[Térence Grover] had a very special coin—a  €1,000 commemorative piece only available to Monégasque nationals. If you want to flip one, normally you’d have to go snatch one up from somebody in Monaco—or you could just do it online!

Yes, he built an automated online coin flipper to flip this very special piece of coinage. A 12-volt solenoid is fired to flip the coin into the air. It then lands on its 3D-printed tray, where a Raspberry Pi-based computer vision system built with OpenCV and a TFLite model classifies whether the result is heads or tails via a machine learning algorithm. An iris mechanism operated by servo motor then centers the coin on the tray, so it sits back over the solenoid, ready to flip once again. [Térence] was eventually able to refine this simple homemade build to the point that it ran autonomously for a full 50,000 flips on a livestream without issue.

The mechanism in this build is not dissimilar to a coin flipper we’ve seen before. We’ve also explored the statistics involved, too. Video after the break.

This week Jonathan chats with Johannes Millan about Super Productivity and Parallel Code! Those are two very different projects, but both aiming for helping us get our work done. Super Productivity is a scheduling and time tracking suite, while Parallel Code is an almost-IDE for managing and isolating AI coding agents. This episode has something for everybody, so check it out!

• https://github.com/johannesjo
• https://super-productivity.com
• https://parallelcode.app/

Did you know you can watch the live recording of the show right on our YouTube Channel? Have someone you’d like us to interview? Let us know, or have the guest contact us! Take a look at the schedule here.

Direct Download in DRM-free MP3.

If you’d rather read along, here’s the transcript for this week’s episode.

Theme music: “Newer Wave” Kevin MacLeod (incompetech.com)

Licensed under Creative Commons: By Attribution 4.0 License

It’s a phrase we use a lot in our community, “Drink the Kool-Aid”, meaning becoming unreasonably infatuated with a dubious idea, technology, or company. It has its origins in 1960s psychedelia, but given that it’s popularly associated with the mass suicide of the followers of Jim Jones in Guyana, perhaps we should find something else. In the sense we use it though, it has been flowing liberally of late with respect to AI, and the hype surrounding it. This series has attempted to peer behind that hype, first by examining the motives behind all that metaphorical Kool-Aid drinking, and then by demonstrating a simple example where the technology does something useful that’s hard to do another way. In that last piece we touched upon perhaps the thing that Hackaday readers should find most interesting, we saw the LLM’s possibility as a universal API for useful functions.

It’s Not What An LLM Can Make, It’s What It Can Do

When we program, we use functions all the time. In most programming languages they are built into the language or they can be user-defined. They encapsulate a piece of code that does something, so it can be repeatedly called. Life without them on an 8-bit microcomputer was painful, with many GOTO statements required to make something similar happen. It’s no accident then that when looking at an LLM as a sentiment analysis tool in the previous article I used a function GetSentimentAnalysis(subject,text) to describe what I wanted to do. The LLM’s processing capacity was a good fit to my task in hand, so I used it as the engine behind my function, taking a piece of text and a subject, and returning an integer representing sentiment. The word “do” encapsulates the point of this article, that maybe the hype has got it wrong in being all about what an LLM can make. Instead it should be all about what it can do. The people thinking they’ve struck gold because they can churn out content slop or make it send emails are missing this.

So we have an LLM, even a small one on our own computer, and looking at it in that light it’s immediately apparent that it can become a function to do almost any processing task, if you wrap the right prompt and API call in a function definition. Of course that’s dangerous, because if I may I would like to coin a new phrase: function slop.

As an example I can call an LLM to do simple numerical addition and it will perform the task, but doing so would be utterly pointless given the existence of the + operator. If you are going to use an LLM to perform a processing function it’s important that it be a function where doing so makes sense, otherwise your function is just function slop. A quick web search tells me that function slop is not yet a thing, so I would like to take this moment to apologise for what I may have unleashed upon the world.

Function slop aside though, using the LLM to do a processing task where it makes sense, shouldn’t be ignored as a useful tool. These things are very good at summarising and categorising information in the way a human might do it, a task that’s often hard in traditional programming, so if the job in hand fits those capabilities then it makes sense to use them.

This has been a three-part series, and unlike Star Wars or The Hitchhikers Guide To The Galaxy, it’s probably going to stay that way. I hope that in our explanation we’ve successfully looked beyond the hype and found something useful in all this. It’s odd though, as the one writing it you might think I would be bubbling over with new ideas, but aside from the previous article’s sentiment analysis I still find myself with not much I find the need to use an LLM for. Which is maybe the point, it’s one thing to know a bit about them, but just because they’re there doesn’t mean you have to use them.

