I highly recommend the feature essay in the latest Quarterly Essay titled "The God we made - the threat and promise of artificial intelligence" by Anna Goldsworthy.
My newsagent's penchant for placing their price stickers over the title makes this look like "Arterly Essay" which is somewhat appropriate as Anna is a music academic but is impressively well read in the AI realm including the amazing Dwarkesh podcasts.
She talks about the impact on writing, music, art, education and work "the bottom rungs of the career ladder have been knocked out".
On education she observes something that I've also thought, that the current crisis in students using AI to answer their assessments is not to be blamed on AI but rather the mass production of assessment.
Apart from her alarm at the high chance that we'll all be wiped out by artificial super intelligence in its quest to make paperclips, she celebrates the capability of AI to help us all educate ourselves.
Spelling checkers didn't make us good spellers. Grammar checkers don't improve our grammar. Will AI Chatbots dumb down our thinking or raise us up to a higher level of enquiry?
Postscript
After reading this I re-watched Spike Jonze 2013 Movie "Her". If Apple ever makes a folding phone he will have been correct in pretty much everything.
Work continues on my little FreeDV RADE V1 native app for macOS. Feedback from some users bemoaned the lack of fancy visualisations so I've addressed that. The latest version also improves the operation of the Reporter window quite a bit and adds a filter by my frequency option.
It's great to see all the third party FreeDV RADE apps starting to appear. We have a list of known ones on the FreeDV site here. Let me know of others that I've missed.
This week on ABC Nightlife we discussed new data from a study by the University of Newcastle, published in the British Medical Journal, that concludes that more than 80% of teens under 16 are still accessing social media sites banned in Australia. The government has moved to increase penalties but our teens are technically pretty advanced it seems.
Social media is currently awash with AI generated content designed to encourage division and presumably promote far right politics.
I've been a bit slack with my dipoles here in the bush, the coax just hangs from the PL259. It's been fine until kangaroos catch the coax and rip the coax out. The repair isn't too hard but I thought I'd look for a way to relieve the strain on the connection and ended up with my own design for 3D printing.
It has a channel that the coax fits in to and two cable ties to grip it.
3mm cord lifts the coax slightly so that there's no weight on the PL259.
I designed it in Tinkercad and you're free to grab it for printing for further modification here. It will be interesting to see what breaks when a Kangaroo goes through it next time.
For a few weeks we've run in to analog traffic when using 7.177 so today we moved to 7.197 and the frequency was clear so I think we might stay here in future. Most stations figured out that we'd moved so we had a good rollup and an interesting discussion about the proposal to report centre frequency rather than dial frequency in future.
It was noted that Icom radios have an option to tune to the centre but most of us have turned that off.
FreeDV reporter was a little unstable today.
Stations on the net included VK2AWN, VK2BLQ (combination digital voice and valve audio with 6AU6 and 6AQ5), VK2KNC, VK3FC, VK3GTP, VK3JCO, VK3KEZ, VK3ZTR, VK5AG, VK5ABE, VK5FD, VK5JPL, VK5RA and me, VK3TPM.
In a video call with a friend recently I noticed that the video was noticably clearer than normal. They were using an iPhone as the camera rather than the camera in the monitor.
It didn't take long to use TinkerCad to design a simple mount for my monitor. I've made it public here.
The phone sits nicely with the camera used in this mode just above the monitor's build-in camera.
Here's the device. After the first prototype I went back and angled the phone forward a few degrees.
I use two of these but you can get by with one. My monitor is about 18mm thick so you might need to adjust the hook part for your monitor.
Amazing work by Jon Dawson at 101 Things and others to bring a very capable SDR to the RaspberryPi PICO.
I've forked the project here and had claude write some very useful documentation on the user interface operation.
It sounds great:
You might notice that my display is offset two pixels to the left. The solution is to select the OLED type of SH1106 instead of the default SSD1306. (It seems they are frequently substituted).
If you're interested in building one of these, I'm not aware of anyone selling these boards - you might try contacting the Manly Warringah Radio Society or it's pretty easy and low cost to upload the gerbers from VK2ARH (he designed this board) git repository and get them made.
I've recently ordered Jon Dawson's board from JLCPCB and the price for 5 was US$5 + shipping. (Shipping costs significantly more than boards).
A terrific net this morning on 40m here in VK3. We had stations in VK2, VK3, VK5 and VK7. Stations were using not just FreeDV 2.3.1 but also the macOS FreeDVNeo and the Thetis SDR software. One was listening on an SDR that decodes RADE at the Nepean Men's Shed.
