Weekly Link Roundup
Seriously...how much can happen in one week?
Do me a favor and don’t skip this first story » OpenAI's Whisper is another case study in Colonisation: This is SUCH an important point in the whole discussion around AI. On one level it reminds me of the difference in how cultures regarded the ability of humans to “land” that enabled the Dutch to “buy” Manhattan Island from the Algonquin. The very idea of land ownership didn’t translate. We’ve already become used to asking what data did the LLM train on - this is similar but could wreck an entire language and culture. Translations are context-dependent and AI has the ability to strip away or misplace context at an unheard of scale. » “A few of our data scientists tried Whisper on te reo Māori videos from YouTube. Their initial reaction was, "Wow it works!" A more critical assessment of Whisper by our Māori data experts saw that it sort of worked but it was terrible. Still, this is concerning, that a non-Māori organisation thought it was okay to create a Māori speech recognition model and open it to the public.”
The Unique History of Japanese Plastic Food Samples: This story has nothing to do with #learninganddevelopment or #innovation but I just love it.
Sora is ChatGPT maker OpenAI's new text-to-video generator. Here's what we know about the new tool: So yes, this was big news this week - and yes, it looks straight amazing. It doesn’t change where we need to be thinking though. This capability already existed although not with this quality output and as Ethan Mollick reminds us, the AI we are using today is the worst AI we will ever use. So we should have known or anticipated that this level of capability would exist. As Dr. Phillipa Hardman points out in this article though, we can either use this new tech to scale what is the old way or we can do the hard work of thinking about a new system that incorporates this tech.
Nvidia’s Chat with RTX is a promising AI chatbot that runs locally on your PC: Sheesh. Did this happen this week too? How long until new hires get their shiny corporate laptop that includes a locally running AI that’s trained on all your corporate data?
This Company Says Conversational AI Will Kill Apps and Websites: Sierra hopes every company will use its empathetic AI chatbots for customer interactions: This quote >> "In the future, a company’s AI agent—basically the AI version of that company—will be just as important as their website,” says Taylor. “It's going to completely change the way companies exist digitally.” » What if instead of building courses, you build the L&D AI Agent. The one agent that the whole company will interact with to learn what they need to know?
Try this 5-point plan to successfully launch the world’s most important technology at your company: Solid article but I want to really point to #3 and #4 which are ones that I tend to talk about all the time. #3 is “Business professionals must upskill on a semi-technical understanding.” AMEN. This is a huge opportunity for #LearningAndDevelopment teams to be seen as educators vs trainers. Senior leaders are going to need and look for trusted advisors within the enterprise who can get them up to speed on the level of technical understanding about AI that they’ll need to assess related business decisions. #4 is “If you aren’t measuring value, you’re not pursuing value.” I said it before and I’ll keep saying it, AI is coming for your To Do list and if you’re not thinking about how to move up the value chain, it’ll come for you too.
Dutch startup secures $25M to bring autonomous bricklaying robots to Europe: No deep insights here, just another signal about tech penetrating all sectors.
True Source unveils AI LLM service based on The Wheel of Time: What’s that? You don’t know the story of Rand al’Thor and the Wheel of Time? Huge series of fantasy books (14 of them!) Also a series on Amazon. Anyway, what’s interesting about this is twofold - #1 I think its a Signal on how content properties will build connection and community with fans/consumers/etc and #2 I think its an example of a model for how companies could construct #knowledgemanagement systems. Enjoy the trip to Tar Valon!
In defense of busywork: I get what the argument is here - that having some kind of productive busywork can help refocus and refresh people for other tasks. I get that. I also get that I keep saying that AI is coming for exactly that kind of work. I think what it will come down to, much like RTO commands, is the good orgs are the ones who will think consciously about which work should go and what should stay.
Largest text-to-speech AI model yet shows ’emergent abilities’: “Researchers at Amazon have trained the largest ever text-to-speech model yet, which they claim exhibits “emergent” qualities improving its ability to speak even complex sentences naturally. The breakthrough could be what the technology needs to escape the uncanny valley.” I love this sentence: “For reasons unknown to us, once LLMs grow past a certain point, they start being way more robust and versatile, able to perform tasks they weren’t trained to.” They do make big uncertain jumps but don’t get that confused with they’re all of a sudden going to go all Skynet on us.
Rasa, an enterprise-focused dev platform for conversational GenAI, raises $30M: Important point here - “it claims to have developed the infrastructure to give developers at large enterprises the ability to build “robust” generative conversational AI assistants so that those interactions feel more personal and meaningful to users. It says it does this by providing the infrastructure CALM (Conversational AI with Language Models) and a low-code user interface.” I keep saying that between AI and low/no code, there’s going to be an explosion of employee-built software inside enterprises that we are woefully unprepared for in a number of ways.
Reddit has reportedly signed over its content to train AI models: So a site that literally runs on nothing but user-generated content, has now monetized that content instead of the community (selling ads). Really adds new meaning to the phrase that if you’re not paying for it, YOU are the product.
Air Canada must honor refund policy invented by airline’s chatbot: Can’t love this enough. If you deploy tech, you are responsible for it.
Animoca Brands’ Yat Siu sees NFTs as a way to protect ownership in the age of AI: I’m actually thrilled to see this. As someone who was involved in large-scale innovation programs and part of our design concern with the program was thinking about the motivation for why someone would submit their great idea…we talked about using NFTs as a way to look at fractional ownership of ideas. This extends that thinking in our now AI-enabled world. So in a low/no code environment with democratized access to content creation tools, maybe this is a way to open up new, internal economies for employees.
Tech companies go dark about AI advances. That’s a problem for innovation: “If we don’t find a way to fund large-scale, public compute, we are going to be in a really bad time, because we will have a really narrow set of actors with a bunch of insights that are hard to publish,” he told Semafor. “It’s a real issue, but we can mitigate this problem.”
How We Sort the World: Gregory Murphy on the Psychology of Categories: How we categorize things, concepts, people; “the glue that holds our mental world together,” is one of the great foundational facets of our world that largely goes unexamined.
Y Combinator puts out a new call for startups in areas like AI, spatial computing, climate tech and more: 9 new categories - Brick and Mortar 2.0, Carbon Removal Technologies, Cellular Agriculture and Clean Meat, Cleaner Commodities, Improving Memory, Longevity and Anti-Aging (YC Bio), Safeguards Against Fake Video, Supporting Creators, and Voice Apps. »Always an interesting signal.
The text file that runs the internet: “For three decades, a tiny text file has kept the internet from chaos. This text file has no particular legal or technical authority, and it’s not even particularly complicated. It represents a handshake deal between some of the earliest pioneers of the internet to respect each other’s wishes and build the internet in a way that benefitted everybody. It’s a mini constitution for the internet, written in code.”
Low-code and no-code development gets a makeover as priorities shift to AI: “It's probably too soon for inexperienced developers to work directly with generative AI. But once AI is embedded into low-code platforms, the technology could serve as a valuable assistant.” »Cool right? Now go ask your CIO and your CISO how they feel? This could be a HUGE unlock or a huge new attack surface or just a huge drain on resources.