In today’s “News from the Dystopia” segment, we have a story about fighting retail theft with drones. It centers on Flock Safety, a company that provides surveillance technologies, including UAVs, license plate readers, and gunshot location systems, to law enforcement agencies. Their flagship Aerodome product is a rooftop-mounted dock for a UAV that gets dispatched to a call for service and acts as an eye-in-the-sky until units can arrive on scene. Neat idea and all, and while we can see the utility of such a system in a first responder situation, the company is starting to market a similar system to retailers and other private sector industries as a way to contain costs. The retail use case, which the story stresses has not been deployed yet, would be to launch a drone upon a store’s Asset Protection team noticing someone shoplifting. Flock would then remotely pilot the drone, following the alleged thief back to their lair or hideout and coordinating with law enforcement, who then sweep in to make an arrest.

Police using aerial assets to fight crime is nothing new; California has an entire entertainment industry focused on live-streaming video from police chases, after all. What’s new here is that these drones lower the bar for getting aerial support into the mix. At a $1,000 per hour or more to operate, it’s hard to justify sending a helicopter to chase down a shoplifter. Another objection is that these drones would operate entirely for the benefit of private entities. One can certainly make a case for a public interest in reducing retail theft, since prices tend to increase for everyone when inventory leaves the store without compensation. But we don’t know if we really like the idea of being tailed home by a drone just because a minimally trained employee on the Asset Protection team of BigBoxCo is convinced a crime occurred. It’s easy enough to confuse one person for another or to misidentify a vehicle, especially on the potato-cams retailers seem to love using for their security systems. We also really don’t like one of the other markets Flock is targeting: residential HOAs. The idea of neighborhoods being patrolled by drones and surveilled by license plate cameras is a bridge too far, at least to our way of thinking.

Are you old enough to remember when having access to a T1 line was a true mark of geek cachet? We sure are, and in a time when the plebes were stuck with 9,600-baud dial-up over their POTS lines, working on a T1 line was a dream come true. Such was the allure that we can even recall apartment complexes in the tech neighborhoods outside of Boston listing T1 lines among their many amenities. It was pretty smart marketing, all things considered, especially compared to the pool you could only use three months a year. But according to a new essay by J. B. Crawford over at “Computers Are Bad”, T1 lines were actually pretty crappy, even in the late 90s and early 2000s. The article isn’t just dunking on T1, of course, but rather a detailed look at the whole T-carrier system, which can trace its roots back to the 1920s with Bell’s frequency-division multiplexing trunking systems. T1 was an outgrowth of those trunking systems, intended to link central offices but evolving to service customers on the local loop. Fascinating stuff, as always, especially the bit about replacing the loading coils that were used every 6,600′ along trunk lines to compensate for capacitance with repeaters.

We’ve heard of bricking a GPU, but ordering a GPU and getting a brick instead is something new. A Redditor who ordered an RTX 5080 from Amazon was surprised to find a plain old brick in the package instead. To be fair, whoever swiped the card was kind enough to put the brick in the original antistatic bag; one can’t be too careful, after all. The comments on the Reddit post have a good selection of puns — gigabricks, lol — and good fun was had by all, except perhaps for the unfortunate brickee. The article points out that this might not be a supply chain issue, such as the recent swap of a GPU for a backpack, which, given the intact authentication seals, was likely done at the factory. In this case, it seems like someone returned the GPU after swapping it out for the brick, assuming (correctly, it would seem) that Amazon wouldn’t check the contents of the returned package beyond perhaps weighing it. How the returned inventory made it back into circulation is a bit of a mystery; we thought returned items were bundled together on pallets and sold off at auction.

Speaking of auctions, someone just spent almost half a million bucks on one of the nine estimated remaining wooden-cased Apple I computers. It’s a lovely machine, to be sure, with its ByteShop-style wooden case intact and in excellent shape. The machine is still working, too, which is a nice plus, but $475,000? Even with a Dymo embossed label in Avocado Green — or is that Harvest Gold? — that seems a bit steep. There’s apparently some backstory to the machine that lends to its provenance, including former ownership by the first female graduate of Stanford Law School, June Blodgett Moore. This makes it the “Moore Apple-1” in the registry (!) for these machines, only 50 of which were ever made. One wonders if the registry makes allowance for basic maintenance of vintage electronics like these machines; does routine recapping affect their value?