There was some SSB partly overlapping our normal frequency of 7.177 so I moved up 2kHz to 7.179. Most people realised but at least one didn't and when they transmitted interfered. In future I'll be careful to move at least 3kHz away. Lesson learned.
Here's some of the stations on the net, if you haven't heard RADEV1 this will give you a good idea of how it typically sounds.
Stations on the net included: VK2AWN, VK2BLQ, VK2DWG, VK3FAR, VK3FC, VK2KNC, VK3GTP, VK3JCO, VK3JF, VK3KEZ, VK3ZTR, VK5ABE, VK5FD, VK5RA, VK7HOB and me, VK3TPM.
Thanks to Stephen, VK2BLQ, for sending this clip of his reception of me in Sydney (about 1000km away).
This week on Nightlife Tech we discus the announcements at Apple's Worldwide Developer Conference. iOS 27, macOS 27 and the others will ship to users in a few months. The Siri features Apple showed two years ago seem to be real this time and the overreach in the Liquid Glass user interface design is being dialled back to be more usable.
In other AI news, after the success of the SpaceX IPO, based largely on future profit from their AI component, the other AI companies - OpenAI and Anthropic - are also planning their own IPOs. One model, "Fable 5", from Anthropic was released and then withdrawn for non-US citizens and this highlights a risk to Australians who rely on AI cloud services not controlled by us. Listen here: https://www.abc.net.au/listen/programs/nightlife/nightlife-tech-talk-with-peter-marks/106805414
A terrific net this morning in south east Australia. Amazingly there has been mobile FreeDV activity during the week! Most people are now on 2.3.1 which resolved a Windows audio issue.
Several stations said that they'd like an option in the FreeDV app to auto-start the Modem at launch. This is particularly helpful when operating on headless, or small screen Raspberry Pi. I've raised a feature request here but it needs some more backers. Please add your email if you support this.
Built for my own entertainment but might be of interest to other macOS hams. FreeDV Neo is enough of an implementation of FreeDV RADE V1 to be usable on air. More info here including a download link.
This app is not officially supported and you should first use the official FreeDV app from the project.
This week on Nightlife Tech Talk with Philip Clark, we discussed the US$1.7 Trillion valuation of SpaceX (if their IPO goes as planned), NVIDIA's announcement that they're getting in to the personal computer CPU market with an ARM based chip that seems influenced by Apple Silicon, and if you go in to the Australian outback there's a good saftey reason to carry a recent cell phone from Google or Apple. Listen here: https://www.abc.net.au/listen/programs/nightlife/nightlife-tech-talk-with-peter-marks/106751536
I'm a big fan of Bill Meara's Soldersmoke podcast and blog. It's always a pleasure to hear from Bill and last week he contacted me asking for help with the website where he keeps an archive of the past editions of the podcast.
Bill has been editing the HTML of the page manually using a web editor that had hit a page size limit of 1MB and would no longer let him add new podcasts so it had fallen behind. Here's how the page looked:
At first glance I thought this was an easy little job, get in via FTP and paste in the new episodes. When I looked more closely at the HTML, and ran a validator on it, my heart sank. It's a credit to the robustness of web browsers that they managed to render this as well as they did. The podcast has been around for 20 years and while some of the HTML was obsolete, worse was that Bill's copy pasting from his blog had all sorts of errors in it. Mostly table tags that weren't balanced etc.
The HTML had ballooned out and had parts that looked like they'd been generated in Microsoft Word's export to Web - huge runs of fonts and styles repeated over and over.
I thought this would take many days of manual clean up and editing so I turned to Claude code for the project. I downloaded the various existing pages to a local folder, explained what we wanted and mentioned the blog as a place to get missing information.
Claude installed BeautifulSoup and proceeded to write python scripts to extract just the needed information from the existing broken HTML. It also looked at the blog to get missing info, looked at YouTube and even got the modification dates of the mp3 files to figure out publish dates. The result is at soldersmoke.com and now looks like this:
The HTML now validates without any complaint and it's much more efficient and compact than before.
AI will put web developers out of work to some extent. On the other hand it's fantastic at doing what would have been an incredibly time consuming menial task. This kind of job might not have been done at all except for this marvelous new technology. It's not perfect but it's amazing what can be achieved.
Thanks to Bill for all his great content, it's great to see the podcast archive remains available and up to date.
My wife unfortunately broke her arm a few weeks ago and can't drive so I've been away from home looking after her. Normally on a Sunday morning I would conduct a FreeDV RADE V1 net on 7.177.
