Where Is The End Of DIY?

October 5, 2024 by Elliot Williams 70 Comments

Al and I were talking on the podcast about Dan Maloney’s recent piece on how lead and silver are refined and about the possibility of anyone fully understanding a modern cellphone. This lead to Al wondering at the complexity of the constructed world in which we live: If you think hard enough about anything around you right now, you’d probably be able to recreate about 0% of it again from first principles.

Smelting lead and building a cellphone are two sides of coin, in my mind. The process of getting lead out of galena is simple enough to comprehend, but it’s messy and dangerous in practice. Cellphones, on the other hand, are so monumentally complex that I’d wager that no single person could even describe all of the parts in sufficient detail to reproduce them. That’s why they’re made by companies with hundreds of engineers and decades of experience with the tech – the only way to build a cellphone is to split the complicated task into many subsystems.

Smelting lead is a bad DIY project because it’s simple in principle, but prohibitive in practice. Building a cellphone from the ground up is incomprehensible in principle, but ironically entirely doable in practice if you’re willing to buy into some abstractions.

Indeed, last week we saw a nearly completely open-source build of a simple smartphone, and the secret to making it work is knowing the limits of DIY. The cell modem, for instance, is a black box. It’s an abstract device that you can feed data to and read data from, and it handles the radio parts of the phone that would take forever to design from scratch. But you don’t need to understand its inner workings to use it. Knowing where the limits of DIY are in your project, where you’re willing to accept the abstraction and move on, can be critical to getting it done.

Of course, in an ideal world, you’d want the cell modem to be like smelting lead – something that’s possible to understand in principle but just not worth DIYing in practice. And of course, there are some folks out there who hack on cell modem firmware and others who could do the radio engineering. But despite my strong DIY urges, I’d have to admit that the essential complexity of the module simply makes it worth treating as a black box. It’s very probably the practical limit of DIY.

What’s The Deal With AI Art?

September 28, 2024 by Elliot Williams 103 Comments

A couple weeks ago, we had a kerfuffle here on Hackaday: A writer put out a piece with AI-generated headline art. It was, honestly, pretty good, but it was also subject to all of the usual horrors that get generated along the way. If you have played around with any of the image generators you know the AI-art uncanny style, where it looks good enough at first glance, but then you notice limbs in the wrong place if you look hard enough. We replaced it shortly after an editor noticed.

The story is that the writer couldn’t find any nice visuals to go with the blog post, with was about encoding data in QR codes and printing them out for storage. This is a problem we have frequently here, actually. When people write up a code hack, for instance, there’s usually just no good image to go along with it. Our writers have to get creative. In this case, he tossed it off to Stable Diffusion.

Some commenters were afraid that this meant that we were outsourcing work from our fantastic, and very human, art director Joe Kim, whose trademark style you’ve seen on many of our longer-form original articles. Of course we’re not! He’s a genius, and when we tell him we need some art about topics ranging from refining cobalt to Wimshurst machines to generate static electricity, he comes through. I think that all of us probably have wanted to make a poster out of one or more of his headline art pieces. Joe is a treasure.

But for our daily blog posts, which cover your works, we usually just use a picture of the project. We can’t ask Joe to make ten pieces of art per day, and we never have. At least as far as Hackaday is concerned, AI-generated art is just as good as finding some cleared-for-use clip art out there, right?

Except it’s not. There is a lot of uncertainty about the data that the algorithms are trained on, whether the copyright of the original artists was respected or needed to be, ethically or legally. Some people even worry that the whole thing is going to bring about the end of Art. (They worried about this at the introduction of the camera as well.) But then there’s also the extra limbs, and AI-generated art’s cliche styles, which we fear will get old and boring after we’re all saturated with them.

So we’re not using AI-generated art as a policy for now, but that’s not to say that we don’t see both the benefits and the risks. We’re not Luddites, after all, but we are also in favor of artists getting paid for their work, and of respect for the commons when people copyleft license their images. We’re very interested to see how this all plays out in the future, but for now, we’re sitting on the sidelines. Sorry if that means more headlines with colorful code!

Against Elitism

September 21, 2024 by Elliot Williams 104 Comments

A while back we got an anonymous complaint that Hackaday was “elitist”, and that got me thinking. We do write up the hacks that we find the coolest, and that could lead to a preponderance of gonzo projects, or a feeling that something “isn’t good enough for Hackaday”. But I really want to push back against that notion, because I believe it’s just plain wrong.

One of the most important jobs of a Hackaday writer is to find the best parts of a project and bring that to the fore, and I’d like to show you what I mean by example. Take this post from two weeks ago that was nominally about rescuing a broken beloved keyboard by replacing its brain with a modern microcontroller. On its surface, this should be easy – figure out the matrix pinout and wire it up. Flash in a keyboard firmware and you’re done.

Of course we all love a good hardware-rescue story, and other owners of busted Sculpt keyboards will be happy to see it. But there’s something here for the rest of us too! To figure out the keyboard matrix, it would take a lot of probing at a flat-flex cable, so [TechBeret] made a sweet breakout board that pulled all the signals off of the flat-flex and terminated them in nicely labelled wires. Let this be your reminder that making a test rig / jig can make these kind of complicated problems simpler.

Once the pinout was figured out, and a working prototype made, it was time to order a neat PCB and box it up. The other great trick was the use of 3D-printed mockups of the PCBs to make sure that they fit inside the case, the holes were all in the right places, and that the flat-flex lay flat. With how easily PCB design software will spit out a 3D model these days, you absolutely should take the ten minutes to verify the physical layout of each revision before sending out your Gerbers.

