Weekly Link Roundup #15
That's up one from last week!
RTO doesn’t improve company value, but does make employees miserable: Study: Wanted make sure that this one didn’t get lost in the mix. “Overall, the analysis, released as a pre-print, found that RTO mandates did not improve a firm's financial metrics, but they did decrease employee satisfaction.” I think this is key but also key is the idea that the “workforce” as some kind of monolith, doesn’t exist. That the “office” can still serve a role and it should serve different roles to different internal demographics but that requires consistent, hard work at looking how to extend the positive attributes of corporate culture virtually and how to make time in the office meaningful. See also - The executive hubris driving five-day in-office mandates.
SambaNova debuts 1 trillion parameter Composition of Experts model for enterprise gen AI: This another story that points to something I’ve been saying - that in this AI-enabled age, people from all levels in the org are going to have get smarter about AI from a technical and technological sense. The example here is the difference between Composition of Experts (refers to keeping each expert model separately trained on its own secure dataset. The security restrictions of the training data propagate to the expert model) and Mixture of Experts (“implies a single expert model is trained on multiple datasets. This can allow data from one dataset to potentially leak into the model, violating the security and privacy of the other datasets”). SambaNova is also making a hardware play here which makes them doubly interesting.
Generative AI's next act: Autonomous agents: Yes AND good example of what I mean about the current and emerging wave of AI and AI-enabled products and services will require leaders throughout an org to be more tech savvy than they’ve ever been before.
Klarna says its AI assistant does the work of 700 people after it laid off 700 people: Well this certainly made a splash when it hit. 1st grain of salt - this company is headed toward an IPO and this is the kind of news, that can drive up the offer price or valuation of a company headed into that kind of event. 2nd grain of salt - the CEO says the fact that the company laid off that many people last year (and did a horrible job at it), isn’t linked to the productivity target. So yeah - not the beginning of mass layoffs (I don’t think) but it does put a point on my idea that we all need to thinking about how we drive value to the org, over and above the typical activities that are on our To Do lists.
The Invisible $1.52 Trillion Problem: Clunky Old Software: Want a target for AI? here’s a 1.5 trillion dollar one. Make no mistake, there’s a real Charybdis/Scylla path to navigate here. Releasing AI into coders’ hands or even wider into employees’ hands with no/low code UI on top and the problem could get much worse, quickly. Just thinking about Marc Ramos’ exhortation that L&D could be the new R&D and how this is one case in which L&D could partner with SDEs/SDMs to raise the bar across the organization on how to prevent growing tech debt.
Your Organization Isn’t Designed to Work with GenAI: The key paragraph “The introduction of the internet, mobile computing, and cloud platforms showed us that extracting full value from groundbreaking technologies lies not in merely integrating them into existing business processes, but in completely reimagining those processes. While GenAI may prove even more transformative than these innovations, it similarly demands redesigning the way work gets done to maximize its potential.” Remember when people used to print out their emails?
This story is one of the reasons why I love scifi. It explores, it pushes, it asks great and terrible questions. Give it a read. > > "MMAcevedo (Mnemonic Map/Acevedo), also known as Miguel, is the earliest executable image of a human brain. It is a snapshot of the living brain of neurology graduate Miguel Acevedo Álvarez (2010–2073), taken by researchers at the Uplift Laboratory at the University of New Mexico on August 1, 2031."
The Intellectual Obesity Crisis: I really like the thinking here around focusing more on what we consume with our brains. “We evolved to crave sugar because it was a scarce source of energy. But when we learned to produce it on an industrial scale, suddenly our love for sweet things became a liability. The same is now true of data. In an age of information overabundance, our curiosity, which once focused us, now distracts us. And it’s led to an epidemic of intellectual obesity that’s clogging our minds with malignant junk.”
The Knowledge Economy Is Over. Welcome to the Allocation Economy: Think about this…there will no more ICs - everyone will be a manager, you just might be managing AI models and/or agents. “Even junior employees will be expected to use AI, which will force them into the role of manager—model manager. Instead of managing humans, they’ll be allocating work to AI models and making sure the work gets done well. They’ll need many of the same skills as human managers of today do (though in slightly modified form).”
H2O AI releases Danube, a super-tiny LLM for mobile applications: When I echo Ethan Mollick and say that the AI you’re using today is literally the worst AI you will ever use, this is what I mean. This is a signal - it’s telling you that you need to be thinking about what can be done when full fledged AI models are running locally on everyone’s devices > “Named after the second-largest river in Europe, the open-source model comes with 1.8 billion parameters and is said to match or outperform similarly sized models across a range of natural language tasks. This puts it in the same category as strong offerings from Microsoft, Stability AI and Eleuther AI.” Add this to the mix: Google DeepMind’s new generative model makes Super Mario–like games from scratch: Here’s a key point “But the Google DeepMind researchers are interested in more than just game generation. The team behind Genie works on open-ended learning, where AI-controlled bots are dropped into a virtual environment and left to solve various tasks by trial and error (a technique known as reinforcement learning).”
Meet the winners of the 2024 Dance Your PhD Contest: I can’t love this enough. Here is exactly all you need to know before you go and watch: “This year's winner is Weliton Menário Costa of the Australian National University for his thesis "Personality, Social Environment, and Maternal-level Effects: Insights from a Wild Kangaroo Population." His video entry, "Kangaroo Time," is having a bit of a viral moment, charming viewers with its catchy beat and colorful, quirky mix of dance styles and personalities—both human and kangaroo.”