I tuned in this morning via the excellent Ironstone Range WebSDR and was pleased to hear that stations can chat without my help ;-).
This screen recording was made by using Loopback to pipe audio from the Safari Web browser into my home brew FreeDV app for macOS called FreeDVneo.
This week on ABC Nightlife Phil and I preview what's coming up in this week's annual Google I/O conference. Headline topics include AI, Android 17, the Chrome browser and a new operating system for laptops dubbed "GoogleBooks".
We make a call for RCS messaging which will smooth out messaging between Android and iOS.
There have been an unusual number of security updates from both Microsoft and Apple - are these vulnerabilities being found by AI?
The author looks at the dramatic change in the way the last four generations get their news and form opinions. Boomers, Gen X, Millennials, and GenZ have different habits and to some extent live in entirely different versions of Truth.
Tony has a list of recommendations that would be valuable for anyone who seeks to transmit a message to the public. Recommended reading.
A good net here in south eastern Australia this morning from 09:30 eastern time on 7.177Mhz. Most people are running version 2.3. The audio problem for Windows users seems to have been figured out and I can see that there are early builds of a 2.3.1 in the wild.
This weekend has been the Dayton Hamvention and FreeDV programmer Mooneer has been busy with a stand at the show.
My own transmit audio was apparently terrible. Nothing much had changed since last week but I was noticing a bit of RF getting back while I transmit so perhaps that was it. I swapped to a USB headset and it cleared up.
Stations participating this morning were VK2BLQ, VK2KNC, VK2EX, VK2KO, VK3FC, VK3KEZ, VK5AG, VK5FD, VK5JPL (ex VK5KVA), VK5RA, VK4GRA, VK3LV, VK3GTP and me VK3TPM.
At the recent deceased estate sale of Graeme VK3NE's gear I purchased a 160m AM transmitter. Crystal locked to 1832kHz. Turns out it's a Drew Diamond VK3XU design from 2006. Rather nicely put together but missing the audio power modulation stages. Anyhow, step one is a suitable antenna.
There's a big dead tree near the shack so I put up an inverted L. Basically as much wire as I could find. Up the tree then horizontally over to another tree. Using a nano VNA to sweep SWR it was much too long and resonant below 1Mhz. After a series of cuts it looks really nice.
I hear a lot of noise on this antenna, S9. When I transmit I can hear myself on the Strathbogie SDR so I know it's getting out.
An audio amplifier is on the way so I'm looking forward to getting on the air. AM on 160m - what memories.
A good net this morning after yesterday's release of version 2.3 of the FreeDV app. Two reports of Windows users having audio trouble with 2.3 that led them to roll back but others are finding it good.
Stations heard included VK2AWN, VK2BLQ, VK2CJB (portable VK4), VK2VCO, VK3KEZ, VK3KQT (On Android), VK3ZD, VK3UBK, VK3ZTR, VK4DKW, VK5AG, VK5FD, VK5KVA, VK2KO and me VK3TPM.
The FreeDV project has just released version 2.3 of the FreeDV GUI app for Linux, macOS and Windows. There's also support for running it directly on a compatible Flex 800 or Aurora series radio.
The 0.1 increase in version number rather understates the change in this release. FreeDV RADEV1 has, until this version, been implemented in Python. It worked well but required the app to have an embedded python interpreter and the required libraries (numpy, scipy, etc).
Version 2.3 drops all the python requirements and uses a C port of that code. As a consequence of this change, the apps are one tenth of the size, start the modem faster, use less CPU and memory.
Aside from the move from python to C, Mooneer has been very busy with many enhancements listed in his latest update.
The availability of the C port means that FreeDV RADEV1 is starting to appear in other apps.
This week on the ABC Nightlife program we discuss the rather extreme claim on Four Corners this week that police can be "tracked" by hackers using the Bluetooth in their tasers and body cameras made by Axon. I think this is an over-reach.
Where we are being spied on is when visiting Microsoft's LinkIn site which, if you're using the Chrome browser, probes for the presence of hundreds of plug-ins. This happens even if you aren't logged in and the amount of information they can derive is surprising.
We started an hour earlier, at 9:30am eastern time, but still had a good roll up. This morning was notable in that there were several stations on early builds of version 2.3 and also we had David, VK3KQT, who was running on an Android version. He sounded good even though he said he was in speaker phone mode.
As an early adopter of home automation I've suffered over the years from a collection of incompatible devices each requiring a proprietary app to control the one brand of light or switch or whatever.
