Weekly Link Roundup #19

Whew....that was a lot

I asked DALL-E to create an image of humans and ai working together on a conveyor belt full of hyperlinks - there are several disturbing items in this pic - can you find them all?!

After AI beat them, professional Go players got better and more creative: This article gives me hope and makes me optimistic that there are already communities in the world who have used their interactions with #AI to increase their value distinct from their activity. “So moves were getting better, but we do not know how that was distributed in the population. Were the moves getting better because some players cheated by using the AI? No. They really were getting better. And it wasn’t simply that they imitated the AI, in a mechanical way. They got more creative, too. There was an uptick in historically novel moves and sequences…moves that deviated from what the AI would do also improved, and these “human moves” accounted for 60 percent of the improvement.”

But Why Is This Paper Important? (A GPT created by Ethan Mollick): OK. The link below goes to a #GPT created by Ethan Mollick and it's purpose is to provide summaries of academic papers and explain why that paper is important. It took me longer to type that sentence then it did for the GPT to summarize the paper linked below. Now to be clear, there are papers that I will read all the way through myself for lots of reasons but I also look at so much information that having this kind of tool is one of those utilities that just blows my mind. In some ways, it reminds of the grad school process of "gutting a book" to say prepare for comps - when you had to master such a wide swath of works that reading them end-to-end was just impossible. The Paper: FLOW MATCHING FOR GENERATIVE MODELING (yes, this is the kind of stuff I read) https://lnkd.in/dAK_i26C The Summary: “The paper introduces a novel approach to training Continuous Normalizing Flows (CNFs), a type of generative model, by proposing a method called Flow Matching (FM). This method enables training CNFs on a larger scale and with higher efficiency than previously possible. By employing FM, the authors overcome the challenges associated with training CNFs, such as long training times and the need for specialized training methods. Their approach not only facilitates faster and more robust training but also achieves state-of-the-art performance in generating high-quality samples, particularly demonstrated on the ImageNet dataset. This advancement opens new avenues for the application of CNFs in generative modeling, enhancing their practical utility in areas like image generation, where model efficiency and sample quality are paramount.”

What we’ve learned from the women behind the AI revolution: Great podcast > “So, TechCrunch’s Dominic-Madori Davis and Kyle Wiggers decided to talk to women working in AI to learn more about their work, how they got into the world of artificial intelligence, and more. The series has been running for some time now, so it was the perfect moment to get the pair onto the Equity podcast for a chat about the project.”

OpenAI prepares to fight for its life as legal troubles mount: When I say we don’t have a clue about what a settled and knowable environment for AI looks like (and we’ll probably never get one), this is part of what I mean. Tech. Legal. Business models. All in flux.

Strategic Foresight & Futures Studies Collection: Wensupu Yang (his LinkedIn profile) (his Substack) has built a page piled high with an amazing array of books, white papers, courses, and more - all focused on Foresight. A huge thank you for this.

AI will not revolutionize business management but it could make it worse: This is a great read but this last paragraph is KEY - “In short, it would be wrong to think that a tool as rational as AI, in an environment as irrational as an organization, will automatically increase efficiency the way managers hope it will. Above all, before thinking about integrating AI, managers need to ensure that their organization is not stupid (in terms of both processes and behavior).”

Symbolica raises $31 mln to develop AI systems to compete with OpenAI: This is a potential, key, foundational shift in this space » “Symbolica has created a framework that will enable it to develop alternatives to the “transformer" deep learning architecture, said Symbolica CEO George Morgan, who formerly worked at Tesla…”

Four Takeaways on the Race to Amass Data for A.I.: Good read since it hits on something that we need to think about more and that is considering the source of the data that whatever AI model we’re using has been trained on. Think of it as a nutritional label or one of side effect label from our meds like in the image below. See also: How Tech Giants Cut Corners to Harvest Data for A.I.

Google Books ingesting all the AI-generated rubbish: This is garbage in / garbage out…but at scale. “AI-generated books—at best useless mish-mashes of written works and at worst lethal nonsense—are flooding Amazon and other venues. Google Books is indexing it indiscriminately, reports 404 Media, and the junk is already having an impact on search.”

