Chris Donahoe’s Post

I've been playing around with the current generation of generative AI and I have to say that notebook.lm is absolutely incredible. I uploaded a single page from "Charlie and The Chocolate factory" and asked the AI to create a podcast about it. Check out the human-like quality of the voices, their pauses and the interjections they both make. It is just mesmerising. This truly is next level AI. Is it amazing for the future or the beginning of the end for human content creation?

Instructions are only as good as the reader's contextual understanding 😅🤦 This also goes for AI - if we assume the AI model has no real concept of the world, this instructs our use of it and the detail level we need to adopt in our use of it - and subsequently the cases we can use it for 🤔 This is the inherent limits to agentic AI - navigating a world requires understanding that world. If these models do not on their own understand restlessly the world around them, we need to ensure they are only exposed to what we have defined their context to cover💡 If not, we will end up with no smoking and way too much safety - and who’d want that! 😅 #ai #technology #llm #context 🤝 #Connect or #Follow for a blend of serious insights and lighthearted humor about everything IT and business 🎧 #Listen to The Only Constant podcast

Curious about how AI can truly transform your business strategy? In my latest podcast episode, I discuss why AI is more about innovating your approach than mastering prompt engineering. We explore: - How to shift your mindset to leverage AI effectively - Why innovation is key in AI adoption - Ways AI can boost your business results beyond just prompts If you’re ready to rethink how you use AI, this episode will guide you in applying it strategically. In today’s fast-evolving world, AI is a tool for innovation. Make sure you’re using it to push your business forward. How will you innovate your approach with AI today? #AIinBusiness #Innovation #DigitalTransformation #AI #GenerativeAI

I believe the #creative industry is ready for #AI integration. 2024 was for being scared. 2025 will be for leaning in. In an upcoming episode of the Virtually Everything! #podcast, we predicted that creative teams would begin creating their own LLMs and even ponder agentic infrastructure. It's already being discussed with many of the clients I've worked with from a creative standpoint. And the most significant focus for many of those companies right now is GOVERNANCE. I know. It sounds sexy. But you can't give AI this immense power over your content unless you give it guardrails. To build out a successful AI Constitution, here's what your clients are working on: • How will their AI model respond to complex and evolving AI policies and requirements within the company? • How will it detect algorithmic fairness and bias? • Ensuring ethical and transparent AI practice. • Developing a flexible and agile technology focused on adaptability and future regulation. • Participating in industry collaborations. My advice? Do some research on AI Constitutions, familiarize yourself with ethical AI practices, and look for the content creation tools that are built through those practices. Then, add it to your creative service's benefits when pitching to clients. So, show your clients you can meet them where they are at with a good understanding of constitutional AI practices. Let me know what you think. Do you agree? Disagree? This is just the start of a very large conversation.

I keep hearing "AI can't do X". ...and then when I try X, it works like a charm. Here's an example: In a recent episode of the Lex Fridman podcast, Edward Gibson of MIT talks about how LLMs get tripped up on a variant of the Monty Hall problem (link in comments). I tried his example with GPT-4, and it got the correct answer and provided the exact reasoning. In the same podcast, he also claimed that LLMs struggle to complete sentences that are heavily nested, for example: "The boy, who the cat, which the dog..." Now, GPT-3.5 provides "chased.", which is wrong because you need to have 3 verbs. One for each of the subjects. GPT-4 answers correctly and coherently: "The boy, who the cat, which the dog had chased, adored, finally came home to his caring family. Interestingly, GPT-3.5 thinks the boy adores the cat in the above sentence, but GPT-4 rightly understands it's the cat who adores the boy. What's the point? There's a few. 1. Don't be so sure that AI can't do something. 2. Don't be so sure that AI can't do something it couldn't do 6 months ago. 3. We need a robust way of measuring these capabilities and we need that measure to be audited by independent people.

