Weekly Link Roundup #68
There's an ebb and flow in the news...this week feels more like deep thinking..
MCP: An (Accidentally) Universal Plugin System: This is slightly nerdy but very interesting and the possibility space could be amazing. Read the article and then come back and tell me if I’m wrong.
Us, Augmented: Improve How We Work, Organize, and Innovate By Augmenting Intelligence, Individual and Collective, With the New AI: There are those of us who nip at the heels of important stories and then there are those who wade in and create a 200 page white paper that is required reading. Gianni Giacomelli is of course the latter. His intro to the white paper - “AI's moving fast. Organizations aren't. When the world shifts, we need not just new knowledge but also the right frameworks to make practical sense of what happens and what to do about it. I believe that many companies are still stuck in outdated organizational paradigms. That's been, for years, the belief at the core of MIT's Center for Collective Intelligence, and mine.” He also has a nice explanation of the work here, and check the comments for a link to the digital twin for this.
Cloudflare tests "pay-per-crawl" system to charge AI firms for scraping website content: This has the potential to be HUGE > > “Cloudflare is experimenting with a new way to prevent AI crawlers from scraping website content. The CDN/security company has announced that it will block them from accessing content without permission or compensation by default. Publishers can allow the crawlers, but the bots' AI firms will be charged.” > > If this takes hold, I wonder what the downstream cost/pricing issues will be for enterprises using LLMs.
IBM’s Dmitry Krotov wants to crack the ‘physics’ of memory: Look, when you read this excerpt, you’ll think - this is way too hard science’y for me to pay attention to. I get that but here’s the thing - the technical and the technological are moving so fast right now, that unless you gain some familiarity with the fundamentals, with the underpinnings of what’s happening, you’ll miss out on a critical vector for doing pattern recognition at the speed we need to be working at. So, read on - “On their own, atoms or neurons behave predictably. But in a network, their collective properties sometimes change as they interact, creating what physicists call emergent behavior. This unexpected shift has been observed in all kinds of neural networks as their amount of training data scales, giving them capabilities that go beyond their explicit training.”
What’s inside Genspark? A new vibe working approach that ditches rigid workflows for autonomous agents: How quickly we’ve moved from vibe coding to vibe working. Pity the life span of a meme these days. Still, it’s interesting - “Vibe coding has been all the rage in recent months as a simple way for anyone to build applications with generative AI. But what if that same easy-going, natural language approach was extended to other enterprise workflows? That’s the promise of an emerging category of agentic AI applications.”
AI avatars are here in full force—and they’re serving some of the world’s biggest companies: Here’s my thinking…you better tell me if I’m talking to an AI. That’s number one. Number two is that you can have the absolute coolest avatar in the world and it can talk and move just like a human but if the experience behind it still sucks, the avatar won’t matter.
Substack Is Having a Moment—Again. But Time Is Running Out: Maybe you can help me figure out which one I am > > “The truth is that the vast majority of the content on Substack is boring, amateurish or batshit crazy,” he wrote.” I’d be happy with batshit crazy. I do think the article has a good point though - that Substack reminds me of WeWork. WeWork was a commercial real estate company that wanted to believe it was a tech company (it wasn’t). And “Substack was a media company trying to be valued as a tech company, which is a familiar fail point for similar companies.” There are already companies like BeeHiv and Ghost that are Substack competitors -the defensible moat here is all network effect. The “subscription fatigue” is something I feel even as I publish. Same with streaming. We need some organizational innovation here.
Human Development Report 2025: A Matter of Choice: People and Possibilities in the Age of AI (United Nations Development Programme): “Going forward, development depends less on what AI can do—not on how human-like it is perceived to be—and more on mobilizing people’s imaginations to reshape economies and societies to make the most of it. Instead of trying vainly to predict what will happen, this year’s Human Development Report asks what choices can be made so that new development pathways for all countries dot the horizon, helping everyone have a shot at thriving in a world with AI.” Nice recap here from Andreas Horn.
What a blast! This AI built an Asteroids remake in five minutes and I still can't believe it: “Six months ago, I was gushing about ChatGPT's ability to build a simple Tic-Tac-Go game (like Tic-Tac-Toe, but with a lot more squares) based on a handful of prompts. That, it turns out, was child's play. Now, I've used five prompts to create a remarkable facsimile of the classic Asteroids game I played as a teen in the 1980s.” > > Still think L&D is somehow safe?
The Best 20th Century Chinese Fiction Books: If I’d only made a habit out of telling you to read scifi - oh wait, I did. Well now here’s a list from a non-English speaking original source. Gosh Mark - could there be new ideas in there?
Commodore acquired for a ‘low seven figure’ price — new (acting) CEO comes from the retro community: Can’t love this enough or root for them hard enough. My first PC (which I still have in its original box) was a Commodore VIC-20 (see below).
After revealing he was ordered to destroy his copy of Fallout's source code, OG lead Tim Cain says we're losing game history because companies "take authority but not responsibility" for preservation: This is a shame. A loss. Really.
I tried a tool from NotebookLM’s creators that makes a daily podcast out of my inbox and calendar, and it’s game-changing: As someone who rides a bus into work at least 3 days a week, this looks pretty useful. > > “Huxe is built around the concept of personalized intelligence — an AI that's essentially made just for you. It learns from your behavior and adapts dynamically over time. By connecting to sources like your calendar and email, Huxe generates a personalized feed of spoken content tailored to your day.”
There Are No New Ideas in AI… Only New Datasets: Loving the background - so important to go over this now before it gets lost to myth-making and change > > The kicker > > “But there’s a missing piece here: each of these four breakthroughs enabled us to learn from a new data source…”
The rise of the creative generalist: why being 'good at lots of things' is becoming a superpower: This is KEY > > “In a world where AI can handle the technical execution, human generalists who can think strategically across disciplines aren't becoming obsolete; they're becoming indispensable.” > > I love the tilt of the article, that generalists, especially in an age of AI, can be incredibly valuable, (read Range by Epstein) but I’ll renew my objection with the use of the term “creatives” to really only describe graphic artists, designers and the like. I don’t want to minimize their contributions at all but I think the term immediately sets boundaries - if you’re not a creative then by definition, you’re not creative. I want to find some term that recognizes the unique abilities of designers and artists but also allows the rest of us to be known for our creative approaches to work like learning, innovation or leadership.
This AI-powered social app aims to end loneliness—by ‘engineering chance’: I think this is a great effort and I could not wish this company any greater chance for success - I hope they crush it. Here’s where my mind goes though - onboarding. I know, how dull but hear my out. I think in general, companies do a poor job onboarding new folks, especially in the sense of them getting to know people outside of an intro 1:1. As WFH and remote work settles in as a default condition, I think its imperative on us to look for these chances to use whatever tools/apps we may have to help people feel some belonging to the team or org that they join. Who knows what connections you’ll find?
One of your better issues