/OpenVoice
·research.myshell.ai·
Open Voice
AI Comes for Music
AI Comes for Music
As the record labels sue AI companies for generating derivative songs, let us ask: What makes a song original and human anyway?
·newsletter.dancohen.org·
AI Comes for Music
Generative AI is a climate disaster
Generative AI is a climate disaster
It can be easy to get lost in numbers, but what they point to is an obvious truth: generative AI is an environmental disaster that’s accelerating natural destruction and the climate crisis at the very moment alarms are sounding about the precious little time that remains to turn things around. Tech companies once pitched themselves as the purveyors of a more ethical form of capitalism. They wanted us to believe they would balance corporate profit with environmental sustainability, such that the digital future was marketed as inherently more sustainable than the analog past. It’s clearer than ever that was a lie.
·disconnect.blog·
Generative AI is a climate disaster
The AI Copyright Trap : Carys J. Craig : Free Download, Borrow, and Streaming : Internet Archive
The AI Copyright Trap : Carys J. Craig : Free Download, Borrow, and Streaming : Internet Archive
As AI tools proliferate, policy makers are increasingly being called upon to protect creators and the cultural industries from the extractive, exploitative, and even existential threats posed by generative AI. In their haste to act, however, they risk running headlong into the Copyright Trap: the mistaken conviction that copyright law is the best tool to support human creators and culture in our new technological reality (when in fact it is likely to do more harm than good). It is a trap in the sense that it may satisfy the wants of a small group of powerful stakeholders, but it will harm the interests of the more vulnerable actors who are, perhaps, most drawn to it. Once entered, it will also prove practically impossible to escape. I identify three routes in to the copyright trap in current AI debates: first is the “if value, then (property) right” fallacy; second is the idea that unauthorized copying is inherently wrongful; and third is the resurrection of the starving artist trope to justify copyright’s expansion. Ultimately, this article urges AI critics to sidestep the copyright trap, resisting the lure of its proprietary logic in favor of more appropriate routes towards addressing the risks and harms of generative AI.
·archive.org·
The AI Copyright Trap : Carys J. Craig : Free Download, Borrow, and Streaming : Internet Archive
Every Google &udm=? in the world
Every Google &udm=? in the world
Appends udm=14 to the Google search URL to instruct Google to only return the Web results. There is even an extension and website built specifically for this use case. This is suitable for users who would like to have clean search results without the clutter from Knowledge Graph, Local Results, Related Questions, etc. Now, there are others udm=x than just 14. I have been doing some searching and compiled all of them here. In fact, you probably have been using some of the common udm. For example, udm=2 is Images.
·serpapi.com·
Every Google &udm=? in the world
Canadian Privacy Library – Open privacy knowledge for all
Canadian Privacy Library – Open privacy knowledge for all
The Canadian Privacy Library is the not-for-profit home to a collection of Privacy Impact Assessments and Open Educational Resources. They are also preserved in the Internet Archive. The majority of this collection consists of Canadian Privacy Impact Assessments (PIAs) from public post-secondary institutions in British Columbia. BC’s Freedom of Information and Protection of Privacy Act (FIPPA) requires public bodies to conduct PIAs, which contain rich knowledge about how technologies work, what privacy risks they carry, and how these risks are being mitigated. Freedom of information requests under FIPPA have been filed to build the collection. Everything is public record. PIAs protect privacy and keep people safe. They are in the public interest. They should be open for education, research, and criticism. This library’s goal is to collect PIAs from every public post-secondary education institution in Canada and make them freely accessible online.
·privacylibrary.ca·
Canadian Privacy Library – Open privacy knowledge for all
Journal of Open Source Education
Journal of Open Source Education
The Journal of Open Source Education is an educator friendly journal for publishing open-source educational materials and software. JOSE, pronounced [hoe-zay], is a sibling journal to the Journal of Open Source Software (JOSS), which publishes open-source research software. JOSE relies on the journal management infrastructure and tools developed for JOSS. Currently, academia lacks a mechanism for crediting efforts to develop software for assisting teaching and learning or open-source educational content and materials. As a result, beyond personal motivation, there is little incentive to develop and share such material. The Journal of Open Source Education (JOSE) is a scholarly journal with a formal peer review process designed to improve the quality of the software or content submitted. Upon acceptance into JOSE, a CrossRef DOI is minted and we list your paper on the JOSE website.
·jose.theoj.org·
Journal of Open Source Education
Academic Fracking: When Publishers Sell Scholars Work to AI
Academic Fracking: When Publishers Sell Scholars Work to AI
Last week, I explored AI and academic publishing in response to an article that came out a few weeks ago about a deal Taylor & Francis made to sell their books to Microsoft and one other AI company (unnamed) for a boatload of money. ... By publishers selling authors’ works to Microsoft AI and other AI companies, it made me think about why it felt familiar and left a sour taste in my mouth. Then, a phrase popped into my head the fit near perfect: “Academic fracking.”