Evolution from Toys to Tools to Transformation The evolution in the use of AI technology has shifted from toys to tools and now to a transformative phase. Despite facing challenges such as technology limitations, governance issues, and ethics, the clear vision of AI's potential is driving thought leaders to explore the possibilities of training models, using generative AI, and achieving the ultimate North Star formula. This evolution has opened up new opportunities for reaching higher levels of achievement through practical applications of AI technology. Snipped from The WARC Podcast - Mapping AI's impact on marketing teams. Via AI-powered podcast app Snipd. https://lnkd.in/d3ft9CXP

Okay LinkedIn fam. Wild?🤯 Crazy? Or, Cool? Last week, I shared an article on how AI will transform businesses in the next 12 months, backed by insights from my LinkedIn poll and the Social Media Examiner's Generative AI Industry report. At the end, I teased something a bit unconventional (for now) - taking that article, running it through a text-to-speech AI tool and turning it into a two-person 🎙️podcast—lickety-split! Well, guess what? I did just that. The result? An AI-generated conversation. https://lnkd.in/gNjYhRBy Imagine listening to two AI personalities discuss my article and the impact of AI on businesses—all created in minutes. This is the future of content. Effortless production, immediate value and a whole lot of fun... Or, Is it just too crazy to imagine AI taking over our podcasts too? You be the judge. Tune in to hear this AI-powered experiment in action. It will blow your mind. https://lnkd.in/gNjYhRBy I'm curious to hear your thoughts P.S. Watch for my next experiment where I will clone my voice for the podcast. Wanna put an AI strategy in place for 2025 to save time and get more done with your marketing? DM me.

AI at Work: The expectation gap The idea that AI is revolutionising how we work feels almost… ironic. It’s not that it’s untrue it’s just that the reality on the ground can be, well, underwhelming. In many organisations, risk, compliance, and security concerns mean employees are stuck with homegrown AI tools that can’t hold a candle to the big beasts like Chat GPT or the proliferation of niche apps that are popping up like weeds. Apps that can turn your leader’s nuggets of wisdom into podcasts? It exists. But can you use it? Or ones that can metamorph your organisation’s vision into a 30 second animated movie that Pixar might be envious of. Nope. And don’t get me started on Otter.ai (I got into awful hot water around that one) The tools with the potential to make a real difference are often the ones we can’t use 😩 And yet, in our personal lives, where we have all the freedom we want to experiment, we don’t have such a burning need. Want to know how to turn those three rotten ingredients into a meal? Yeah. Ish So is there a frustrating gap between AI’s promise and its practical application where we actually need it: at work? #AI

Using AI in Marketing: Discover the importance of using AI in marketing ethically and responsibly. Learn about the principles of legality, accountability, transparency, objectivity, and neutrality in AI marketing. Avoid harmful practices and false information. that's Naully Nicolas on the Humane Marketing podcast this week: https://lnkd.in/eZ6sbRyv #AIinMarketing #EthicalAI #MarketingPrinciples #TransparencyMatters #AccountabilityInMarketing #LegalAI #ObjectivityInMarketing #AIResponsibility #AIinBusiness #MarketingEthics

Keeping up with the rapid advancements in AI can feel like a full-time job! Here's how I stay informed about AI headlines, politics, advancements, and the latest tools:   🎧 The AI Daily Brief Podcast: Every day, I tune into the AI Breakdown podcast. In just 15 minutes, I get a concise summary of the day's top AI headlines, the main story, and insights from political, financial, and geopolitical perspectives. It's my go-to for staying updated efficiently.   📚 Learning with Besuper.ai: When I want to explore new AI tools and quickly see how they work I go to besuper.ai, a fantastic learning platform with hundreds of tutorials on various AI tools. They’re constantly adding new content, making it easy to stay on the cutting edge of AI technology. ($20/month subscription)   For anyone looking to stay ahead in the AI field, I highly recommend these resources. They’ve been game-changers for me! How do you stay up to date with AI? I'd love to know! #AI #AIBreakdown #AIAdvancements #Learning #BesuperAI #TechInnovation