I'm afraid to say I rather fell into it. Or rather I joined a part of the UK Civil Service which champions modern software development methods. The reason that the Digital sections of Government Departments were created was (to my mind) twofold. One was as a reaction to some terrible IT projects that had plagued Government IT for years. Second was a desire to be far more focused on User Needs and to ensure that the user was put foremost in any new development.

Security.
The challenge of ensuring that the development practices we want to see and know deliver far better outcomes are done in a way that is not only secure, but trusted to be secure. Security needs to an ally in development and not the blocker that it can all to easily become.

I don't work at the bleeding edge by any means, but I'm hopeful that (if it can stop being labelled as Artificial Intelligence which it is not) the developments in natural language models will open up technology to people who would never think of themselves as digital natives. In the broadest sense I would apply this to things like Github co-pilot, but at a simpler level as voice based assistants expand and improve, the idea of getting Alexa to communicate with an API to file a tax return should be possible.

We've been focusing on Security of our Platform for some time, but only recently have we realised the power we have to look for vulnerabilities and improve the security not just of the platform but every service we host. It seemed like a natural topic for a talk because we think we've discovered some interesting things on our journey.

• Security can't be done effectively by ranking CVS scores.
• Consistency is a security advantage if you leverage it.
• Knowledge is power. Obviously.

Sorry - I know it's starts tomorrow, but I've not really looked yet. The day job is likely to keep me quite busy for parts of it - the only real downside to being a virtual attendee, I haven't cleared my calendar.

I fear this talk could be quite niche and people might take the view that it's only worth listening to if you running a huge platform and dealing with vast numbers of sensitive transactions. I hope that is not the case and that we're able to show that there are general principles that would be useful to consider whatever the scale of your project.