Rory Madden

I do this but then ask codex to critique the implementation that Claude does. Works really well

But they are different opinions to the ones I like. You can't shore those "other" opinions. If you share the ones I like then it is ok.

Great surprise seeing one of my articles in a newsletter I read every week. If you don't subscribe yet, you should.

My LLM codegen workflow atm A detailed walkthrough of my current workflow for using LLms to build software, from brainstorming through planning and execution.

I've been hearing a lot about different programming workflows to make full use of LLMs, but I want in-depth accounts of how it works. This blog by Harper Reed is exactly what I've been looking for.

harper.blog/2025/02/16/m...

The research equivalent of "if a tree falls in the woods, does it make a sound?". If your research doesn't get used did you do any research at all?

I've found that after about two times to fix a bug, it will just make the program worse. But what works is too nice between llms. Chatgpt - Claude - deep seek etc

I think "chat up" is one of those sayings that means different things in different countries.

Makes the tweet a lot funnier though. Great to see you're working the neighborhood 🤣

One thing my wife didn't factor in was exhaustion. 30 hours without sleep makes it tough to do anything. The epidural was amazing

It doesn't seem to be on Amazon at the moment but you can buy epubs on other platforms www.kobo.com/ie/en/ebook/...

How to Run a Productive Generative Research Interview Running an interview is not difficult. But running a productive interview that uncovers genuine insights to inform your product is a different challenge. This article introduces a proven approach to help you maximise the value of your interviews, drawing on methods from the Jobs-to-be-Done (JTBD) framework.

Want to run better research interviews?

Check out the full article here: https://blog.zeroblockers.com/p/how-to-run-a-productive-generative-research-interview

Always try to interview in pairs!

We all fall into bad habits. One person can rescue the conversation if it drifts into opinion territory. Plus you can play good cop/bad cop to dig deeper into confusing answers. 7/8

People can explain their reasons for actions, but sometimes things don't line up quite right.

Want to know what people really value? Give alternatives:
"Why not choose the cheaper option?"
"Why not hire someone else?"

People struggle to explain their choices until you give them contrasts. 6/8

Pro tip: When interviewing, energy matters more than words.

When someone gets excited or bored about a topic, that tells you more than what they're actually saying.

Dig in to find out why they are so excited, or why the have no interest. 5/8

Broad questions are great to get started, but don't trust these answers! You want to move quickly to evidence, or as Bopb says get into the details.

"Show me your workflow"
"Can I see your browser history"
"What browser tabs do you have open right now?"

Real behavior > opinions every time
4/8

Start the research interview with broad, struggle-based questions:
Push: "What are you frustrated with?"
Pull: "What are you hoping for?"
Anxiety: "What worries you?"
Habit: "What do you love about your current solution?"

This will help you zone in on the most important areas. 3/8

First, the setup is crucial. You want the person to be relaxed so explain there are no "right" answers. Next since you will ask very detailed questions frame the interview like shooting a documentary. This gives you permission to ask detailed questions without making people uncomfortable. 2/8

Want to start running #ux research interviews but don't know how?

When it comes to understanding people's true motivations there is no better interviewer than Bob Moesta. He outlined his approach in the book Demand Side Selling.

Here is a quick overview 1/8

I went to a fortune teller and they said that 2025 was going to be a fantastic year. I will find love, I'll get richer beyond all my dreams and I'll get into better shape.

I stood up and slapped them.

They were really confused but I always want to strike a happy medium

Messy workflows slowing you down?
Priceline’s Aadi & Anthony reveal how #Server-DrivenUI transformed their process into a "#Product superhighway." Alignment, governance & cultural shifts made it possible. Watch Aadi & Anthony break it down: https://uxdx.com/session/server-driven-ui-at-priceline/

I use Figma to store these snapshots. I have a research folder and store all of the snapshots in chronological order.

In terms of organising it for future decisions: I don't. I've tried complex research repositories but data decay is a real challenge. I think continuous, low effort research is best

How to Document Research Interviews After doing a customer interview you need to synthesise your insights right away before your memory fades. And because time is always tight, we'll share a quick and simple format so you can get back to other work.

A structured approach to documenting insights not only makes your write-ups clearer - you end up with better context for each step in their journey.

Check out the full article for more info: https://blog.zeroblockers.com/p/how-to-document-research-interviews

A sample interview snapshot showing 5 distinct areas: interviewee details, quote, experience map, insights and opportunities. The Insights section is further broken into five areas: Push Forces, Pull Forces, Habit Forces, Anxiety Forces and Trade Offs

Wait - five categories of insights - what if we didn't ask questions covering all of these areas?

Don't worry if you're missing insights for some categories. Leaving sections blank is WAY better than making assumptions.

Bad assumptions = bad product decisions! 7/8

We can also add a fifth insights category: Trade offs.

What compromises are they making?
What are they willing to give up to solve their problem?

This reveals their true priorities. 6/8

Push forces: What they don't like about their current approach

Pull forces: Capture what's attracting them to new solutions

Habit forces: Note what keeps them using their current approach

Anxiety forces: Record their fears about changing

These help understand the motivations behind actions
5/8

When running interviews we should ask questions around the jobs-to-be-done emotional forces that drive decisions: push, pull, habit and anxiety.

Structuring your insights using these forces gives clear categories to document against. 4/8

Opportunities are the unmet needs, pain points, and desires that emerge from the interview.

The term "insights" though is more vague; defined as everything interesting that is not an opportunity. Without more guidance, I think we risk missing or misinterpreting important information. 3/8

Teresa Torres (@producttalk.bsky.social) has a great interview snapshot template with five parts:

Interviewee details
Memorable quote
Experience (Journey) map
Opportunities
Insights

Most of these are straightforward, but I think there's a problem with how we document insights... 2/8

#Documentation needs more love!

After doing customer #ux interviews, you need to document insights before your memory fades. But what's the best way to do it? Let's break down how to create better interview snapshots 1/8

DIS = Digital Interaction Score
RIS = Retail Interaction Score

I can't remember where I saw it but someone gave the example that by the time a report when through 3 levels of hierarchy the real meaning was lost. Execs have really poor visibility of what is really happening on the ground. because everyone wants to downplay issues