UK AUG: Access Development with AI
Links and resources for my presentation today to the UK Access User Group.
FYI: Today's virtual presentation is only available to paying members of the UK Access User Group:

Tools to Code with AI
VCS Add-in

Agentic AI Coding Tools


Markdown Viewers
Obsidian is best for managing a folder full of Markdown files–such as those that will accumulate in your code repositories over time when working with AI coding agents:

MarkText is a lightweight Markdown viewer and editor that is ideal as the "default" application for opening .md Markdown files on Windows. (Obsidian does not allow you to open a random .md file on your filesystem; it insists on creating Markdown "vaults" at the folder level.)
Building Tight Feedback Loops for AI
The best way to unlock AI's potential in its current form is to provide it with as tight a feedback loop as possible and then let it iterate.



Voice-to-Text: When Typing Can't Keep Up
This is not your phone's built-in voice-to-text transcription. This is a context-aware magician that can actually save you time, especially when interacting with AI via a typical chat UI or an agentic tool like Claude Code or Cursor.

FULL DISCLOSURE: This is an affiliate/invite link. If you sign up via this link and dictate 2,000 words, I get a free month of WisprFlow. Doesn't cost you anything. Doesn't cost me anything (besides this disclaimer note). And if I didn't have this invite link, I would have posted a regular link to WisprFlow.
Feature Development with AI



Future of AI
Future of Software Development with AI
However, we’re fairly confident that the overall trend is roughly correct, at around 1-4 doublings per year. If the measured trend from the past 6 years continues for 2-4 more years, generalist autonomous agents will be capable of performing a wide range of week-long tasks.

Solve 1 Problem 3+ Ways
Coding tools that can run unattended for days at a time will require new techniques in software development. One such technique is to generate multiple algorithms to achieve the same result. This greatly reduces the chances of logic errors in code. I wrote about the technique years ago, but writing multiple solutions by hand to the same problem is not practical in most scenarios.
But if one had access to a tool that excelled at such tedious tasks...

Jevons Paradox
In short, the Jevons Paradox is the observation that the more efficient a resource is to use (code generation, in our case), the faster demand for the resource will rise–often leading to increased overall resource consumption despite the gains in efficiency.
In other words, AI will make writing code so efficient (and relatively cheap) that we will find many more uses for it–leading to greater long-term demand for software developers, not less.

Past Presentations
If you enjoyed my presentation today, you may enjoy some of my previous presentations, as well.