And finally, continuing with the vintage theme, we’ve been following the adventures of [Buy It Fix It] over on YouTube as he attempts to revive a Williams Defender arcade machine from the 1980s. We remember this game well, having fed far too many quarters into the one at the Crazy 8s Pool and Arcade back in the day. This machine is in remarkably good shape for being over 40 years old, but it still needed some TLC to get it running again. The video documents a series of cascading failures and maddening intermittent faults, requiring nearly every tool in his kit to figure out. At the end of the second video, [Buy It] reckons he put 60 hours into the repair, a noble effort with fantastic results. Enjoy!

youtube.com/embed/IpuPvdyxeW0?…

youtube.com/embed/PyCM_PB92r4?…

hackaday.com/2025/09/28/hackad…

Over the years, we’ve brought you many stories of the creative artwork behind electronic event badges, but today we may have a first for you. [Spencer] thinks nobody before him has made a badge powered by a Z80, and we believe he may be right. He’s the originator of the RC2014 Z80-based retrocomputer, and the badge in question comes from the recent RC2014 Assembly.

Fulfilling the function of something you can write your name on is a PCB shaped like an RC2014 module, with LEDs on all the signal lines. It could almost function as a crude logic analyser for the system, were the clock speed not far too high to see anything. To fix this, [Spencer]’s badge packs a single-board RC2014 Micro with a specially slow clock, and Z80 code to step through all memory addresses, resulting in a fine set of blinkenlights.

Thus was created the first Z80-based event badge, and we’re wondering whether or not it will be the last. If you’re curious what this RC2014 thing is about, we reviewed the RC2014 Micro when it came out.

hackaday.com/2025/09/28/decora…

These days, video cards are virtually supercomputers. When they aren’t driving your screen, they are decoding video, crunching physics models, or processing large-language model algorithms. But it wasn’t always like that. The old video cards were downright simple. Once PCs gained more sophisticated buses, video cards got a little better. But hardware acceleration on an old-fashioned VGA card would be unworthy of the cheapest burner phone at the big box store. Not to mention, the card is probably twice the size of the phone. [Bits and Bolts] has a look at several old cards, including a PCI version of the Tseng ET4000, state-of-the-art of the late 1990s.

You might think that’s a misprint. Most of the older Tseng boards were ISA, but apparently, there were some with the PCI bus or the older VESA local bus. Acceleration here typically meant dedicated hardware for handling BitBlt and, perhaps, a hardware cursor.

It is fun watching him test these old cards and work on them under the microscope, too. Since the PCI bus was new when this board was introduced, it apparently had some bugs that made it incompatible with certain motherboards.

We recall being blown away by the color graphics these boards provided when they were new. Now, of course, you wouldn’t see graphics like this even on a cheap video game. Still, fun to take a walk down memory lane with these old boards.

[Bits and Bolts] definitely has a hobby. We love that these were high-tech in their day, but now designing a VGA card is well within reach for anyone adept at using FPGAs.

youtube.com/embed/2aapI_uyvII?…

hackaday.com/2025/09/28/a-walk…

Dato che si inizia a parlare incessantemente di robot umanoidi e intelligenza artificiale, gli hacker hanno voluto dare una guardatina a questa nuova tecnologia che sempre di più invaderà lo spazio del nostro futuro. E non è andata bene.

Il 27 settembre 2025 sono emerse nuove preoccupazioni riguardo ai robot prodotti dalla cinese Unitree Robotics, dopo la segnalazione di serie vulnerabilità che potrebbero esporre migliaia di dispositivi a rischi di controllo remoto e utilizzo malevolo.

Secondo quanto riportato da IEEE Spectrumgiovedì 25 settembre, i ricercatori hanno individuato una falla critica nel sistema Bluetooth Low Energy (BLE) utilizzato dai robot dell’azienda per la configurazione iniziale della rete WiFi. Tale debolezza consentirebbe a un aggressore di ottenere i privilegi di root sul sistema Android dei dispositivi, acquisendone il controllo totale.

Il ricercatore di sicurezza Andreas Makris ha spiegato che, una volta compromesso un robot, l’infezione può diffondersi automaticamente ad altri dispositivi Yushu nel raggio di azione del Bluetooth, trasformandoli in una botnet in grado di replicarsi senza intervento umano.

Il meccanismo di autenticazione appare particolarmente fragile: i robot Unitree consentono l’accesso semplicemente attraverso la crittografia di una stringa hardcoded, “unitree”. In questo modo un attaccante potrebbe inserire codice arbitrario mascherato da SSID e password della rete WiFi. Nel momento in cui il robot tenta la connessione, il codice verrebbe eseguito con privilegi di amministratore, senza alcuna verifica aggiuntiva.

Makris ha aggiunto che un exploit di questo tipo potrebbe persino impedire l’aggiornamento del firmware da parte dell’utente, lasciando i dispositivi permanentemente vulnerabili e aprendo la strada a un controllo di massa. Tra i modelli interessati risultano i cani robot quadrupedi Go2 e B2 e i robot umanoidi G1 e H1. È la prima volta che una falla di tale portata viene resa pubblica su una piattaforma robotica umanoide commerciale.

