Mahmoud GadGuest
Paul TurnerHost
Jan BebbingtonHost
In just three years, you can see how much Mahmood's work has moved on and how the use of AI has allowed him to build upon what he already had and take it to another level and get so much more use out of it.

And it sounds like that's both the models moving on, but then the approaches to using the models as well.

And I really liked his analogy of an engine and it's what you've built around it that makes it robust.

2023, it was producing so many essentially mis-truths and fictions, and they've worked on it over three years.

And if you imagine with the AI that we're all using in our lives, how much generalities there is there, whereas this is on a very narrow subject and it's taken them three years of concentrated work on this narrow area to get it to a stage where those fabrications are no longer there and everything's being produced to that great level of accuracy.

That demonstrates the great task there must be with AI on the global scale of making sure that it's not making up anything.

And I think it's that differentiation between like a general, you know, ask and see what you get and actually having it under these guardrails and supervising it very closely, which I thought was amazing.

I really liked some of the findings they had at the end of this process as well.

And one of the concepts I found really engaging is the idea of a collapsing funnel.

Because I've seen collapsing funnels in all sorts of other areas of work as well.

So looking at Tim Lamont's work that looked at the restoration practices of companies and what they say about it, that had this collapsing process going on as well.

But likewise, looking at biodiversity disclosure by ocean companies, by indicators disclosure around the TNFD.

So collapsing funnels is something I think, you know, for a select number of our listeners will go, oh, yeah, that's going to use that phrase because it captures it.

So if you are unable to disclose this particular piece of information, either because you don't know what's going on or you decide not to say anything about it, then everything else sort of wanes off.

Some of the numbers he was giving out there where you've got from 67% of firms who are doing the reporting bit and doing it okay down to there's only 2% I think was the last number he gave where he'd gone through various things such as whistleblowing and stuff like this and then mediation and then satisfactory mediation was the last one and 2%.

In just three years, you can see how much Mahmood's work has moved on and how the use of AI has allowed him to build upon what he already had and take it to another level and get so much more use out of it.

And it sounds like that's both the models moving on, but then the approaches to using the models as well.

And I really liked his analogy of an engine and it's what you've built around it that makes it robust.

2023, it was producing so many essentially mis-truths and fictions, and they've worked on it over three years.

And if you imagine with the AI that we're all using in our lives, how much generalities there is there, whereas this is on a very narrow subject and it's taken them three years of concentrated work on this narrow area to get it to a stage where those fabrications are no longer there and everything's being produced to that great level of accuracy.

That demonstrates the great task there must be with AI on the global scale of making sure that it's not making up anything.

And I think it's that differentiation between like a general, you know, ask and see what you get and actually having it under these guardrails and supervising it very closely, which I thought was amazing.

I really liked some of the findings they had at the end of this process as well.

And one of the concepts I found really engaging is the idea of a collapsing funnel.

Because I've seen collapsing funnels in all sorts of other areas of work as well.

So looking at Tim Lamont's work that looked at the restoration practices of companies and what they say about it, that had this collapsing process going on as well.

But likewise, looking at biodiversity disclosure by ocean companies, by indicators disclosure around the TNFD.

So collapsing funnels is something I think, you know, for a select number of our listeners will go, oh, yeah, that's going to use that phrase because it captures it.

So if you are unable to disclose this particular piece of information, either because you don't know what's going on or you decide not to say anything about it, then everything else sort of wanes off.

Some of the numbers he was giving out there where you've got from 67% of firms who are doing the reporting bit and doing it okay down to there's only 2% I think was the last number he gave where he'd gone through various things such as whistleblowing and stuff like this and then mediation and then satisfactory mediation was the last one and 2%.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.