When Apple, Google and Amazon got together to standardise Matter and Thread I jumped in and purchased some bulbs from Nanoleaf. Matter is the protocol which can run over different transport including Wifi and Thread is a mesh radio system that works in Wifi black spots and seems quicker.
Apple has been shipping Thread border routers built in to devices such as the Apple TV and HomePods so I have a few scattered around already. Thread devices also act as mesh relays so it's easy to cover the house thoroughly. It's also in some phones which greatly helps with configuration.
When I got the bulbs I took advantage of the ability to use Apple's Home app for configuration thinking it would be smoother than the dodgy proprietary apps. Big mistake.
I've had a rough time until this weekend. Setting up new bulbs was frustrating and often took several goes at it and often lights, which I had set to come on at dusk, would often come on very dim and purple, requiring manual intervention.
This weekend I got a notification that there was a critical software update to do. It was actually two updates in a row. It didn't go well and I ended up deleting all the devices and setting them up again using the Nanoleaf app which then did the firmware update.
Only some of my bulbs support Thread. You can see the little icon on those ones above. The Thread devices definitely work better than the Wifi only ones.
I'm afraid we're not free of proprietary apps but at least things are working reliably for me now.
My advice:
Choose Matter & Thread capable home automation devices
Do use the manufacturer's app and apply the software updates
After all that add the devices to Apple's Home app (or Google's equivalent)
There was a feature I wan't too keen on, it makes the light warmer in colour as the night goes on, but having lived with it a bit I quite like it now.
Claude Code has been my helper for a few months now. Others have observed that while it can help a non-programmer to create software it can be even more help to an experienced programmer.
Knowing what to ask for, being able to evaluate the generated code, and thinking up tasks are things that a seasoned coder can employ to get fantastic results.
In February 2026 I had the bright idea of asking Claude to port the python in FreeDV RADEV1 over to C. To my amazement it did that daunting task in under a day of prodding. This has saved our proper programmers months of tedious work and has led to a trickle of other FreeDV applications. I made one myself that is a simple native application in SwiftUI for macOS. Here it is receiving Anthony VK3YSA. Download here to try it out.
Another app is a simple client that talks the excellent digital chat mode Olivia 8/250, again native SwiftUI for macOS. Download here to try it out.
In the shack I run Linux and have been getting Claude to make little utility apps using GTK+3. (They will also build on macOS but look rather out of place).
Presumably because of the extensive amount of open source software to train on for Linux and GTK, it's particularly good at whipping up little apps.
The job of the programmer has changed probably for ever. I'm dreaming up projects that I never would have considered before. I wake in the night with ideas and have anxiety about not using the available quota for the day.
It was proposed that we move to an hour earlier for improved propagation and with a few exceptions this was fine with everyone. I'll update the list on the freedv.org site to reflect this change.
Two of us were running an early build of version 2.3 which has the C port. It seems solid to me and Leighton, VK3FC, and I look forward to it being released as it's smaller and faster.
Stations included VK1RF (Running the Flex AppImage!), VK2KNC, VK2BLQ, VK3CCR (welcome to the net Brian), VK3FC, VK3GTP, VK3JCO, VK3KEZ, VK3ZD, VK4EV (not heard), VK5AG, VK5KVA, VK5RA, VK5ST, VK2AWJ and me VK3TPM.
It's amazing to me that a 500GB disk can get full but that happened to me on the Mac Mini yesterday. I have a lot of large developer tools including Xcode and Android Studio. Like any unix system, macOS doesn't do well when disk space is low. I've long had an external drive and store source code there but a lot of space was taken up inside my Home directory in ~/Library so I thought I'd try putting home on the external drive.
Doing this is pretty easy:
Go to Settings > Users & Groups
Control Click on your username and you see an "Advanced Options..." menu item
Scroll down to Home directory and Choose.. the disk you want to use
You'll be prompted to reboot and when the machine comes up it will be like a new user setup with all the Apple ID login etc. Once all that is done you can copy over files from the old home directory, things like Documents, Developer, Movies, etc.
I left it to iCloud to sync over my music and pictures.
Perhaps there's a way to move over the old home but I decided not to risk it.
All is working well now but the external disk is noticably slower than the fast internal SSD but Applications, which are still on the boot disk, still launch quickly.
Was chatting with FreeDV regular Joe, VK3SRC, who's been pushing for an appliance like implementation of FreeDV RADE V1 for Raspberry Pi, mainly for portable operation. It occurred to me that a better way would be to have a client app for the phone - we all carry phones and they are powerful enough these days.
