HW089: Spectrum Analysis and Vibe Coding
The Fat Pipe - All Packet Pushers Pods
September 22, 2026
Keith sits down with Kjetil Hansen to explore spectrum analysis and vibe coding for wireless networking tools. They also discuss building custom tools, using AI for 3D network modeling, and how Kjetil uses OpenIntent in his projects. Episode Links: Kjetil Hansen’s Website Kjetil Hansen on YouTube
Speakers Keith Parsons, Kjetil Hansen
TopicsTechnology
Keith Parsons (0:01)
Welcome back to another episode of Heavy Wireless, part of the Packet Pushers Podcast Network. My name is Keith Parsons, and today I have with me someone I met. Well, I met him a long time ago, but we were just together last week in Manchester at the WICO Fest, and he gave a presentation.
Kjetil, how are you?
Kjetil Hansen (0:20)
Yeah, I'm good. Very good.
Keith Parsons (0:22)
So what was your experience at WIFI Fest, WICO Fest?
Kjetil Hansen (0:29)
Yeah, it was really cool since I wasn't able to come to Prague this year. So like WICO Fest, that sounds cool. And then I see your name there and Nick Turner and Peter McKenzie, and Matt Parker and Dan Jones, and all the same kind of people. I always meet and greet at Prague. So I say, Oh, I need to go here and I need to show something, a presentation that I've been working with for so long time. And I only wanted it to be shown at WICO because it isn't recorded. So for exactly this time of presentation. Usually I love to end up on YouTube. I love it, Keith. But this time I was a little bit unsure, but I really wanted to show what I was supposed to show.
Keith Parsons (1:13)
Well, we all got to see it and cheer and applaud.
Well, for our, this is an audio podcast, so they can't see what you did, but let's talk them through. So first, what was the reason behind why you even prepared that demonstration?
Kjetil Hansen (1:28)
I think it comes down all the time to behind my very first presentation at the WPC, where I had to stand on stage to explain how the Echo Psychic worked. And knowing my brain usually, when I don't know exactly how things works, I dig deep until I know how it works.
And I always looked like I have other types of specmalizers that I have. And when I can't figure that one to be similar, very similar to how the spec sheet of a Psychic 2 behaves, because the spec sheets for the Psychic is massive, like 50 sweeps a second, 19, 39 kilohertz, it's an insane amount of quality. But it doesn't resonance when I look at AI Pro. So I just felt something is up. I need to see what I can do to help out, because I want to see all three bands at once, that I've been asking for since we got three bands.
Keith Parsons (2:24)
I think everyone's been asking for that.
One of the things that, for a long time ago, this isn't for Psychic 2, and I don't have any inside information, but Ekahau did say that their Spec-N is delivering way more data than they can show on the screen, so they had to kind of slow it down. And that was something public from a long time ago. So that might be part of it. But you took that little desire to figure out what was going on, and so what did you build?
Kjetil Hansen (2:52)
Yeah, I build my own tool. Similar thing that I did with the NetAlly NXT, I saw that the G3 was a bottleneck, so I built my own tool to visualize how that looks. And that is free and publicly available that you can download if they have that one. And I just want to see if I do the similar thing to see if I can, some way or another, get access or some of the data, to see if I can visualize it without slowing it down, with showing everything at once.
And that is the thing that I was able to show at WICO. And yeah, I think people like that.
Keith Parsons (3:33)
You kind of got them to stop for a second and just go like, what is he showing? So you knelt down, you picked up a sidekick to, plugged it in, and then your app, and how was your app coded, and where was it running when we got to see that demo?
Kjetil Hansen (3:50)
It runs, it has previously run on Windows.
And then because the MacBook is so relaxed to have a new lap in your living room, it doesn't just work, I just made it over to Swift instead, because it's so easy to code on a MacBook with Claude. But all kinds of AI, but mostly a lot of it is manual, tedious work to know how things work. It's really manual because parsing a lot of data to AI is not feasible, and giving it too much data, it doesn't want to help you. So a lot of things is manual. But I visualized it on a Mac, and an iPad. I should have just showed it on my phone, but my phone was on my chair in the audience. So I wasn't able to do that.
18 more minutes of transcript below
Thousands of transcripts fetched by people building searchable podcast archives
Fetch the whole transcript
The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000791159832