Weekly Link Roundup #42

What can you say about #42? It is of course, the Answer to Life, the Universe, and Everything. It's also a Pronic and a Harshad number. Also 42 degrees is the angle that we see rainbows at.

Started writing an item for here that turned into a whole other post: Agents, and Architecture, and Access - oh my! This is one of the articles that kicked off my thinking: Why Agent Orchestration Is The New Enterprise Integration Backbone For The AI Era. The other article: From selling access to selling work (and what it means for you). In reading this, I’m thinking that the cost equation for deploying AI is much less complicated when you’re a B2C and just adjust the cost of your product or service. It gets more complicated when the AI solution will at best have an indirect relationship to revenue. Zoho turns to Nvidia NeMo to build proprietary LLMs.

Bluesky and enshittification: I never tire of Cory Doctorow just nailing it. “Any system where users can leave without pain is a system whose owners have high switching costs and whose users have none. An owner who makes a bad call – like removing the block function say, or opting every user into AI training – will lose a lot of users. Not just those users who price these downgrades highly enough that they outweigh the costs of leaving the service. If leaving the service is free, then tormenting your users in this way will visit in swift and devastating pain upon you.” And if you’re not familiar with “enshittification”…

A Practical Guide to Gaining Value From LLMs: Thank you Rama Ramakrishnan for writing a useful article that never veers into the realm of the “ULTIMATE CHEAT SHEET” category. Worth the read alone just for this image and its explanation.

Walt Disney forms business unit to coordinate use of AI, augmented reality: This is an interesting approach - “Bergman noted the unit will focus on fast-moving areas of technology, such as AI and mixed reality, which blends the physical and digital worlds. It will not centralize work on these projects, but rather, ensure the various projects around the company fit with its broader strategy.” I’ve got questions like, what authority will the team have to do course corrections inside other teams that it thinks aren’t aligned with the broader strategy? What mechanisms will they have in place to try to make sure that everyone is on the same page re the strategy? What will the 100 people actually do? I mean, I get the impetus and it sounds like it’d a great gig with a lot of visibility but also fraught with no direct tie to a business unit and if everyone aligns, then it could be out of work.

YC-Backed Edtech Startup Receives €2.5 Million Funding: I always want to watch startups like this. Not all will succeed but almost all will show us something different - “Turing College’s approach combines AI-powered personalized learning with human mentorship, creating a scalable solution for the growing demand for tech skills.”

I’m the CEO of an AI company, and this is the BS behind AI: Kind of love this because it focuses on the questions to ask. “The solution for combatting false information and BS in AI and life in general lies in thinking like a scientist.”

Microsoft’s new Magnetic-One system directs multiple AI agents to complete user tasks: Agentic architecture is going to be huge. “To this end, Microsoft researchers recently unveiled a new multi-agent infrastructure called Magnetic-One that allows a single AI model to power various helper agents that work together to complete complex, multi-step tasks in different scenarios. Microsoft calls Magnetic-One a generalist agentic system that can “fully realize the long-held vision of agentic systems that can enhance our productivity and transform our lives.”

From challenges to opportunities: navigating the human response to automated agents in the workplace: Managers will have to start thinking about what agents they want on their teams: “Our findings suggest that lower-efficiency AA might outperform higher-efficiency ones due to the constraining influence of trust on adoption rates. Additionally, we find that lower initial trust in AA could lead to increased usage in certain scenarios and that stronger emotional and social responses to the use of AA may foster greater trust but result in decreased AA utilisation.”

Treating AI Agents as personas: Too soon? “AI agents are emerging as a new class of users, independently navigating the interfaces we design and performing complex tasks on our behalf. This marks the beginning of a new era of interaction, the Agent-Computer Interaction, where user experience encompasses not only human-computer relationship, but also the experiences of the AI agents.”

Study: Growth of AI adoption slows among U.S. workers: “The survey queried 17,372 workers in Australia, Brazil, Canada, France, Germany, India, Italy, Japan, the Netherlands, Singapore, Spain, Sweden, Switzerland, the U.K. and the U.S., and took place between Aug. 2 and Aug 30. Slack also found that globally, nearly half of workers (48%) said they were uncomfortable telling their managers they use AI at work. Among the top reasons cited were a fear of being seen as lazy, cheating or incompetent. What they're saying: "Too much of the burden has been put on workers to figure out how to use AI," Slack senior VP of research and analytics Christina Janzer said in a statement.” I can’t LOVE this enough > > "To ensure adoption of the technology, it's important that leaders not only train workers, but encourage employees to talk about it and experiment with AI out in the open."

Golly - do you think that purpose-built AIs built on niche/custom models will be big? Dunno but here are three clues:

• Generative AI startup Writer raises $200M at a $1.9B valuation:

• DataRobot launches Enterprise AI Suite to bridge gap between AI development and business value:

• Microsoft-backed startup debuts task optimized enterprise AI models that run on CPUs:

OpenAI launches ChatGPT with Search, taking Google head-on: Forget Google, it’s taking on ad-supported search as a business model.

How virtual cows could help improve human-robot interactions: “A video game in which participants herded virtual cattle has furthered our understanding of how humans make decisions on movement and navigation, and it could help us not only interact more effectively with artificial intelligence, but even improve the way robots move in the future.”

Aging spacecraft starts up a radio transmitter it hasn’t used since 1981 from 15 billion miles away: Go bore someone else with how reliable your product or service is; Voyager 1 has set the highest bar.

Patronus AI launches world’s first self-serve API to stop AI hallucinations: Another layer to the AI tech stack? “Today, Patronus AI, a San Francisco startup that recently secured $17 million in Series A funding, launched what it calls the first self-serve platform to detect and prevent AI failures in real-time. Think of it as a sophisticated spell-checker for AI systems, catching errors before they reach users.”

The first wooden satellite launched into space: That’s a headline you don’t see every day.

The Alexa Skills revolution that wasn’t: I don’t know maybe its because the process to build, deploy, and access the skills was arcane and byzantine. And maybe because voice was the wrong UI.

LinkedIn launches its first AI agent to take on the role of job recruiters: Yes, please continue to work on behalf of *checks notes* recruiters and no job seekers. Also, so happy you spent time/money *checks notes again* making games.

Meta’s former hardware lead for Orion is joining OpenAI: I can only think of one reason you hire a renowned hardware engineer.