Weekly Link Roundup #37
https://vote.gov - unless you're dead. (also, Share and Subscribe ;-))
Soft skills dead—long live ‘skills’: LOVE this > > “Here’s a little thought exercise: Suppose Harvard University, the Carnegie Foundation, and the Stanford Research Center all got together to find out what exactly made a person successful at work. Suppose they discovered that we had two complementary skill sets. One accounted for about 15% of one’s ability to succeed at work, and the other accounted for the remaining 85%. Which skill set would be deemed more important?” The answer is that the skill set that accounts for 85% of someone’s ability to succeed at work is typically and derisively known as “soft skills.” Lots of useful thinking in this short piece but I love these two points: “Forget knowledge and look for experience” and “skip that degree; seek competence.” Both these are logical and sound and absolutely at odds with (especially American) notions of degrees as requirements and ageism.
How can we make the best possible use of large language models for a smarter, more inclusive society?: The write-up article linked to above is good but this quote is from the paper itself, and these are the questions that NEED to asked and researched - “Large language models, however, are transforming how information is aggregated, accessed and transmitted online. Here we focus on the unique opportunities and challenges this transformation poses for collective intelligence. We bring together interdisciplinary perspectives from industry and academia to identify potential benefits, risks, policy-relevant considerations and open research questions, culminating in a call for a closer examination of how large language models affect humans’ ability to collectively tackle complex problems.”
Meta has a major opportunity to win the AI hardware race: I ABSOLUTELY think glasses combined with AR and AI are the right combo to be winners. Regular RX glasses, augment your sight, they correct it. Sunglasses allow you the capability to see in conditions you normally would have trouble with. It makes sense to add inputs into that visual context. Certainly makes more sense than some dumb pin.
LinkedIn Is Training AI on User Data Before Updating Its Terms of Service: In case you missed this HUGE blunder from LinkedIn. As this issue goes to ‘print’ - I believe the ToS have been updated but what a boneheaded move on LI’s part. “Multiple LinkedIn users on Wednesday noticed a setting that showed LinkedIn was using user data to improve its generative AI. LinkedIn told 404 Media it will update its terms of service “shortly.”
On the Diagram of Thought: I’m going to keep saying it - someone in L&D, in every company that is considering an AI pilot or will be affected by AI (that’s just about everyone), could be taking articles like this one, proactively working with a SME and translating this into a module in an ever-expanding course on AI that L&D proactively builds for the whole org > > - “We introduce Diagram of Thought (DoT), a framework that models iterative reasoning in large language models (LLMs) as the construction of a directed acyclic graph (DAG) within a single model. Unlike traditional approaches that represent reasoning as linear chains or trees, DoT organizes propositions, critiques, refinements, and verifications into a cohesive DAG structure, allowing the model to explore complex reasoning pathways while maintaining logical consistency.” See also: Speculation Brims That OpenAI o1 Leverages Multiple Chain-Of-Thought And Meta-Reasoning.
Uniphore unveils X-Stream, a unified knowledge offering to build RAG apps 8x faster: So L&D wants a seat at the table, cool - can L&D explain to sr leadership why this company’s offering is important? No? Then no seat for you. If L&D is to be a differentiator, a competitive differentiator, then it needs to help the org learn what it needs to know and do that faster. The tech mentioned in the article is actually really important and could have a large impact on AI projects and their success or failure. Here’s a key point where L&D could help “Uniphore offers a usage-based pricing model, and customers typically see a 4x-6x return on investment in weeks from going live,” he noted.” Cool. What does usage-based pricing imply? What kind of usage does your L&D pilot that will use AI require? Answering these questions in advance could be the kind of value (not activity) that would raise the visibility of L&D.
AI can’t understand cultural nuance - This is how leaders can use the right prompts to mitigate risk: Please tell that people already knew this “For example, an LLM trained primarily in English data might suggest marketing strategies without cultural nuance that inadvertently offend cultural sensibilities in another culture. English makes up 48% of training data, and when combined with other European languages, this figure climbs to 86%” and wow, brilliance of insights like this > > “the best approach is for you and your teams to understand its cultural awareness capabilities so you can leverage AI effectively while recognizing its limitations. Researchers say the best approach is to understand AI’s cultural awareness capabilities, so you can leverage it effectively while recognizing its limitations.”