So was this a 1337 hack? Maybe not. But was it worth reading for these two sweet tidbits, regardless of whether you’re doing a keyboard hack? Absolutely! And that’s exactly the kind of opportunity that elitists shut themselves off from, and it’s the negative aspect of elitism what we try to fight against here at Hackaday.

Fun And Failure

September 7, 2024 by Elliot Williams 20 Comments

My sister is a beekeeper, or maybe a meta-beekeper. She ends up making more money by breeding and selling new queen bees to other beekeepers than she does by selling honey, but that doesn’t mean that she doesn’t also process the sweet stuff from time to time. She got a free steam-heated oscillating hot knife, used for cutting the waxy caps off of the tops of the cells before spinning the combs down to extract honey, and she thought it might be easier to use than her trusty hand-held electric hot knife.

The oscillating knife, which was built something like a century ago, hadn’t been used in decades. All of the grease had turned to glue, and the large v-belt wheel that made it go was hard to turn by hand, and the motor was missing anyway. So she gave it to my father and me as a project. How could we resist?

We found the original manual on the Internet, which said that it would run from any 1/2 hp motor, or could be optionally driven by a takeoff wheel from a tractor – unfortunately not an option in my sister’s honey house. But we did find a 3/4 hp bench grinder at Harbor Freight that conveniently fit inside the case, and bought the smallest v-belt pulley wheel that would fit the grinder’s arbor. We thought we were geniuses, but when we hooked it all up, it just stalled.

We spent more than a few hours taking the mechanism apart. It was basically an eccentric shaft with a bearing on the end, and the bearing ran back and forth in the groove of a sliding mechanism that the knife blade attached to. As mentioned above, everything was gunked, so we took it all apart. The bearing was seized, so we freed that up by getting the sand out of the balls. The bearing couldn’t move freely in the slide either, but we filed that down until it just moved freely without noticeable play. We added grease from this century, and reassembled it. It turned fine by hand.

But with the belt and motor attached, the mechanism still had just enough friction to stall out the motor. Of course we wrapped some rope around the shaft and pull-started it, and it made a hell of a racket, nearly vibrated itself off the table, and we could see that the marvelous zinc-coated frame that held it all together was racking under the tension. It would require a wholly new housing to be viable, and we hadn’t even figured out a source of steam to heat the knife.

In short, it was more trouble than it was worth. So we packed up the bench grinder in the original container, and returned it no-worse-for-wear to the Freight. But frankly, we had a fantastic time playing around with a noble machine from a long-gone past. We got it “working” even if that state was unworkable, and we were only out the cost of the small v-belt pulley. Who says all of your projects have to be a success to be fun?

Your Data In The Cloud

August 17, 2024 by Elliot Williams 33 Comments

I try not to go off on security rants in the newsletter, but this week I’m unable to hold back. An apparent breach of a data aggregator has resulted in a monster dataset of US, UK, and Canadian citizens names, addresses, and social security numbers. As a number of reports have pointed out, the three billion records in the breach likely contain duplicate individuals, because they include all the addresses where you’ve lived, and there have only been on the order of 450 million US social security numbers issued anyway.

But here’s the deal. Each of these data aggregators, and each of the other companies that keep tons of data on you, are ticking time bombs. Maybe not every one of them gets breached, but there’s certainly enough incentive for the bad guys to try to do so. (They are looking to sell the NPD dataset mentioned above for $3.5 million.)

My gut feeling is that eventually all of the information on everyone will be released. Maybe then it will cease to be interesting to new crops of crooks, because there’s nothing new to learn.

On the other hand, the sheer quantity of identity thefts that this, and future breaches, will unleash on us all is mind-boggling. In the case of legitimate data aggregators like this one, requesting to have had your data out of their dataset appears to have been a viable defense. But for every one legit operator, there are others that simply track you. When they get hacked, you lose.

This breach is likely going to end in a large lawsuit against the company in question, but it almost certainly won’t be big enough to cover the damage to everyone in the affected countries. Is it time that companies that hold large datasets will have to realize that the data is a liability as well as an asset?

How About Privacy and Hackability?

August 3, 2024 by Elliot Williams 32 Comments

Many smart electric meters in the US use the 900 MHz band to broadcast their usage out to meter readers as they walk the neighborhood. [Jeff Sandberg] used an RTL-SDR dongle and some software to integrate this data into his own home automation system, which lets him keep track of his home’s power usage.

Half of the comment section was appalled that the meters broadcast this data in the clear, and these readers thought this data should be encrypted even if the reach is limited to the home-owner’s front yard. But that would have stopped [Jeff] from accessing his own data as well, and that would be a shame. So there’s clearly a tradeoff in play here.

We see this tradeoff in a lot of hardware devices as well – we want to be able to run our firmware on them, but we don’t want criminals to do the same. We want the smart device to work with the cloud service, but to also work with our own home automation system if we have one. And we want to be able to listen in to our smart meters, but don’t necessarily want others to do so.

The solution here is as easy as it is implausible that it will get implemented. If the smart meters transmitted encrypted, each with their own individual password, then everyone would win. The meter reader would have a database of passwords linked to meter serial numbers or addresses, and the home owner could just read it off of a sticker, optimally placed on each unit. Privacy and usability would be preserved.

This issue isn’t just limited to electric meters. Indeed, think of all of the data that is being sent out from or about you, and what percentage of it is not encrypted and should be, but also about what data is sent out encrypted that you could use access to. The solution is to put you in control of the encryption, by selecting a password or having access to one that’s set for you. Because after all, if it’s your data, it should be your data: private and usable.