A cyberattack on supermarket refrigeration systems overnight from October 11 to 12 forced equipment at hundreds of French stores into prolonged defrost cycles, spoiling fresh and frozen food. France’s Paris prosecutor has opened an investigation into the losses and reported fires. A final nationwide count and damage total are still unavailable, according to [AFP’s October 13 report](https://www.afp.com/en/french-supermarkets-suffer-refrigerator-shutdown-after-unprecedented-cyberattack). In its [October 13 report](https://www.lemonde.fr/en/economy/article/2026/10/13/a-disastrous-day-for-french-supermarkets-after-massive-cyberattack-targets-refrigeration-systems_6758535_19.html), Le Monde put the affected footprint at around 300 Intermarché and Netto stores, about 100 Coopérative U shops, 80 Carrefour stores, several E.Leclerc locations and roughly 30 Monoprix stores. [Reuters, reporting the previous day](https://www.reuters.com/business/retail-consumer/french-supermarket-chain-intermarche-hit-by-cyberattack-150-stories-targeted-2026-10-12/), cited a lower Intermarché estimate and Monoprix’s figure of 5% of its roughly 490 French shops. These are dated, attributed snapshots rather than a final consolidated count. Le Monde also reported that wholesaler Metro was affected through an external service provider, while Metro said it checked that products remained safe; Picard said it was not affected. Some stores discarded all affected refrigerated stock to protect the cold chain. Le Monde reported that the national loss was not yet known and could reach several million euros or more, an estimate rather than a confirmed total. One shop owner it interviewed put his own loss at €21,000 at purchase cost. Several stores had to clear shelves or close temporarily; the reporting does not establish a national food shortage. The technical picture has sharpened. Le Monde reported that the attack involved Danfoss refrigeration-control software and that some settings were changed to raise temperatures and lengthen defrost cycles, delaying alarms. Danfoss told the newspaper that most affected units were obsolete and no longer received software or security updates. Separately, CERT-FR says it knows of numerous compromises of Danfoss AK-SM and AK-SC controllers, and urges operators to disconnect or tightly filter access. For suspected compromise it recommends restricting access to trusted service providers, factory resetting, applying fixes, rotating secrets, reviewing accounts and reconfiguring systems. [CERT-FR’s alert](https://cert.ssi.gouv.fr/alerte/CERTFR-2026-ALE-012/) This is more specific than the first reports, but investigators and Danfoss have not publicly identified a particular CVE or confirmed that a previously disclosed flaw was the route used in the supermarket attack. [Danfoss previously issued security advisories for its AK-SM 8xxA controllers](https://www.danfoss.com/en/service-and-support/coordinated-vulnerability-disclosure/danfoss-security-advisories/dsa-2025-03-01/); those disclosures alone do not establish how this incident happened. The attackers’ identity also remains unknown. The Paris prosecutor’s cybercrime office and France’s cybersecurity agency are investigating. For shoppers, the confirmed effects are discarded food and temporary disruption at affected branches. The next answers are whether the compromised controllers share a common exposure, how many sites and other cold-storage businesses were hit, and what the final losses are. Current reporting does not show that consumer data was stolen or that the Dutch retailer Ahold Delhaize incident was connected.

[Anthropic’s October 12 study](https://www.anthropic.com/research/self-replicating-ai) shows that AI models can perform parts of a self-replicating worm’s setup across connected devices, but it does not report an AI worm spreading in the wild. The tests point to a future IoT security risk; they do not show that today’s smart TVs are being infected by AI. The company’s researchers ran one 26-billion-parameter model across 15 inexpensive Android TV boxes, each costing about $55. The devices shared the workload at roughly 0.39 tokens per second. In a separate test, a 9-billion-parameter model got inference running on eight of nine hardware types. These experiments show that modest devices can host model workloads when linked together, not that the devices were compromised. For the end-to-end replication test, Anthropic used eight Android emulators in a firewalled Google Cloud network. GLM-5.2 reached all eight nodes in four out of five trials and seven in the fifth. But the researchers deliberately simplified initial access: the emulator network exposed an unauthenticated ADB port. They say this makes the result an upper bound on real-world success. No real consumer device was infected in this test, and the team did not demonstrate a model that was both small enough for the tested TV boxes and reliable enough to complete the whole cycle on its own. Anthropic estimates that small models could reach the tested capability by the end of 2026 if recent improvement continues. That is a trend-based extrapolation, not a confirmed release date or warning that a new worm is imminent. [Al Jazeera’s October 13 report](https://www.aljazeera.com/economy/2026/10/13/ai-could-self-replicate-on-home-devices-within-months-anthropic-says) also emphasizes the controlled setup and quotes an outside AI security expert who described the current result as concerning in trajectory, but not an existential threat. The near-term takeaway is familiar IoT security: devices with weak or outdated protections can be attractive botnet targets, whether malware uses AI or not. Anthropic says defenders could monitor unusually large model downloads, sustained traffic between devices that do not normally communicate, and scanning activity. For owners, keep device software updated, change default passwords and replace hardware that no longer receives security fixes. The research adds a possible layer of adaptability to a longstanding problem; it does not establish an active outbreak. *Conceptual illustration of connected devices; not a photograph or depiction of a real incident.* 