The project noted recently an effort by talented developer Lee, BX4ACP, who has started work, open source, on both iOS and Android clients.
I forked the iOS code and started looking around. Lee commented that there would be problems with the iPhone audio system but his app looked quite polished so I tried to get it going.
It turns out that while you can attach other audio devices to an iPhone including bluetooth or USB audio devices, the phone basically takes the device as an input and output pair. My use case is to have the phone connected to both a transceiver and probably a headset or some sort.
Audio must come from the transceiver, be decoded and play into the headset. For transmit we'd sample audio from the headset's mic and send encoded audio to the transciever.
I fought the OS for a few days and finally gave up. It seems that Apple doesn't want developers changing audio devices programmatically. They want the user to be in control.
Talking with Mooneer this morning, he raised the idea of special USB audio hardware, rather like what was done in the EzDV box which even uses the separate channels of a stereo pair.
I dialled back what I was trying to do and now have a RADE V1 decoder that takes audio from the input of a USB sound dongle and sends decoded audio out to the same device.
This works well. The phone easily handles the load. It's a pity 3.5mm patch cords are needed but I think a portable decoder like this might have some applications.
Another problem on the iPhone is that Apple don't let you talk to USB serial devices unless they've been specifically approved by them. A friend suggested that a bluetooth to serial adapter would be a work-around.
In the end though, I think an Android app might be the solution for portable RADE operation - until we get it built in to a radio.
This week on ABC Nightlife we discussed the stepping aside of arguably the most successful CEO of any company in history with the move of Apple's Tim Cook to Executive Chair of the Board. The AI model so powerful it's too dangerous to release and the boom in electric vehicle sales around the world due to fuel prices - and what needs to be done to make them practical in Australia. https://www.abc.net.au/listen/programs/nightlife/ai-revolution-arrives-in-software-development/106589944
Good conditions for the Sunday morning 10:30am net but a smaller group probably because I've been away for the past two weeks. I had reports that I could be heard directly in VK4 and another person who listened via Web SDR.
There were deep fades on the OFDM waterfall but all signals could be decoded well enough. I was particularly pleased to hear that David, VK3KQT, was talking with my C port client on a Raspberry Pi and Peter, VK3PCC was using a pre-release build of the FreeDV client using my library.
Participating were: VK2AWJ, VK2DWG, VK2KNC, VK2BLQ (once he got the audio working!), VK3KEZ, VK2UMZ, VK3KQT, VK3PCC, VK3UBK, VK3ZTR, VK5ABE, VK5AG, VK5FD, VK3JF, and me VK3TPM.
Today I had the pleasure of speaking at Redfest north of Brisbane. I spoke about how AI is being used in FreeDV and the exciting development (with the assistance of AI coding) of the port to C of the existing Python code in the Radio Autoencoder.
I attended a fascinating talk about Allstar.
There were several commercial vendors displaying their gear in the main hall plus an outdoor area with lots of great second hand equipment at good prices.
My thanks for the invitation and congratulations to the organisers for their hard work and professionalism.
There's been a bit of discussion at our radio club about the origin of the pair of FSK signals on the 40m band in recent months. Here's how they look and sound:
They sounds a bit like morse code but as you listen you soon notice that they aren't. One of our members was convinced, due to the strength, that they come from Australia. I did some internet sleuthing and the conclusion seems to be that they come from the hotspots on this map:
From Google’s AI overview:
These mystery FSK signals on the 40m ham band (approx. 7.061–7.068 MHz and 7.100+ MHz) are persistent, paired, non-amateur, wide-shift (~7 kHz) FSK signals likely originating from Eastern Europe/Russia, potentially related to regional conflicts. They resemble slow, garbled CW (20 baud) but are actually digital, military-style 5N1 encrypted data.
Key Details of 40m Mystery FSK Signals:
· Location & Characteristics: Usually found as a two-channel, high-power system with a 7 kHz separation (e.g., pairs near 7.0615/7.0685 MHz or higher in the band).
· What they sound like: They sound like slow, rhythmic, or "gibberish" Morse code to the ear, but common CW decoders will not read them properly.
· Likely Origin: The signals are believed to be related to military, government, or state-level monitoring, possibly coming from the Sevastopol/Crimea region or near the Russia/Ukraine conflict area.
· Modulation: They are 20-baud FSK (Frequency Shift Keying) with a 7,000 Hz shift, often characterized by a constant, repeating 255-bit, 17-bit frame, 15-frame superframe, designed to appear like noise or to be jamming-resistant.
