I wrote the post and then I played about with creating the podcast.
I’d heard of Google’s Notebook Language Model so I thought I’d give it a try.
In short, the tool can be used to create a podcast with two speakers to talk through whatever content you give it.
Something like ChatGPT, a large language model, has consumed a big bite out of the internet so it’ll bring that knowledge with it. However, Notebook LM works only on what you’ve brought it. So, if you have a long report or a series of links on a subject you can chuck them in and ask them to summarise for you.
Seeing as though I’d just blogged about the coverage of Liam Payne’s death on Black Country Live on Facebook I thought I’d chuck that into the tool.
And for wider content I thought I’d also add a link to the UK Press Gazette coverage of Facebook moving away from news as a source it shares in people’s timelines.
Firstly, Notbeook LM provided a text summary of what it thought of the two sources. And then you get the chance for it to provide a summary as a podcast. After waiting two or three minutes, this was produced.
You can listen to the results here.
If you are not happy with the results or the emphasis, yes, you can go back and change that in the original text setting.
The results? Well, there are the ahs, and ums you would expect from a podcast as two presenters talk through the key points. It’s ;like liustening to two people who have read the material and are talking it through. I also love how they’ve added some value of their own. There is criticism in the original piece of Black Country Live’s coverage and the hosts come up with their own ideas for how the title could report on the issue.
With that in mind, is this a useful way of summarising two documents for you? It can be. The American accident is a clear impediment to a British audience but if you can work through this it may be a useful way for you to consume something on the go.
The added value the hosts give can also be useful so long as you know where the joins are.
At the moment, Google’s Notebook LM is at a pilot stage with free access. You can find it here.