I ricercatori hanno contattato Unitree Robotics già a maggio 2025, ma dopo diversi tentativi di comunicazione senza risultati, l’azienda avrebbe smesso di rispondere lo scorso luglio. La mancanza di collaborazione ha spinto alla divulgazione pubblica della vulnerabilità. Makris ha ricordato inoltre di aver già identificato in passato una backdoor nel modello Yushu Go1, sollevando dubbi sull’origine di tali falle: se siano frutto di negligenza nello sviluppo o di implementazioni intenzionali.

Un’ulteriore segnalazione è arrivata da Victor Mayoral-Vilches, fondatore di Alias Robotics, secondo cui i robot Yushu inviano ai server cinesi dati di telemetria che potrebbero includere informazioni audio, video e spaziali. Mayoral-Vilches ha evidenziato come questi dispositivi siano ampiamente diffusi a livello globale, ma molti utenti non siano consapevoli dei rischi legati al loro utilizzo. In attesa di risposte ufficiali, l’esperto consiglia agli utilizzatori di collegare i robot esclusivamente a reti WiFi isolate e di disattivarne la connettività Bluetooth come misura di protezione immediata.

Le preoccupazioni non riguardano solo la sfera privata. Ad agosto 2025, la città di Taipei ha impiegato il modello Go2 per attività di pattugliamento urbano, scelta che ha suscitato interrogativi sulla sicurezza dei dati. Già il 5 maggio 2025 la Commissione speciale della Camera dei Rappresentanti degli Stati Uniti sulla concorrenza strategica con la Cina aveva inviato una lettera al Segretario alla Difesa, al Segretario al Commercio e al presidente della Federal Communications Commission, avvertendo che Yushu “rappresenta una minaccia crescente per la sicurezza nazionale”.

Secondo quanto riportato, i robot dell’azienda sarebbero già stati adottati in contesti sensibili come prigioni, corpi di polizia e basi militari statunitensi. La presenza di backdoor e la possibilità di sorveglianza remota hanno spinto alcuni osservatori a definirli “cavalli di Troia con telecamere”.

Ad oggi, Unitree Robotics non ha rilasciato alcun commento ufficiale.

L'articolo Falle critiche nei robot cinesi. Una bonet di robot zombie può essere controllata a distanza proviene da il blog della sicurezza informatica.

We are proud to sponsor The Abolition and Alternatives Conference (AAC) on October 3–5. The conference is organized and hosted by The Black Response at their offices at 245 Main Street, Cambridge, MA, 02142 on Friday and Saturday. On Sunday, it will be at The Foundry – 101 Roger Street Cambridge, MA – Kendall Square. The conference schedule is available.

We encourage all Pirates to attend and support this conference, especially, but not exclusively, the ShotSpotter and Police Surveillance track. If you can not attend, or even if you can, please consider giving a donation to The Black Response or print out their poster and put it up in your neighborhood. See you next week!

Details on the conference are reproduced below. Edits are only for clarity:

This free, in-person event will bring together community members, organizers, and advocates for a weekend of in-depth learning and discussion focused on alternative public safety and community care, housing justice, and the impacts of surveillance technologies like ShotSpotter. It will include keynote addresses from Fatema Ahmad (Muslim Justice League), Stephanie Guirand (The Black Response), and Spencer Piston (Boston University).

Food will be provided, childcare will be available, and we encourage attendees to share any additional access needs via the conference interest form. TBR will be reaching out to invite participation as speakers and facilitators. For questions, please contact Stephanie at general@theblackresponsecambridge.com.

Throughout the conference, participants will have the opportunity to choose from panels in four tracks:

Housing Justice

This track features panels led by the Cambridge Housing Justice Coalition (CHJC). CHJC is a coalition of activist groups and concerned Cambridge residents who believe housing is a basic human right. The panels and workshops on this track will focus on housing justice and its intersections with the prison industrial complex.

ShotSpotter and Police Surveillance

This track will be led by the #StopShotSpotter Coalition Camberville. In this track, coalition members will provide an introduction to ShotSpotter, the audio-surveillance technology. We will examine its impact in Cambridge, the national landscape, and broader conversations about surveillance tech.

Alternatives and Community Care

This track will be led by members of the Massachusetts Community Care Network (MCCN). This track will include panels of responders, program directors, and organizers working to make alternatives to policing real. It includes a panel on the movement with Daanika Gordon, Spencer Piston, and Minali Aggarwal.

Community Concerns (Anti-Racism, Immigration Justice, Justice for Palestine, and Black Lives Matter)

This track will discuss concerns that come directly from the communities we serve and work with. These concerns also intersect with the movement for abolition and alternatives. They include Justice for Palestine, Immigration Justice, and Anti-Racism. In this tract we intend to learn from organizers leading these movements in Massachusetts.

masspirates.org/blog/2025/09/2…

[Simone]’s AI assistant, dubbed Max Headbox, is a wakeword-triggered local AI agent capable of following instructions and doing simple tasks. It’s an experiment in many ways, but also a great demonstration not only of what is possible with the kinds of open tools and hardware available to a modern hobbyist, but also a reminder of just how far some of these software tools have come in only a few short years.