· Timing: They appear as "key-down/key-up" pairs, meaning when one frequency is transmitting, the other is silent, suggesting a 2-channel diversity system.
· Distinction from Amateur Activity: While some users initially suspected Winlink (a, b) or heavy traffic, these signals are widely considered "intruders" and not standard amateur digital modes like FT8 or RTTY.
These signals are a well-documented annoyance to amateur operators in Europe and North America and have been appearing in bursts, especially between 2023 and 2026.
On Facebook: https://www.facebook.com/groups/hfunderground/posts/1377589986956733/
What is this random strong CW signals appearing in 40m band simultaneously in three frequencies like 7050 7065 7070. We believe it may be from Russia.
These are two fsk transmissions with 7 kHz shift and 20 baud. They change frequency after about 10 minute transmission with 1.5 kHz.
It is Iranian.. encrypted message system built by the Russians. been tracking it for weeks. It is used for sleeper cells to give them instructions. And you can take that to the bank. Jack
Also on Facebook: https://www.facebook.com/groups/22212809595/posts/10163697406274596/
For the past week+ there has been some intruder on the 40M CW band easily heard on both coasts of North America.
It is occasionally QSY'ing anywhere from 7.04MHz to 7.07MHz and may be outside that range goo. It is often on a whole multiple of 500Hz.
Although we are most often hearing one CW frequency with our rigs set at a normal bandwidth, I can often find an "evil twin" which is key-down while the other one is key-up, so I think it's more likely to be super-wide shift FSK. Below for example is me putting receiver 1 on 7.063MHz and the other on 7.070MHz circa 0400Z on Jan 30 2026. You see when one is key up the other is key down.
The basic signal rate seems to 0.05seconds (20 Baud). Our CW-seeking brains seem to try to interpret it as CW because if we here one element at 0.05seconds we think it's a dit and another at 0.15 seconds our brains.
I do not recognize 7kHz shift 20 baud FSK as a standard encoding anywhere but someone might know better than me.
I'm gonna record some longer patches and see if I can recognize any framing.
Tim N3QE
From SOTA Reflector: https://reflector.sota.org.uk/t/some-weird-cw-found-on-40m/38289
While trying to impove my indoor loop on 40-80m (coil added in a middle makes a noticeable difference on both TX/RX!) I’ve came across those strange CW signals. Perhaps they are trying to emulate a mechanical key with no audio control (although a vertical wouldn’t sound like that) But what could be a purpose? Another weird thing is that those sigs are present on four frequencies simultaneously with S9+:
7041.50 kHz
7048.50 kHz
7063.00 kHz
7070.00 kHz
Weird CW on 40m
Is that something known or a new thing? There is plenty of strange stuff on air now, but I haven’t heard such lovely Morse yet :slight_smile:
Also
Cheers David, using that “wide-band FSK” hint I’ve found another Youtube sample and also some explanation, which I quote here for reference:
“Sevastopol; Extremely powerful two channel 20b/7000 Hz FSK paired signals re-appear from Apr 23rd on 40m band ~ 7045cf and 7065.5cf, each channel seems to be 7 kHz from mark and space, chosen for it’s immunity from being easily jammed, at times the pair will shift frequency; originating from Eastern EU / Western RU region, now Crimea with transmitters near Sevastopol. The modulation rate of 20b with a shift of 7000 Hz. -The signal frame is 17 bits, the first of which is a 0. -The superframe is 15 frames (255 bits) and is constantly repeated so it does not appear to carry any useful information Military encryption 5N1 use in related signal connection to the conflict between Ukraine and Russia is likely. Contributor ANgazu thank you for your analysis! Recorded 12 June 2025 by Tyler Stampfli, KA0KA …”
As Artemis II makes its lap around the moon, astronauts on board keep in touch with their control room and their loved ones using a different type of internet on NASA's Deep Space Network.
ACMA announces new rules for telco companies surrounding their mobile coverage maps ensuring clearer and more accurate results.
Recently I purchased a rather nice backpack which features an area that can hold velcro patches. I thought it would be nice to have one with my callsign on it and went hunting for one off emboydered patches. An idea occurred to me that a 3D printed banner could work. Here's the result:
Close up you can see that I've coloured in the engraved letters with a permanent marker. Some sort of paint would be better.