Here Come the AI Worms: Security researchers created an AI worm in a test environment that can automatically spread between generative AI agents—potentially stealing data and sending spam emails along the way: Did you really think that you can hide a virus inside an Excel file or a PDF but it wouldn’t happen with AI? Have you read no William Gibson? Even the name of the attack vector sounds like its from a story: “adversarial self-replicating prompt.”
Google brings Stack Overflow’s knowledge base to Gemini for Google Cloud: This is a classic example of a ‘be careful what you wish for moment.’ On the one hand, Stack Overflow is the goto place for devs to get answers to tricky coding problems and making that knowledge base more accessible is good, the other hand says look back at the tech debt story and the one about worms and realize that this could also help bad actions occur faster. Finally, on our third hand, remember that this is a great example of ‘better tools help build better tools faster.’ This is what I mean: Welcome to the Era of BadGPTs: And keep in mind that while we guard against large-scale bad stuff, it’s also on us to keep an eye for the smaller, well-intentioned acts that can create, oh, I don’t know, let’s say $1.5 trillion worth of tech debt.
Why does AI have to be nice? Researchers propose ‘Antagonistic AI’: As a huge fan of sarcasm, I support this development. What if “AI systems that are purposefully combative, critical, rude and even interrupt users mid-thought” can be found to help build “resilience; provide catharsis and entertainment; promote personal or collective growth; facilitate self-reflection and enlightenment; strengthen and diversify ideas; and foster social bonding.” Bring on the snarky LLMs!
WordPress says VIP customer data won’t be shared to AI firms without consent: Here’s your newest upsell…if you’re a premium customer, we won’t use your data to train an LLM.
From Eliza to ChatGPT: why people spent 60 years building chatbots: “What Eliza showed, and what other developers and engineers have spent the next six decades working on, is that we treat our devices differently when we think of them as animate, human-like objects. And we are remarkably willing to treat our devices that way. (Have you ever felt bad for your robot vacuum as it bonks its way around your living room, or thanked Alexa for doing something for you?)”
Jamix emerges with $3M pre-seed and model-agnostic enterprise AI assistant: I’m going to class this another signal - a ping from the emerging future if you will - that while this one company may not be the eventual winner (it might be), what’s important is that more and more companies will spring up in this intermediation space and orgs will have be increasingly sophisticated and clear about how they evaluate potential partners and what problems they are looking to solve.
Writer unveils Palmyra-Vision, a multimodal AI to reimagine enterprise workflows: Now this one looks much more robust than the prior story. It also kind of makes the point that people aren’t buying drills, they’re buying holes and what people want now is an AI product that can solve problems vs just adding AI to the mix. “According to Habib, the model can extract insights from images, classify objects, interpret charts and graphs, answer specific questions, and generate text descriptions. On several benchmark tests, Palmyra-Vision achieved state-of-the-art accuracy, outperforming other lauded multimodal models like OpenAI’s GPT-4V and Google’s Gemini 1.0 Ultra.”
AI for MBAs? One Harvard Business School lecturer is giving it a shot: Wonder how this makes TA’s, who’ve been a source of unpaid or underpaid labor in academia for decades feel? “The 250 Harvard Business School students who signed up for the popular “Launching Tech Ventures” class last fall got the usual lectures filled with case studies of startups and visits from successful entrepreneurs. But they also had access to a specially trained digital assistant to help with their coursework, one powered by the same artificial intelligence technology behind ChatGPT. Students could ask the bot, dubbed ChatLTV, to help with analysis of a case study, find definitions of unfamiliar terms, or even see when professors were holding office hours.”
And just because I love writing notes and thoughts in books and because this poem is amazing, here ya go - enjoy
Sometimes the notes are ferocious, skirmishes against the author raging along the borders of every page in tiny black script. If I could just get my hands on you, Kierkegaard, or Conor Cruise O’Brien, they seem to say, I would bolt the door and beat some logic into your head.
Other comments are more offhand, dismissive – “Nonsense.” “Please!” “HA!!” – that kind of thing. I remember once looking up from my reading, my thumb as a bookmark, trying to imagine what the person must look like who wrote “Don’t be a ninny” alongside a paragraph in The Life of Emily Dickinson.
Students are more modest needing to leave only their splayed footprints along the shore of the page. One scrawls “Metaphor” next to a stanza of Eliot’s. Another notes the presence of “Irony” fifty times outside the paragraphs of A Modest Proposal.
Or they are fans who cheer from the empty bleachers, hands cupped around their mouths. “Absolutely,” they shout to Duns Scotus and James Baldwin. “Yes.” “Bull’s-eye.” “My man!” Check marks, asterisks, and exclamation points rain down along the sidelines.
And if you have managed to graduate from college without ever having written “Man vs. Nature” in a margin, perhaps now is the time to take one step forward. We have all seized the white perimeter as our own and reached for a pen if only to show we did not just laze in an armchair turning pages; we pressed a thought into the wayside, planted an impression along the verge.
Even Irish monks in their cold scriptoria jotted along the borders of the Gospels brief asides about the pains of copying, a bird singing near their window, or the sunlight that illuminated their page–anonymous men catching a ride into the future on a vessel more lasting than themselves. And you have not read Joshua Reynolds, they say, until you have read him enwreathed with Blake’s furious scribbling.
Yet the one I think of most often, the one that dangles from me like a locket, was written in the copy of Catcher in the Rye I borrowed from the local library one slow, hot summer. I was just beginning high school then, reading books on a davenport in my parents’ living room, and I cannot tell you how vastly my loneliness was deepened, how poignant and amplified the world before me seemed, when I found on one page
a few greasy looking smears and next to them, written in soft pencil– by a beautiful girl, I could tell, whom I would never meet– “Pardon the egg salad stains, but I’m in love.”