I smiled when writing this title because even the idea of being “up to date” with AI feels almost impossible. So much is changing daily, hourly, or sometimes by the minute. Staying on top of everything is not a realistic aim.

So with that in mind, how do I approach learning about AI and evolving how I use it?Reflecting back, there are 5 approaches I’ve been following:

  1. Podcasts
    This is my go-to - I listen to AI podcasts daily.The default recommendation is the AI Daily Brief, it has been my partner in this journey for the last 18 months, learning so much from it.Sitting at 20-30 minutes a day, it’s a simple, digestible format and covers the core news and deep-dive topics.Honestly starting here is the best place in my humble opinion.Some episodes are more relevant than others, but I still listen to all of them. It gives me a clear view of what is going on, from the latest models and upcoming product changes through to the wider political and economic climate.Yes, that is important too.Supporting the podcast is a newly launched newsletter if reading is more your preference - it includes handy links too, making it easy to get information in the way you want it alongside exercises to try out guided learning.

    I also continue to be a fan of Lenny’s Podcast.While not strictly ‘AI’ specific, his content continues to evolve and inevitably pull in thoughts about AI.Listening to other professionals outline what they think and how they are approaching it is reassuring as none of us can predict how the future will develop and what the potential impacts will be on our roles and digital development overall.

    Those are my two go-to’s.There are others I dip in and out based on the topic.The a16z Show is another that has great episodes.My recommendation is find the ones that resonate, you learn something from and build from there.The AI Daily Brief was a single episode recommendation that has built to a daily habit since.
  2. Hands on learning
    The purpose of this website is exactly this: trying by doing.There’s a huge concept called the ‘Capability Overhang’, it’s the difference between what the models can do vs what they are being used for.The overhang is the gap.The best way to stay relevant and continue to stay on top is through trying the tools out yourself.This means not only trying new things in a model you defer to, it’s also trying how that produces different outcomes in other models - or simply trying new models entirely.

    While I tend to default to ChatGPT simply due to the paid-for subscription, I regularly use other models.Be that Lovable for creating websites to Copilot for work or Gemini for different tasks, learning what the models can do well and where to switch is fundamental.It helps build your knowledge and means you lose the emotional attachment to a single provider and instead focus on the outcomes and benefits each model provides.Podcasts fuel the areas and experiments I want to try, by giving me inspiration and guidance on what to try next and why.

    A good example is Gemini.I’ve been trying to 3D print some of my daughter’s drawings and wanted to add colour without changing the drawing itself.Numerous attempts in ChatGPT failed, every single one made the image ‘better’.Moving to Gemini, it one-shotted it and produced a nice colour version of my daughter’s exact drawing.

    Overall, seeing what the tools can do and how they behave will only make you better informed.Don’t just limit yourself to work use cases, personal ones also show how to use the tools differently.
  3. LinkedIn
    I’m sure I’m not alone in saying LinkedIn has become a pretty volatile place.Either AI is coming for all our jobs or simply can do everything.The truth, as is always the case, is somewhere in the middle.I take LinkedIn with a pinch of salt but do find there are some amazing people out there publishing incredibly helpful knowledge.Those are worth following and using it to bolster your knowledge while giving practical examples or templates you can apply.Ones I follow are below and I’m sure there are others:
  • Neil Hoyne - Chief Strategist at Google
  • Nathaniel Whittemore - Superintelligent
  • David Pereira - Product Leadership Coach & Advisor
  • Ethan Mollick - Associate Professor
  • Ant Murphy - Product Coach
  • Pawel Huryn - AI PM
  • Melissa Perri - Board member, author, CEO and many other things
  • Lenny Rachitsky - Lenny’s Podcast
  • John Cutler - Head of Product
  • Teresa Torres - Author, Speaker and Product Discovery Coach
  • Allie K. Miller - AI Startup Advisor, Board Member and many other things

My advice would be to find the voices you trust and treat their outputs as inputs, not answers.AI continues to evolve and the one thing that is clear, no-one has it figured out yet.

  1. Company blogs
    This one surprised me. I’ve found it genuinely helpful.Reading company announcement pages has helped shape where I get the information from.This isn’t just limited to the ones from OpenAI or Anthropic, it’s also the other companies e.g. Salesforce.For now I delve in and read them directly, one of my future tasks is building an agent to scour the key ones and pull back anything that’s changed within my own daily/weekly brief.

    What I like about this approach is seeing not just what companies are doing, but how they are choosing to position themselves publicly.As this is all figured out, seeing how changes are positioned by different companies helps me understand what is on their mind as I work out the right direction for my own role and area.
  2. YouTube
    This is a channel for a future use case to consolidate information using an agentic workflow.I’ve used YouTube in a more ad hoc way to learn specific tools, such as n8n or Claude Code.It has not yet become a daily habit because finding good quality content is almost a role in itself.What I have found is there is great content out there, especially when there are screen shares to show what is going on.I’ve yet to find a creator that captivates me in the same way the AI Daily Brief does for podcasting, that may also be a podcast is a lot easier to consume on the commute to work where YouTube requires dedication and focused time - which can be a lot harder to find.

This list will no doubt evolve and other sources be added.I expect you will be reading this and asking why X wasn’t included or Y.My aim here is to show how I’ve found a way to continue to learn, in a way that works for me and the time available.If I had to boil it down, my two go-to’s would be a) podcast as this fits around my available time and it does an amazing job of taking the core developments and breaking it into manageable chunks for me to take away and try and b) hands-on learning.This is the reason for this blog, to show what I’m trying.AI is evolving so fast, the only way to see what’s possible is to find out for yourself.

What I will be looking to try next

While I do have my two favourite methods, there are others I’m keen to try:

  1. Twitter/X research agent
    The aim here is to build an agent to source the key voices in AI and bring me a daily report on what is being covered, good and bad.
  2. YouTube transcript agent
    The second is also a use case for an agentic use case - taking key and authoritative voices from YouTube, analysing the transcripts and producing a key report.This will require a detailed workflow being planned out to determine what the right tools are to use.NotebookLM is one I can see helping out massively here.

The above is by no means exhaustive.It will also change as I delve more into building out more agents and giving them core tasks to do.There will be other ways to keep learning as the AI landscape continues to change. I’m keen to try them all.