I'm able to write OpenSCAD code but it turns out so can Claude Code. This is what we came up with:
// Rectangular plate with embossed text
// Dimensions
plate_length = 90; // mm
plate_width = 20; // mm (adjust to taste)
plate_height = 2; // mm plate thickness
text_depth = 0.8; // mm how proud the text stands
rim_width = 1.5; // mm wall thickness of raised edge
rim_height = 1.0; // mm how tall the rim stands above the plate
Today at the Macedon Ranges Amateur Radio Club coffee morning, Eric, VK3EJD, brought along a new gadget that he hopes will remove noise he experiences from a neighbour's solar system.
I used to have a noise reducing speaker from BHI but was never terribly happy with it. He reminded me that I'd started building a live noise reduction application for macOS using RNNoise so when I got home I did some tinkering and got it working. Here's a demo - although the noise is not terribly bad. (I switch it in and out during the video.
Here's an example of removing solar panel noise from an AM transmission:
It's promising but needs to be trained for your specific noise. In this I'm using the default model. This runs easily on an Apple M4. I wonder when we'll start to see this built in to amateur radio gear?
I had a nice capture of a busy 40m band done using a HackRF Mayhem portapack and I wanted to play it back using a HackRF connected to a computer using hackrf_transfer.
Even though the native format of the hackrf IQ is 8 bit signed interleaved samples, for some reason the portapack pads these out to 16bit so conversion is needed.
At this week's Radio Old Timers Club lunch in Melbourne I presented a talk outlining the good success I've had recently using Claude Code for ham radio applications. Also there's a bit of an overview of how FreeDV RADEV1 uses machine learning already.
A few AV hitches and the crowd was finishing lunch but feedback was pretty good.
This week on ABC Radio's Nightlife program we discuss a recent post by Microsoft acknowledging a list of things that users complain about. They promise to address these in the year ahead. The complaints about Windows comes at a time when Apple CEO Tim Cook reports that, with the launch of the low cost MacBook Neo, Apple has seen the highest number of new users to macOS in the company's history.
Continuing a trend in recent months, layoffs in tech firms are being blamed on AI being used to replace workers - but is that really the case or just an excuse to downsize?
Finally there are rumours that Amazon is having another go at a product in the mobile space. Code named "Transformer" it's likely to be an AI driven portable device. Inevitably something will replace the mobile phone but is this it?
This week on ABC Nightlife I spoke with Rod Quinn about Apple's new MacBook Neo, a sub-$1000 laptop that competes very well with Windows laptops in that price range. Apple has made some compromises but kept their quality standards up.
We discussed the US AI controversy where Antrhopic, makers of Claude, asked the Department of Defence to not use its products for widespread domestic suveillance or autonomous weapons. DOD would not agree to those limitations and has done a deal with OpenAI, who presumably are ok with that.
Finally, the ban on social media for Australians under 16 introduced in December seems to be spreading with Indonesia announcing they'll introduce it gradually from March. They join Denmark, France, Germany, Malaysia, Spain, the UK and Greece in either announcing or at least considering a ban.
Too early to know how effective it's been in Australia but we do see that VPNs are booming...
The Russian FSK transmissions were very strong this morning which ruled out using 7045 so I parked the net on 7177 and didn't expect too many stations to find me but there was no problem.
I've been working on my C port of the FreeDV app's main functionality over recent weeks. It's really just a proof of concept but yesterday I got reporting to qso.freedv.org working which made it practical to run today's net. That feature, combined with a minimal implementation of Hamlib, just enough to read transceiver status and turn transmit on and off, has made the app quite usable.
The app worked smoothly and the following stations were seen: VK5RT, VK5LN, VK5LO, VK5KVA, VK5AG, VK5ABE, VK3ZUM, VK3XCI, VK3UBK, VK3PCC, VK3KEZ, VK3GTP, VK3CDH, VK3BAL, VK2KO, VK2DWG, VK2CJB, VK2BLQ, VK2AWY, VK2TTL and me, VK3TPM.
This week on ABC Nightlife, I discuss the final removal of the truly evil UK Post Office Horizon software system that led to 13 suicides among innocent sub postmasters, job offers for the OpenClaw developer, and the latest video creation models that are rocking movie producers - or is it just an evolution of the Computer Generated Images we've had for decades. https://www.abc.net.au/listen/programs/nightlife/actors-to-be-sacked-for-ai/106382876
A very well attended Sunday FreeDV net on 7045 this morning from 10:30 eastern time. Stations on frequency were: VK2AWJ, VK2BX, VK2UMZ, VK3BAL, VK3CDH, VK3DJB, VK3FC, VK3GTP, VK3ID, VK3KEZ, VK3PCC, VK3PTR, VK3SRC, VK5ABE, VK5AG, VK5KHZ, VK5KVA, VK5MH, VK7DMH, VK2KO, VK2KNC, VK2YW, VK2CJB and me VK3TPM. 24 stations!