·aiedusimplified.substack.com·
Academic Fracking: When Publishers Sell Scholars Work to AI
AI-Enabled Transformation of Information Objects Into Learning Objects - The Scholarly Kitchen
AI-Enabled Transformation of Information Objects Into Learning Objects - The Scholarly Kitchen
The past year has seen remarkable development in the AI-enabled services embedded in information tools across the scholarly communications industry. Looking at library-licensed content and tools, AI is powering a wide range of services for our users. Retrieval-augmented generation technologies are particularly prominent, powering experiments in search and discovery. Other applications assist readers by integrating information from multiple sources into a single synthesized text. Yet others support readers in investigating a single text, particularly longer documents, to gain understanding, perspectives, and insights. Finally, tools are emerging that leverage AI as a coach, scaffolding and supporting readers – particularly students – to foster growth in knowledge and skills as they engage with content.  I find these latter two – AI for insight and AI as coach – particularly intriguing. By transforming information objects into learning objects, these tools unlock the contents of articles and books and expand their reach beyond experts who can already relatively easily make sense of what they are reading. In this post, I’d like to share my experiences with and thoughts about these tools by exploring three examples: Alethea, Papers AI Assistant, and the JSTOR interactive research tool.
·scholarlykitchen.sspnet.org·
AI-Enabled Transformation of Information Objects Into Learning Objects - The Scholarly Kitchen
The Web We've (Never) Lost
The Web We've (Never) Lost
Doesn't it feel like the web is getting worse every day? Do you miss the days when Google wasn't a garbage factory, Twitter wasn't a cesspool of Nazis, and you weren't treated like a pair of eyeballs with a wallet? Don't worry, the web of yore is still alive and kicking – in fact, it's thriving. You just need to know where to look. Let's take a dive into the non-mainstream web, where people create blogs and websites for the sheer joy of it, algorithmic feeds are nowhere to be found – and you can join, too!
·bitoff.org·
The Web We've (Never) Lost
The Memex Method – Pluralistic: Daily links from Cory Doctorow
The Memex Method – Pluralistic: Daily links from Cory Doctorow
These repeated acts of public description adds each idea to a supersaturated, subconscious solution of fragmentary elements that have the potential to become something bigger. Every now and again, a few of these fragments will stick to each other and nucleate, crystallizing a substantial, synthetic analysis out of all of those bits and pieces I’ve salted into that solution of potential sources of inspiration. That’s how blogging is complimentary to other forms of more serious work: when you’ve done enough of it, you can get entire essays, speeches, stories, novels, spontaneously appearing in a state of near-completeness, ready to be written.
·pluralistic.net·
The Memex Method – Pluralistic: Daily links from Cory Doctorow
The Library Map of Vermont | Before Your Time
The Library Map of Vermont | Before Your Time
The “Library Map of Vermont” was created in 1914 to track all 225 brick and mortar libraries as well as 267 traveling library stations around the state. In this episode we’ll ask; Can a map truly show what it means for a community to have a well-supported library… and when communities lack that?
·beforeyourtime.org·
The Library Map of Vermont | Before Your Time
The Academics That Think ChatGPT Is BS - Better Offline - Omny.fm
The Academics That Think ChatGPT Is BS - Better Offline - Omny.fm
In a paper released earlier this year, three academics from the University of Glasgow classified ChatGPT's outputs not as "lies," but as "BS" - as defined by philosopher Harry G. Frankfurt in "on BS" (and yes I'm censoring that) - and created one of the most enjoyable and prescient papers ever written. In this episode, Ed Zitron is joined by academics Michael Townsen Hicks, James Humphries and Joe Slater in a free-wheeling conversation about ChatGPT's mediocrity - and how it's not built to represent the world at all.