Ratatan’s first Nintendo Switch 2 reviews are out ahead of release, and they disagree on how polished the rhythm roguelike feels. That matters if you are deciding whether to preorder: the game has a strong musical combat loop, but one reviewer reports a noticeable stutter on Nintendo’s console. Pocket Tactics awarded the Switch 2 version 7/10. Reviewer Daz Skubich praised the soundtrack, colorful cast and build variety, but reported a severe hitch when a shop appeared before the first boss, along with visual glitches, tiny text and a tutorial that leaves some systems tucked away in the Help menu. [CGMagazine’s review](https://www.cgmagonline.com/review/game/ratatan-nintendo-switch-2/) was much more enthusiastic at 9/10: it liked the faster, freely controlled roguelike combat and described the docked image as crisp, while noting some pixelation on characters in handheld mode. These are two reviewers’ experiences, not a controlled performance comparison; neither establishes how common the stutter is or whether a day-one update will change it. The standard edition is listed at $34.99, with the publisher advertising a 10% preorder price of $31.49. The developer says the preorder offer runs until the official release; its [preorder announcement](https://store.steampowered.com/news/posts/?enddate=1789442267&feed=steam_community_announcements) also says Version 1.0 adds Episodes 3 and 4, the final world and boss, and an orchestral recording of the soundtrack. [Nintendo’s US store listing](https://www.nintendo.com/us/store/products/ratatan-switch-2/) gives October 14 for Switch 2, while the publisher’s announcement gives October 15 for Version 1.0. Check your local store for the platform-specific release time. Ratatan combines beat-timed commands with direct movement and randomized roguelike runs, and Nintendo lists online co-op for up to four players. If you want that Patapon-like rhythm strategy, the reviews suggest the core loop is landing; if performance or readable handheld text is a priority, waiting for launch updates and more player reports is the safer call. Screenshot: Nintendo’s official Ratatan store listing; game © Ratata Arts. ~Gaming ~Nintendo ~Switch2 

/media/assets/4a2e377b-dfa2-464d-87b6-568174ef83b7/media.mp4 

Nintendo will release a Japan-only Nintendo Switch 2 bundle on November 12 for ¥63,980 (about $403), including a download code for one game chosen from three options. At Japan’s current recommended price of ¥59,980 (about $378) for the console on its own, the bundle adds ¥4,000 (about $25) for the game code. Dollar equivalents use the October 12 USD/JPY rate and are approximate; they are not U.S. retail prices. Buyers can choose Mario Kart World or Nintendo Switch Sports Resort for Switch 2, or Tomodachi Life: Living the Dream for the original Switch. That last option makes this a choice across two generations, rather than three Switch 2 games. Nintendo says the code must be redeemed by March 31, 2027, and downloading requires a Nintendo Account and internet access. [Nintendo’s announcement](https://www.nintendo.co.jp/corporate/release/en/2026/261006.html) lists the full package contents. The regional limitation matters as much as the price. Nintendo’s Japanese-Language System (Japan Only) supports Japanese as its system language, and only accounts set to Japan can be linked. That makes the bundle a poor fit for international buyers looking for a lower-cost way into Switch 2. [Nintendo’s system information](https://www.nintendo.com/jp/en/hardware/switch2/lineup/index.html) spells out those restrictions. The Japanese console’s recommended price rose from ¥49,980 (about $315) to ¥59,980 (about $378) on May 25. The new package adds one game choice for ¥4,000 (about $25) over the current hardware price. [Gematsu reported the bundle announcement](https://www.gematsu.com/2026/10/switch-2-japanese-language-system-choose-your-game-bundle-launches-november-12-in-japan), while [Nintendo’s price notice](https://www.nintendo.co.jp/corporate/release/en/2026/260508.html) documents the increase. The dollar equivalents use the [October 12 USD/JPY rate](https://www.investing.com/currencies/usd-jpy-historical-data). ~Nintendo ~Switch2

