Ross McLean’s Post

I'm officially 6 months into my deep dive on Generative AI! 🎓 In celebration, I give you... 🍾 "Easy Ways to Get Comfortable with Generative AI Activity # 2:" Use the Fathom note-taker service on a video call with a colleague. Fathom is a free service that you can invite to a Google Meet, Zoom or Teams video call. Fathom records the meeting, transcribes it and provides a short Generative AI-based summary of what was discussed in the meeting. https://fathom.video Here’s what you can learn from doing this: - Fathom uses AI-based transcriptions, which are now 90+% accurate, even better with a clear audio recording. You’ll get a clear sense of AI transcription capabilities. - Fathom creates a Gen AI summary of the session for you. Because it's a summary of a conversation that you've just participated in, you'll get a great sense of how Gen AI summarization's strengths and weaknesses. If you try this out, feel free to share your experience/thoughts as a comment or DM me. I'd genuinely love to hear. * If you use the free Fathom service, make sure your video call doesn't include anything confidential* Thanks Larisa Mats (Fathom is not a client of mine, and I am not a paid endorser)

Since I started doing a deep dive on Generative AI and taking about it on Linkedin, I've had a lot of people ask me how they can start getting comfortable with AI. So, I'm going to start sharing a few simple things people can do to start exploring and getting comfortable. Hope it helps. Feel free to share tricks that work for you. Get Comfortable with Generative AI Activity # 1: Start using https://www.perplexity.ai/ instead of whatever search engine you currently use (note: I don't work for Perplexity and they are not a client). Perplexity is a Gen AI-driven search engine that gives you AI-generated answers to your searches rather than just giving you a list of links like Google. Here's what I learned from doing this: 1. Perplexity does a good job of checking multiple sources and clearly shows the source of the materials it quotes in the answers it gives you including a direct link to the source. This helps me see how providing great sourcing lets me be confident that the answer I'm being given by Generative AI is based on reputable data and isn't a hallucination, which is a big deal in the Generative AI space. 2. "Answer engines" like Perplexity give answers. Most search engines just give a list of links (mostly paid/sponsored) that I have to sort through to answer the question myself. Answer engines like Perplexity are going to seriously disrupt search companies like Google and interacting with Perplexity makes it clear why.

This post by Roger Jackson and the comments are like a ring-side seat to the debate/slugfest that's going on in the market research community about synthetic data usage. Grab some popcorn and enjoy!

4 Reasons it’s Crucial that Qualitative Researchers Pay Attention to GenerativeAI: 1. GenAI’s effect on qualitative research will be profound. Up until now, technology has struggled to provide qualitative researchers with useful analysis tools. GenerativeAI will change that. One of qualitative research’s greatest challenges is taking the massive amounts of unstructured data we collect and actually having enough time to have a smart human parse it for what’s important and insightful. One of the things that GenAI is fundamentally good at is helping extract meaning from massive amounts of unstructured data. The more I learn about GenAI and the more I experiment with it in the qual research space, the more confident I become that GenAI will change qual profoundly. 2. GenAI will allow qualitative research to retain/regain relevance by allowing it to deliver more powerful insight faster. This will ultimately benefit the qual research business by bringing qualitative back into an increasingly high-paced innovation and product development cycle from which it is currently being dropped or excluded because it takes too long. 3. Qual researchers have a huge amount to gain by having a basic understanding of GenAI. We need to know what it can and can’t do, and how to apply it. AI can’t replace a human who’s great at qualitative, but it can make it much faster for you to deliver better work to your clients. 4. If quallies get on board and contribute, we’ll get powerful Generative AI tools. If not, tech will do it. Poorly. If the development of AI for qualitative research is influenced by quallies, we will end up with powerful AI tools that truly help us do our jobs. If we delay/resist, technology experts will build GenAI for qualitative without truly understanding what qual researchers actually need from it.

If you present data as part of your job, this is worth a quick read.

Great thought on LLMs from Kevin Hartman.

Great perspective on Generative AI and LLMs from Ben Evans. Excellent weekend long-read, dense with insight. "We’ve had ChatGPT for 18 months, but what’s it for? What are the use-cases? Why isn’t it useful for everyone, right now? Do Large Language Models become universal tools that can do ‘any’ task, or do we wrap them in single-purpose apps, and build thousands of new companies around that?" https://lnkd.in/gEc9MH8J

Want to dive into generative AI? A great place to start is to learn Prompt Engineering so you can get better, more reliable, hallucination-free responses from Large Language Models. I have just completed this course and I highly recommend it. It's currently free on Coursera.