This is genius: A new graphic novel imagines conversations between Einstein and Kafka: Ordered > > “Part biography, part historical fiction, Krimstein playfully explores the possibilities, building, with footnotes, on a thorough archive of letters, diaries, and other research. The result, a thought-provoking work made up of comics suffused in a gentle mix of aquamarine watercolors, is equal parts joyful and ruminative.”
Fake AI “podcasters” are reviewing my book and it’s freaking me out: These kind of advances are why, while the “hype” may be dying down around AI, I don’t think the use or applications of AI are even close to being done. “I still wasn't prepared for how disarmingly compelling it would be to listen to Google's NotebookLM condense my recent book about Minesweeper into a tight, 12.5-minute, podcast-style conversation between two people who don't exist.”
AI’s New Role in NYC Schools? Chancellor Banks Teases Personalized Learning and College Counseling: Hey Mark - what kind of statements drive you crazy? Thanks for asking…here’s one: “AI can revolutionize how we function as a school system,” Banks told the audience of administrators, elected officials, and union leaders at Frank Sinatra School of the Arts High School in Queens as he outlined his plans for the nation’s largest school system. Still, Banks acknowledged that the Education Department has no concrete plans, timelines, or cost estimates for those AI projects.” AI won’t revolutionize the function of the school system unless we change its underlying financial/funding model, much like how AI won’t revolutionize HR unless we change the way people are accounted for. I hate these statements that ignore foundational issues - its like building a boat without thinking about the water.
New Science Fiction & Fantasy Releases for September 20, 2024: I’m going to put this in front of the next article because it seems like the dummies from the next article should maybe read some scifi.
10 Books to Read Now: Science & Technology Changing Our Lives: Then, after they read some scifi, here’s some actual science they could read as they re-think this insanely stupid idea in the next article.
The crypto bros who dream of crowdfunding a new country: I include this article reluctantly and only for mockery and ridicule. I try not to type swear words here but this most arrogant of morons is really making it hard. He’s a VC talking about nation-building. Instead of governments formed by the will of the governed, you get companies. Company Towns is thinking too small - think at SCALE - think Company Countries. What foundational, dystopian, scifi novel from Snowcrash to Neuromancer fails to start from a world in which companies rule? I don’t know the units of measurement to describe how stupid this idea is. “He calls his idea the “network state”: startup nations. Here’s how it would work: communities form – on the internet initially – around a set of shared interests or values. Then they acquire land, becoming physical “countries” with their own laws. These would exist alongside existing nation states, and eventually, replace them altogether.” > > Oh good, there’s a school for it now: A Mysterious School for the Network State Crowd Is Now in Session. “Aspiring Network School participants put down deposits of up to $2,000 without even knowing the Network School’s location. Srinivasan has still not disclosed it publicly, although social media posts and WIRED reporting indicate that it’s Forest City, Malaysia.”
SocialAI offers a Twitter-like diary where AI bots respond to your posts: OK, so it looks like this is a good week for incredibly stupid ideas. A “social” network populated only by you and by bots that you get to decide will either by sycophants or trolls has got to be one of the signs of the Apocalypse.
What Lies Beneath Canada’s Former Indigenous School Sites Fuels a Debate - Despite possible evidence of hundreds of graves at former schools for Indigenous children, challenges in making a clear conclusion have given rise to skeptics: “Ground-penetrating radar had found possible signs of 215 unmarked graves at a former residential school in British Columbia run by the Catholic Church that the government had once used to assimilate Indigenous children forcibly taken from their families. It was the first of some 80 former schools where indications of possible unmarked graves were discovered, and it produced a wave of sorrow and shock in a country that has long struggled with the legacy of its treatment of Indigenous people. Three years later, though, no remains have been exhumed and identified. Many communities are struggling with a difficult choice: Should the sites be left undisturbed and transformed into memorial grounds, or should exhumations be done to identify any victims and return their remains to their communities?”
Fascinating Lifespan Map Reveals Which Famous Figures Overlapped in Time: I love this “Another more obscure overlap reveals that J.R.R. Tolkien (1892–1973), author of The Lord of the Rings, coincided by one year with rapper Eminem (born 1972). And it’s fascinating to be reminded that explorer Christopher Columbus (1451–1506), polymath Leonardo da Vinci (1452–1519), and theologian Martin Luther (1483–1546) were all alive at the same time.”