When [OGS Mechanics] got a Mercedes EQC 300 battery-electric car in for repair, it was found to have a bit of a weird issue: after sitting in a garage for a while, its range on battery had suddenly reduced significantly without clear cause. Although the typical response here is to just mark the battery pack as ‘faulty’ and replace the whole unit, [OGS] decided to dig into the pack to see what was going on.

The short version is that this particular battery pack consists of two individual batteries, each with its own BMS, one of which had reported a condition to the master BMS that triggered the ‘replace battery module’ error observed with the scan tool. From this it could also be seen that the first battery was at a 10% state-of-charge (SoC), and the second at 95%, making them incredibly unbalanced. Unfortunately the dealer procedure to rebalance did not work here, with only the second battery wanting to charge even after draining both to the same initial level.

To diagnose the underlying issue in earnest required gently prying open the battery pack like a massive glued-shut smartphone. Going by the theory that it is a software glitch, since the first battery was still at a healthy voltage level, it was decided to manually charge it. With both batteries now fully charged, the BMS for the first battery was then removed to have its memory overwritten with that of a known good BMS module, clearing the ‘replace battery module’ error.

Although in the preview for the next video it’s hinted that there’s also an internal balancing issue in the first battery pack, this could be another symptom of its BMS glitching out. Either way, it would seem that BEVs battery modules are both heavily dependent on software, as well as afflicted by the same throw-away culture that has people just buying a new smartphone when the battery fails.

If you read a headline that signs of intelligent life were found on the moon, you might suspect a hoax. But they are there! Humans have dumped a lot of stuff on the moon, both in person and via uncrewed rockets. So after the apocalypse, what strange things will some alien exo-archaeologist find on our only natural satellite?

The Obvious

Of course, we’ve left parts of rockets, probes, and rovers. Only the top part of the Apollo Lunar Excursion Module left the moon. (See for yourself in the Apollo 17 ascent video below.) The bottoms are still there, along with the lunar rovers and a bunch of other science instruments and tools. There are boots and cameras, as you might expect.

But what about the strange things? As of 2012, NASA compiled a list of all known lunar junk that originated on Earth. The list starts with material from the non-Apollo US programs like the Surveyor and Lunar Prospector missions. Next up is the Apollo stuff, which is actually quite a bit: an estimated 400,000 pounds, we’ve heard. This ranges from the entire descent stage and lunar overshoes to urine bags. There are even commemorative patches and a gold olive branch.

After that, the list shows what’s known to be on the surface from the Russian space program, along with objects of Chinese, Indian, Japanese, and European origin.

The Sentimental

Charles Duke on Apollo 16 left a framed family photo on the Moon’s surface with an inscription on the back. We figure if you go looking for it now, the sun will have bleached it white, but we appreciate the sentiment.

There are several objects meant to commemorate fallen astronauts and cosmonauts, including an Apollo 1 mission patch. You may recall that a fire during training killed all three of Apollo 1’s crew.

Lunar Prospector brought a portion of the ashes of Gene Shoemaker, a geologist who trained Apollo astronauts, to the moon. The capsule of ashes holds a quote from Romeo and Juliet:

And, when he shall die

Take him and cut him out in little stars

And he will make the face of heaven so fine

That all the world will be in love with night,

And pay no worship to the garish sun.

To date, Shoemaker is the only person who has remains on the moon.

While not exactly sentimental, NASA did send a silicon disc to the moon with Apollo 11 containing goodwill messages from 73 countries. The whole thing is about the size of a US half dollar, so if you want to read the messages, you might be better off reading the associated document.

Making tiny silicon wafers with finely-detailed etchings was pretty high tech in the late 1960s. GCA Corp used a reduction camera to make a negative photomask containing all the letters plus an inscription around its edge at its final size. This mask was given to Sprague, who etched it.

The Odd

One of the strange things on the NASA list is a falcon feather. That was left by Apollo 15’s Davis Scott, who carried out the classic experiment of dropping a feather and a hammer to note that they fell at the same speed, even in the weak gravity of the moon. The feather was from Baggin, the Air Force Academy’s mascot, and remains on the lunar surface today.

Speaking of Baggin, there are 96 bags of human waste sitting up there. Probably best not to bring that up the next time you and your partner are gazing at the romantic moon overhead.