Max Headbox is not just a local large language model (LLM) running on Pi hardware; the model is able to make tool calls in a loop, chaining them together to complete tasks. This means the system can break down a spoken instruction (for example, “find the weather report for today and email it to me”) into a series of steps to complete, utilizing software tools as needed throughout the process until the task is finished.

Watch Max in action in the video (also embedded just below). Max is a little slow, but not unusably so. As far as proofs of concept go, it demonstrates that a foundation for such systems is perfectly feasible on budget hardware running free, locally installed software. Check out the GitHub repository.

The name is, of course, a play on Max Headroom, the purportedly computer-generated TV personality of the ’80s who was actually an actor in a mask, just like the person behind what was probably the most famous broadcast TV hack of all time (while wearing a Max Headroom mask).

Thanks to [JasonK] for the tip!

youtube.com/embed/ZnuXIrbRQiE?…

hackaday.com/2025/09/28/fully-…

Dalla fine di luglio 2025 è stata registrata una nuova ondata di attacchi informatici che colpisce le organizzazioni dotate di firewall SonicWall, con la diffusione attiva del ransomware Akira.

Secondo i ricercatori di Arctic Wolf Labs, l’attività malevola ha subito un incremento significativo e continua a persistere. Gli aggressori ottengono l’accesso iniziale attraverso connessioni VPN SSL compromesse, riuscendo a eludere l’autenticazione multifattore (MFA). Una volta entrati nella rete, passano rapidamente alla fase di crittografia: in alcuni casi, il tempo di permanenza prima del rilascio del ransomware è stato di appena 55 minuti.

La vulnerabilità sfruttata e il ruolo delle credenziali rubate

Gli accessi abusivi sono stati collegati al CVE-2024-40766, una vulnerabilità di controllo degli accessi resa pubblica nel 2024. La principale ipotesi è che i criminali abbiano raccolto in passato credenziali da dispositivi esposti e vulnerabili, ora sfruttate anche contro apparati già aggiornati. Questo spiega perché siano stati compromessi sistemi completamente patchati, circostanza che inizialmente aveva alimentato l’ipotesi di un nuovo exploit zero-day.

Un altro elemento critico riguarda la MFA OTP di SonicWall: gli attaccanti sono riusciti ad autenticarsi anche con account protetti da questa funzione, aumentando la gravità della campagna.

Tecniche e strumenti utilizzati

Una volta ottenuto l’accesso tramite VPN SSL, gli aggressori:

• avviano la scansione della rete interna per identificare porte esposte come SMB (445), RPC (135) e SQL (1433);
• utilizzano strumenti di ricognizione e movimento laterale tra cui Impacket, SoftPerfect Network Scanner e Advanced IP Scanner;
• creano nuovi account amministrativi e innalzano i privilegi di quelli già presenti;
• installano software di accesso remoto come AnyDesk, TeamViewer e RustDesk per garantire la persistenza;
• stabiliscono connessioni nascoste tramite SSH inverso e Cloudflare Tunnels.

Per ridurre le possibilità di rilevamento, gli operatori della minaccia tentano di disattivare le soluzioni di sicurezza degli endpoint, come Windows Defender ed EDR.

In alcuni casi ricorrono alla tecnica BYOVD (bring your own vulnerable driver) per compromettere i sistemi a livello di kernel ed eliminano le copie shadow del volume per impedire eventuali ripristini.

Dalla raccolta dei dati al rilascio del ransomware

Prima di avviare la crittografia, gli attaccanti esfiltrano informazioni riservate: i file vengono compressi con WinRAR ed estratti attraverso strumenti come rclone e FileZilla. Successivamente distribuiscono il ransomware Akira, tramite file eseguibili denominati akira.exe o locker.exe, crittografando le unità di rete e avviando la richiesta di riscatto.

Raccomandazioni per le organizzazioni

Gli esperti di Arctic Wolf invitano tutte le realtà che utilizzano dispositivi SonicWall ad adottare misure immediate.

In particolare, si raccomanda di reimpostare le credenziali delle VPN SSL, inclusi gli account collegati ad Active Directory, soprattutto se i sistemi in passato hanno eseguito firmware vulnerabili a CVE-2024-40766. La semplice applicazione delle patch non è considerata sufficiente se le credenziali sono già state compromesse.

L'articolo Ransomware Akira: una nuova campagna colpisce i firewall SonicWall proviene da il blog della sicurezza informatica.