While working on my C port of FreeDV RADE1 I wanted a reliable radio signal to test with. While I have off air recordings from the FreeDV GUI app I thought it would be nice to generate my own files to transmit with the HackRF Portapack Mayhem which can record to C16 files and play them back. There's a TXT file with each recording that specifies the sample rate and output frequency.
I've made a python utility that can read a WAV file recording and produce C16 sideband. Here's how it sounds.
The HackRF is a wonderful tool for radio experimentation. I'm using the standalone Mayhem but it works just as well from a computer via USB with hackrf_transfer.
An excellent FreeDV RADE1 net here in South East Australia with 23 stations on frequency. Some for the first time. Stations: VK5KHZ, VK5KFG, VK5AG, VK3ZUM, VK5KVA, VK3UBK, VK3SRC, VK3KEZ, VK3JCO, VK3GTP, VK3FC, VK3DJB, VK3CDH, VK3BQO, VK3BAL, VK2YW, VK2XOR, VK2UMZ, VK2KNC, VK2DWG, VK2AWJ, VK5HM and me VK3TPM.
Normally I call in stations in alphabetical order but today I thought I'd trick Jack VK5KVA, by going in reverse.
Today marked a milestone in my investigation into using Claude Code with a complex codebase. I had a FreeDV RADE contact with another station using code it wrote under my instruction in just one week - very much part time. Here's a receive demo:
My thanks as always to Joe, VK3SRC, for assistance in testing and feedback. Also thanks to Britt, VK3AOB who responded to my first call which I really expected to not be decodable.
Here's how Claude Code looks while working on a bug in the code (click to enlarge):
I'm developing on Linux Mint. The app is C++ using GTK and Claude seems very knowledgable in this scenario. I suspect that the large amount of open source software around is a great advantage for this sort of thing.
The code is here: https://github.com/peterbmarks/radae_decoder As you can see the name of the project is no longer correct as it encodes as well.
Have you heard of a tool called OpenClaw? Apparently, you install it on your computer and it acts like your personal assistant. it's generating a lot of buzz and users are saying it's like they're 'finally getting a glimpse of the future'.
Tech Talk takes a look at this and how the big four tech companies are faring after they've spent big on AI.
Peter Marks, mobile software developer and technology commentator from Access Informatics, joined Philip Clark and listeners to Nightlife with the latest tech news. Listen here: https://www.abc.net.au/listen/programs/nightlife/nightlife-tech-talk-with-peter-marks/106328266
An excellent Sunday FreeDV net this morning with VK2, VK3 and VK5 stations all with good SNRs.
Some interest was expressed in the C port of the RADAE python code which is encouraging. On the frequency, (not everyone transmitted), VK2AMF, VK2AWJ, VK2DWG, VK2GMH, VK2YE, VK3DJB, VK3FC, VK3GTP, VK3JCO, VK3KEZ, VK3XCI, VK3YV, VK3ZUM, VK4TEW, VK5ABE, VK5KHZ, VK5KVA, VK5LN, VK5ST, VK6YR, VK3BAL (running three copies of FreeDV!), VK3AOB, and me VK3TPM.
One of the objectives of the work on RADEV2 is to replace the currently required Python runtime from the distributed FreeDV app. This would make it easier to install, smaller, and hopefully more efficient. (Other objectives are improvements in how it works). Python is an important part of the project and is used for prototyping and testing - that won't change.
Re-writing all the python in C is a daunting, tedious, expert task, estimated to take many months that would result in an app that works pretty much the same as it does today.
I've been experimenting with Anthropic's Claude Code recently and it occurred to me that one of the things LLMs are particularly good at is translating between languages, human or programming. The result of just a few days of guiding Claude Code is an experimental RADEV1 decoder app.
The intial work, which includes command line tools to decode and encode from and to WAV files is on github here.
The proof of concept Linux application shown above is in a Github repo here.
While it seems to work and does pass important tests, this is not code supported by the project at this time.
Claude Code enabled me to attempt something I wouldn't dream of until now. It was quite a big job for it (I hit my quota on two separate days and had to wait to resume).
Working with an LLM to write software does still require enough knowledge to know what to ask for and to guide the work.
I hope this work will encourage others to develop innovative FreeDV RADE software along side the official FreeDV App which remains the reference release. My sincere thanks to Dr David Rowe for his fantastic work and his help and encouragement with this little experiment.