Fedora 45 is still a beta as of 12 October 2026. The Fedora Project currently lists 20 October as its target for the final release, but that date can move. For gamers, the useful question is not whether a new Fedora number automatically adds frames per second; it is which parts of the Linux gaming stack are changing, and whether those changes help a particular game and GPU. [Fedora’s beta announcement](https://fedoramagazine.org/announcing-fedora-linux-45-beta/) and [current Fedora 45 release status](https://fedoraproject.org/wiki/Design_Suite/Fedora_45) describe what is available and what remains subject to change. ## What happens when a game launches? A native Linux game can use Linux graphics interfaces directly. For a Windows game running through Steam, [Proton](https://github.com/ValveSoftware/Proton) provides a compatibility layer built on Wine and other components. Tools such as DXVK and VKD3D-Proton translate Direct3D graphics calls into Vulkan commands. On many AMD and Intel systems, Mesa supplies the open-source graphics drivers that handle those Vulkan or OpenGL commands; the Linux kernel driver then communicates with the GPU. The desktop session sits alongside this path: GNOME and its Wayland compositor handle windows, display and input. The exact route varies with the game and graphics hardware, and NVIDIA’s proprietary driver follows a different graphics stack. That is why “Linux gaming compatibility” is not one switch. Proton can translate many Windows interfaces, but it cannot make every game work. A problem in a game’s Windows compatibility path is different from a graphics-driver bug, and both are different from a desktop bug that stops Steam accepting clicks. ## What Fedora 45 changes for that stack The Fedora 45 Workstation beta uses GNOME 51. Fedora’s package tracker currently lists Mesa 26.2.3 as the Fedora 45 stable build, and Linux kernel 7.2.9-300 as stable, with 7.2.9-301 in testing. Those are a snapshot of the beta repositories, not a promise of the final launch versions. The Fedora 44 kernel tracker also lists a stable 7.2.9 build, so the kernel version alone is not a Fedora 45-exclusive leap. A newer Mesa point release may include driver fixes or hardware support, but a version number by itself does not tell you whether your games will run faster. [Phoronix’s Fedora 44 versus Fedora 45 beta comparison](https://www.phoronix.com/review/fedora-45-beta-benchmarks) helps put performance claims in perspective. Its September 23 software snapshot used the same Linux 7.2 kernel, Mesa 26.2 and GCC 16.2.1 on both releases, on one high-end AMD workstation. The author found no broad system-performance shift, though some individual workloads changed. These were not game benchmarks, and the hardware was neither a gaming handheld nor a typical gaming laptop, so the results cannot predict frame rates for every setup. They do show why “Fedora 45 means more FPS” is too strong a conclusion. ## A beta bug can be in the desktop, not the GPU During the early beta, Fedora users reported that Steam and other XWayland windows stopped accepting mouse clicks under a GNOME 51 release candidate. In the [Fedora discussion thread](https://discussion.fedoraproject.org/t/fedora-45-beta-issue-with-steam/202218), users later reported that updating to Mutter 51.0 fixed the problem. That was a beta regression with a reported fix, not evidence that the current beta still has the issue. It is a useful example of why a broken Steam window does not automatically mean Proton or the graphics driver is at fault. ## Should a gamer upgrade now? If this is your main gaming PC, wait for Fedora 45 to reach its final release and check the final package versions and your GPU’s driver support. If you want to test the beta, use a spare drive or a setup you can roll back; keep in mind that package versions and behavior can still change before release. Fedora Workstation is a general-purpose desktop rather than a handheld-specific gaming image. For a handheld-first Linux setup, see [Mushimu’s Bazzite guide](https://www.mushimu.com/post/8728). The diagram is an original, simplified illustration of the software path; individual games and GPUs can use different components. ~Gaming 