Big tech unites around AI workforce training for 100M workers: This is some cutting edge journalism - “The stakes are high. Nearly all IT workers and technology executives say AI initiatives will fail without skilled teams, according to a Pluralsight report.”

Poe introduces a price-per-message revenue model for AI bot creators: Here’s the part that amazes me; we will sit inside our corporate walls and be amazed at the creativity that people can show when you provide them with compelling incentives. What we fail to do however, is think that is some way that we could bring those dynamics in house. I mean really, what would it take to create internal creator economies? > “Bot creators now have a new way to make money with Poe, the Quora-owned AI chatbot platform. On Monday, the company introduced a revenue model that allows creators to set a per-message price for their bots so they can make money whenever a user messages them. The addition follows an October 2023 release of a revenue-sharing program that would give bot creators a cut of the earnings when their users subscribed to Poe’s premium product.”

MIT License text becomes viral “sad girl” piano ballad generated by AI: Crazy, intriguing, and potentially scary “We've come a long way since primitive AI music generators in 2022. Today, AI tools like Suno.ai allow any series of words to become song lyrics, including inside jokes (as you'll see below). On Wednesday, prompt engineer Riley Goodside tweeted an AI-generated song created with the prompt "sad girl with piano performs the text of the MIT License," and it began to circulate widely in the AI community online.”

How Photos Were Transmitted by Wire in 1937: The Innovative Technology of a Century Ago: Super cool 1937 documentary - now there’s a sentence you don’t say every day. Pieces like this always make me about what we’re doing today that the future will look back on and if their judgement will be kind.

This AI exoskeleton is designed to help you hike easier and faster: I have mixed feelings here. Great for more people to enjoy hiking but I can see some folks turning this into gear for speed running mountain trails. “The Dysns is leading the race in advanced AI walking assistance. A chain clasps belt made from 3D knitted material secures around the user's waist and legs. The device has a motor that offers up 83lbs of weight reduction. Using its built-in motion sensor the Dysns recognizes any posture at a thousand times per second.”

The Elder Scrolls Online dev says the Metaverse is sinking because it ignored 20 years of games doing the exact same thing: 'it's not new, and they should stop treating it like it's new': This. A thousand times this. “Frior, who rather sensibly believes that "virtual worlds" like the Metaverse need to be backed up by communities (rather than speculative, ugly virtual property going for thousands of real dollars), says that the sales pitch of the Metaverse really only worked in the sphere of big tech investments.”

Playboy image from 1972 gets ban from IEEE computer journals: C’mon man - it’s just for the articles.

Getty Images CEO Calls for Industry Standards Around AI: Standards will be needed to stabilize the market and let orgs’ General Counsels feel better about not getting involved in copyright lawsuits.

Caribbean nation of Aruba backs itself up to Internet Archive: The historian and anthropologist in me rejoices - “The Internet Archive is now home to the Aruba Collection, which hosts digitized versions of Aruba’s National Library, National Archives, and other institutions including an archaeology museum and the University of Aruba. The collection comprises 101,376 items so far—roughly one for each person who lives on the Island—including 40,000 documents, 60,000 images, and seven 3D objects.”

First languages of North America traced back to two very different language groups from Siberia: “Johanna Nichols, a linguist at the University of California, Berkeley, has used her pioneering work in the field of language history to learn more about language development in North America. She has found that it can be traced back to two language groups that originated in Siberia. Her paper is published in the American Journal of Biological Anthropology.”

Chechnya is banning music that's too fast or slow. These songs wouldn't make the cut: Not sure I have any words here - except it makes me think about Gibson’s quote - “The future is here, it’s just not evenly distributed.” Sometimes the uneven distribution is easier to see. “"Musical, vocal and choreographic" works will be limited to a tempo of 80 to 116 beats per minute (BPM) to "conform to the Chechen mentality and sense of rhythm," said Dadayev, according to the Russian state-run news agency TASS.”