The Unconfirmed

Forrest Myers created a small ceramic wafer with tiny artwork from six artists, like Andy Warhol, titled “Moon Museum.” The tile features six drawings, including a stylized “AW” (Warhol), a line (Robert Rauschenberg), a black square (David Novros), a diagram (John Chamberlain), Mickey Mouse (Claes Oldenburg), and an interlocking design (Myers). Apparently, Novros and Chamberlain were inspired by circuit diagrams of some kind.

Bell Labs created the wafer. However, NASA failed to approve the project, and Myers sought an alternative.

Reportedly, Myers gave the chip to an unnamed Apollo 12 engineer who affixed it to the leg of the lunar module. However, NASA has not confirmed this, so we don’t know for sure if it is up there or not. Perhaps if you get to the neighborhood, you can check it out and let us know?

To the Dump

You might wonder why so much stuff was left, but if you think about it, it makes sense. The rockets can only bring back so much stuff. Every camera you leave behind means more moon rocks you can bring home. You can buy a new camera, but you can’t buy more moon rocks.

According to the Lunar Legacy Project, Apollo 11, and presumably the other missions, had designated toss zones. (We guess “dumps” didn’t sound good.)

If you are looking for a more up-to-date list, the Wikipedia article can help fill in the gaps, at least for vehicles. There’s been quite a bit added since the NASA list, including items from the UAE, Israel, and Luxembourg. Plus, there are many new additions from other countries.

With the advent of high-resolution orbital cameras, you can see some of the landing sites better than ever. For example, the video below shows the Apollo 17 site imaged by the Lunar Reconnaissance Orbiter Camera.

Of course, we are on our way back to the moon, and so are other space programs. So there will probably be even more human debris on the moon soon. It is only a matter of time before lunar waste management becomes a hot topic.

Title image “Map of artificial objects on the Moon” by [Footy2000]

If you wanted to host a website, you could use any one of a number of online services, or spin up a server on a spare computer at home. If you’re a bit more daring, you could also do what [Tech1k] did, and run one on an ESP32 microcontroller.

The site in question is available (or at least, should be) at HelloESP.com. The first revision ran entirely on an ESP32, serving pages from a SPIFFS filesystem. The device was also fitted with a BME280 environment sensor and an OLED screen. It had an uptime of 500 days before the board failed.

The site has since been relaunched, running on a board that is framed on [Tech1k]’s wall. It runs on an ESP32-WROOM-32D, paired with a BME280 again, along with a CCS811 CO2 and air quality sensor and a DS3231 RTC for accurate timekeeping. The ESP32 is setup to hold an outbound WebSocket to a Cloudflare worker, with the Worker routing HTTP requests to the site via that route. This avoids the need for port forwarding for the ESP32 to be visible to the outside world, and the Cloudflare Worker will also serve a static version of the page in the case of WiFi dropouts or other temporary failures.

It’s true that this isn’t a completely unheard of project—microcontrollers have been working as simple web servers for a long time now. Still, [Tech1k] did a great job of making this as robust as possible and more like a real functional webserver rather than just something that runs on a local network to serve up a config page. That’s worthy of note.

You can run webservers on all kinds of chips these days, even the Raspberry Pi Pico. If you’re doing web stuff on something weird, you know we always wanna hear about it on the tipsline!

There’s a reason that the standards specifications for various wireless communications protocols are extremely long and detailed. [Made by Dennis] found this out first hand when he decided to build a wireless button from scratch.

The major issues with wireless devices is one of power consumption. If reliable power is available from a wall plug or solar panel, this isn’t as serious of a concern. But [Dennis] is using batteries for his buttons, so minimizing power consumption is a priority. He’s going with the nRF52, a microcontroller designed for low power and which has a built in wireless radio, and configuring it in a way that uses the least amount of energy possible.

From there, [Dennis] turns to the wireless communication. He goes into detail about how the microcontroller is woken up, how it sends its data packets to another wireless-enabled microcontroller, and how they handle handshakes and acknowledgements of data. For something as simple as a button press, it gets quickly more complicated especially when adding some basic encryption and security to the communications protocol.

With all the design decisions out of the way, the system can be built. [Dennis] has created custom PCBs for his devices, and also included some expansion I/O for other sensors and peripherals beyond just a pushbutton. All of the schematics and code are available on the project’s GitHub page and the STL files can be found at Printables.

For those new to offline home automation or who are turning away from cloud-based services lately, there are some easy entry points that don’t require much extra hardware or expenditure.