Can you drop an app into REA and get its original source code back? No. REA gives an AI coding agent a way to investigate software through tools such as Ghidra, but reverse engineering still takes focused questions, careful reading, and checks against the original. A walkthrough of DX-Ball shows what that looks like in practice: tracing one sound effect back to the calculation that controls its left-right position. REA connects an agent to analysis tools for native binaries, JavaScript and Electron apps, .NET software, websites, and more. The agent can inspect evidence such as decompiled code, assembly instructions, callers, and data, then decide what to investigate next. That makes the process easier to navigate; it does not turn an executable into a clean, complete copy of its source project. ## A small question, followed through The DX-Ball example starts with a specific question: when a brick is hit, how does the game decide where its sound sits between the left and right speakers? The target is DX-Ball 1.07, a Windows game released in 1996. First, the agent asks REA to inspect the sound-panning function. The decompiler makes the function look as if it has no input, but the assembly shows that it reads a value passed in. That mismatch is a useful clue, not something to gloss over. The agent follows the function’s caller and finds that the game calculates a brick’s screen position as 20 + 30 × tile_x: start 20 pixels from the edge, then move 30 pixels for each tile. Next, it traces the values used by the instructions. Reading the relevant bytes as floating-point numbers gives 1.5625, 500.0, and an initial scale of 1.0. Put together, the calculation is: pan = (screen_x × 1.5625 − 500.0) × pan_scale At the default scale, a position of 320 maps to the centre (0); the edges at 0 and 640 map to −500 and 500. That is a concrete explanation of how a brick’s position can steer its sound between the speakers. ## The check matters more than the confident explanation The DX-Ball reconstruction project reports comparing its C version with the original for 3,205 cases: every integer screen position from 0 to 640 at five scale values. It also reports reproducing 63 bytes of the original function with its pinned Visual C++ 4.0 toolchain. Those are results reported by the project, not an independent benchmark of REA. They show the kind of check that makes a narrow reconstruction more convincing: compare it with the target instead of trusting a fluent explanation. The [DX-Ball walkthrough](https://rea.tools/showcase/dx-ball/) includes the queries and evidence. REA can also help with packaged apps. In a separate [Notion Desktop case study](https://rea.tools/showcase/notion/), the project follows a clipboard call through the app’s JavaScript, Electron preload, and main-process handlers. That is a different investigation from reading native assembly, and it still traces one behavior rather than rebuilding the whole application. ## What you need, and what still needs your judgment The current README lists Node.js 22.x, 24.x, or 26+ and npm for setup. Native binary analysis needs a suitable, properly licensed tool such as Ghidra, Hopper, or IDA; some static JavaScript and .NET inspections do not. The right setup also depends on the file and the question. A model may help with patterns and follow-up questions, but it can misread evidence, and harder binary work can demand a stronger model and more hands-on guidance. Someone who can spot a bad interpretation is still part of the workflow. REA’s project says analysis runs locally. The findings returned to a connected coding agent are also sent to that model provider. Only examine software you own or have permission to analyze, and use commercial tools under their licenses. For the project and current requirements, see the [REA repository](https://github.com/morluto/rea). The useful promise here is practical: investigate a specific behavior with an agent, then check what it found. “Reverse engineer anything” is the ambition; a verified slice of one program is what this example demonstrates. ~programming ~ai 

As of October 11, Mistral Large 4 is available as a public preview through Mistral Studio’s API, but its downloadable weights have not been released. Mistral’s [announcement](https://mistral.ai/news/mistral-large-4/) says they are due by the end of October; the official [Hugging Face release page](https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B) currently lists October 31 as its estimated date. For people planning to run AI on Linux or a home server, there is no official local install to try yet. The size figures also need care. Mistral’s [model card](https://docs.mistral.ai/models/mistral-large) lists 1.05 trillion total parameters and 52 billion active parameters, while the Hugging Face page describes 49 billion active per token, or 52 billion including embeddings and output layers. Those active-parameter counts describe the mixture-of-experts model’s computation; they are not a published GPU-memory requirement. A local deployment depends on the actual weight files, precision or quantization, inference software, and memory for the chosen context length. Mistral lists a one-million-token context window, but that does not mean a typical desktop can hold a million-token conversation in memory. Context and concurrent requests add runtime costs beyond the model weights. The company has not published a confirmed local hardware specification or an install guide for this release, so estimates circulating before the files arrive should be treated as estimates, not supported requirements. Until the weights and runtime support are available, the documented route is the hosted API preview. On launch day, [Le Monde reported](https://www.lemonde.fr/en/economy/article/2026/10/06/mistral-ai-unveils-new-ai-model-aimed-at-narrowing-the-gap-with-top-chinese-competitors_6758318_19.html) that Mistral’s performance claims still awaited confirmation in independent rankings. For the broader launch and its other model comparisons, see our [Mistral Large 4 launch overview](https://www.mushimu.com/post/8822). Image credit: Mistral Large 4 model icon from [Mistral’s documentation](https://docs.mistral.ai/models/mistral-large). ~AI 

