Coté

Home Depot builds DIY GenAI model that tells human employees what to say - Better search/knowledge management: ‘Home Depot’s team built an LLM with RAG capabilities that answers queries by retrieving knowledge documents and processing them alongside data from previous chat history to “generate response suggestions for customer service agents."'

Embracing AI To Augment, Not Replace, The Status Quo - Ideating AI apps: ‘1. Be on the lookout for processes that don’t scale well by adding more resources, are somehow constrained by human involvement, or could be improved if the human actor had more information. 2. Once you identify the process ask yourself either “How can I improve this process?” or “What do people hate most about this process?” 3. With that information, start to triage the type of AI that could help you. If you need more scale or more specific data, start with machine learning or vision apps. If you need more summarized data or connected information, start with generative AI.'

Oscar Health’s early observations on OpenAI o1-preview - Experiments with using AI in healthcare, specially the chain of reasoning stuff in o1-preview.

To Improve Your Mean Time to Recovery, Start at the Beginning - How to think about securing your cloud native apps, and apps in general.

The impact of AI on the workforce: Tasks versus jobs? - “While many businesses use AI to replace worker tasks, there is little evidence that AI use is associated with a decline in firm employment.” // I’m starting to think that what’s going on here is that we haven’t found the right applications (uses of) for generative AI. Keep exploring!

Did a digital obsession ‘Just Do It’ in for Nike’s John Donahoe? - Getting direct versus channel (partners, “middlemen”) right is difficult. It’s especially risky when you dramatically change it (move from selling direct to selling through middle-people, or the other way around).

The Camera is the Filter - “Digital cameras are, themselves, a filter—each one reflects the technology and aesthetics of its time. A photo taken on an early 2000s digital camera, for instance, captures not only the subject but also the era’s unique imperfections—grainy textures, soft focus, lower resolutions, and muted colors. These qualities are part of the photo’s authenticity, a snapshot of history shaped by the camera’s limitations. Adding filters that emulate an old type of film erases that context, replacing the original look with a trendy (at the time), artificial aesthetic that distorts the photograph’s integrity. To me, that defied what I thought the purpose of the job was, which was to document these factories in that particular moment in time."

Six technology myths and the Big Tech-lash - “So why do people think that it’s getting faster? Well, one of the reasons is they sort of conflate applications with platforms. So sure, ChatGPT gets 100 million people in a couple of months. But why is that? Well, they didn’t have to buy anything. I didn’t have to install anything. I could just go use it. And to me, that’s no different than saying that the Milton Berle show on TV went from no users to 50 million users in a year because all you had to do was turn on the station. And so the applications, whether it’s Facebook or Google or ChatGPT, it’s much easier to put an application on top of a radio program, a television or a computer or the internet than to buy a new platform.” And: “Data isn’t [the] new oil. I wish it were true. If only the world revolved around computers and data and information and connectivity and networks, life would be better. But the reality is that energy is still the main force in the global economy."