·omny.fm·
The Academics That Think ChatGPT Is BS - Better Offline - Omny.fm
Federal Court asked to declare only humans can be authors under copyright law | CP24.com
Federal Court asked to declare only humans can be authors under copyright law | CP24.com
The Federal Court of Canada is being asked to declare that only humans — and not artificial intelligence — can be considered authors under Canada’s copyright law. It’s the first court case in the country testing how the Copyright Act treats artificially generated content, like the text, images and videos created by systems such as ChatGPT. David Fewer, director and general counsel at the University of Ottawa’s Samuelson-Glushko Canadian Internet Policy and Public Interest Clinic, says one of the aims of the clinic’s application is to lay "down in bedrock" that only humans are authors under the law. h/t @downes posting in OLDaily https://www.downes.ca/post/76813
·cp24.com·
Federal Court asked to declare only humans can be authors under copyright law | CP24.com
The AI + Open Education Initiative
The AI + Open Education Initiative
The mission of MIT Open Learning is to transform teaching and learning at MIT and around the globe through the innovative use of digital technologies. We fulfill our mission by supporting MIT faculty and students in bold experiments to enhance our residential education; promoting and enabling rigorous, empirical, interdisciplinary research on teaching and learning; providing platforms for digital education; sharing research and best practice by convening and partnering with schools, universities, companies, NGOs and governments; and extending MIT’s knowledge to the world.   The remarkable growth of Artificial Intelligence poses new benefits and new challenges to the field of open education. To meet the moment, MIT Open Learning is now soliciting rapid response papers and multimedia projects from stakeholders in the United States and internationally to articulate how generative AI might accelerate (or hinder) the promise of open education to offer engaging learning experiences. Our long-term vision is to bring together the energy and attention of MIT Open Learning, MIT faculty, leaders in open education, leading industry representatives, and leaders in philanthropy to more closely examine the challenges and opportunities that AI presents for open education. We welcome you to join us in identifying measurable next steps for understanding the role of Artificial Intelligence in the future of open education. h/t to @heatherb in OEG Connect for sharing the 2024 call for abstracts https://aiopeneducation.pubpub.org/2024call
·aiopeneducation.pubpub.org·
The AI + Open Education Initiative
Literary Hub
Literary Hub
Literary Hub is an organizing principle in the service of literary culture, a single, trusted, daily source for all the news, ideas and richness of contemporary literary life. There is more great literary content online than ever before, but it is scattered, easily lost—with the help of its editorial partners, Lit Hub is a site readers can rely on for smart, engaged, entertaining writing about all things books. Each day—alongside original content and exclusive excerpts—Literary Hub is proud to showcase an editorial feature from one of its many partners from across the literary spectrum: publishers big and small, journals, bookstores, and non-profits.
·lithub.com·
Literary Hub
Internet Phone Book
Internet Phone Book
We are creating a physical directory for exploring the vast poetic web. It features the personal websites of hundreds of designers, developers, writers, curators, and educators.
·internetphonebook.net·
Internet Phone Book
Creating Accessible Figures and Tables – DIS 2023
Creating Accessible Figures and Tables – DIS 2023
In this section, we have provided guidance for creating accessible figures and tables. By figures, we are referring to visual content such as charts, diagrams, images, designs, schematics and text blocks that may be embedded as images.  Since visual content including figures contributes to a paper or pictorial’s communication of knowledge, the visual design and alt text should also reflect that intellectual merit. So we have developed guidance for creating accessible content to support authors and reviewers. 
·dis.acm.org·
Creating Accessible Figures and Tables – DIS 2023
HTTP Keanu
HTTP Keanu
A fun way to learn about HTTP status codes, inspired by http.cat! uses movie references to the epic catalog of Keanu Reeves. whoah
·httpkeanu.glitch.me·
HTTP Keanu
Spurious Scholar
Spurious Scholar
Because if p < 0.05, why not publish? This is a spoof of having an LLM generate fake research papers bast on pure random correlation. Step 1: Gather a bunch of data. Step 2: Dredge that data to find random correlations between variables. Step 3: Calculate the correlation coefficient, confidence interval, and p-value to see if the connection is statistically significant. Step 4: If it is, have a large language model draft a research paper. Step 5: Remind everyone that these papers are AI-generated and are not real. Seriously, just pick one and read the lit review section. Step 6: ...publish " The silliness of the papers is an artifact of me (1) having fun and (2) acknowledging that realistic-looking AI-generated noise is a real concern for academic research (peer reviews in particular). The papers could sound more realistic than they do, but I intentionally prompted the model to write papers that look real but sound silly. Also: every page says 'This paper is AI-generated" at the bottom and the first letters of the names of the authors always spell out C-H-A-T-G-P-T.'"
·tylervigen.com·
Spurious Scholar
Digital infrastructures for education: Openness and the common good | UNESCO
Digital infrastructures for education: Openness and the common good | UNESCO
"Major technology companies have catered to the education market for a very long time, offering hardware and software products schools and institutions could purchase and retain. But something significant has changed over the past decade. Many of these businesses have shape-shifted from software and hardware companies to media and advertising1 behemoths. After contributing to significantly refashion work and social life, these corporations, now known as big tech, began offering, promoting, and selling – not products – but subscriptions to cloud-based platforms and services, many of which are not inspired by the needs of teachers, students and administrators in educational institutions. Today big tech businesses are, in many cases de facto mediators to fundamental rights such as access to information and the right to education." h/t to article author @tamiel See also Mapping Surveillance Capitalism in Education https://educacaovigiada.org.br/
·unesco.org·
Digital infrastructures for education: Openness and the common good | UNESCO
Pivot to AI
Pivot to AI
It can't be that stupid, you must be prompting it wrong Pivot To AI is written by Amy Castor and David Gerard, who also collaborate writing about the foolishness of cryptocurrency and blockchains.
·pivot-to-ai.com·
Pivot to AI