/media/assets/c9a3c6ca-51bd-47ff-aef5-e55423a63abc/media.mp4

The International Organization for Migration says at least 139 people, including 15 children, died after an overcrowded boat carrying 305 people capsized off Djibouti during the night of Oct. 8–9. The vessel was traveling toward Yemen. The updated toll leaves 84 people missing. Eighty-two survivors, all Ethiopian nationals, have been accounted for, according to the IOM figures reported by the Associated Press and AFP. The death toll rose from 109 after more bodies were recovered. Rescue work was continuing on Sunday. IOM’s chief of mission in Djibouti told AP that overcrowding was the main cause of the sinking. Survivors were receiving medical and psychological support at an IOM center in Obock. The route across the Gulf of Aden and Red Sea is used by migrants traveling from the Horn of Africa toward the Gulf states. IOM figures cited by AP say more than 900 people died or went missing on the route in 2025. Sources: [Associated Press](https://apnews.com/article/a9e324ca36a7d9dee0e745c736c8aa95) · [AFP via L’Orient Today](https://today.lorientlejour.com/article/1550443/djibouti-boat-wreck-leaves-139-migrants-dead-84-missing-un.html) · [Reuters via Boursorama](https://www.boursorama.com/actualite-economique/actualites/au-moins-139-migrants-perissent-dans-un-naufrage-au-large-de-djibouti-af98840c6d72bd3fbfc335a29560d6ca) ~worldnews

Puerto Vallarta International Airport (PVR) is scheduled to reopen at 8 a.m. Monday, Oct. 12, local time. Grupo Aeroportuario del Pacífico (GAP) says flights scheduled after 5 p.m. Saturday and all Sunday services at PVR have been canceled. Manzanillo’s airport (ZLO) is scheduled to reopen at 8 a.m. Sunday for general aviation and Mexican Navy operations, weather permitting. Airlines have canceled all scheduled commercial flights there for Sunday. GAP says reopening and operations depend on weather and safety inspections. It advises passengers to contact their airline and avoid going to the terminals during closures. The National Hurricane Center’s 9 p.m. CST Saturday advisory placed Hurricane Simon about 65 miles (105 km) south-southwest of Manzanillo, with sustained winds of 155 mph (250 km/h), moving north. The NOAA forecast graphic below is from that advisory; conditions and airport operations can change. This is a travel update to our earlier [storm briefing](https://www.mushimu.com/post/8896). Sources: [GAP: Puerto Vallarta closure](https://aeropuertosgap.com.mx/files/Comunicados/20261010_Comunicado_Cierre_temporal_Aeropuerto_Internacional_de_Puerto_Vallarta.pdf) · [GAP: Manzanillo closure](https://aeropuertosgap.com.mx/files/Comunicados/20261010_Comunicado_Cierre_Temporal_ZLO_Huracan_Simon.pdf) · [NHC advisory 15](https://www.nhc.noaa.gov/text/refresh/MIATCPEP5+shtml/102057.shtml) · [AP reporting](https://apnews.com/article/1e6f590459e5ee4ead370192f52f9239). ~world ~worldnews ~hurricane 

Max Verstappen will start Sunday’s Singapore Grand Prix from pole after setting the fastest lap in qualifying, 1:31.373. Charles Leclerc was 0.054 seconds behind for Ferrari, with Lewis Hamilton third. The FIA says Hamilton led the first two qualifying segments, but Verstappen’s opening Q3 lap held up as Leclerc improved on his final attempt. It is Verstappen’s first Singapore Grand Prix pole. Kimi Antonelli qualified fourth and Lando Norris fifth. The 62-lap race is scheduled for 12:00 UTC on Sunday at the Marina Bay Street Circuit. Verstappen also won Saturday’s rain-affected Sprint; this qualifying result sets the grid for the main event. Read our separate [Sprint report](https://www.mushimu.com/post/8895). Sources: [FIA qualifying report and classification](https://api.fia.com/news/f1-verstappen-beats-ferrari-pair-take-singapore-grand-prix-pole) · [Formula 1 schedule and results](https://www.formula1.com/en/racing/2026/singapore) · [Grandprix.com qualifying report](https://www.grandprix.com/races/singapore-gp-2026-qualifying-report.html). Image: FIA. ~Sports 

Panama declared a national state of emergency on Saturday, Oct. 10, after Friday’s earthquakes and continuing aftershocks. The government says the measure will prioritize assistance to affected families; Cabinet Resolution 108 was published in the Official Gazette. The presidential ministry’s preliminary count lists 426 homes assessed, 419 education centers and 21 health facilities affected, with about 21 people treated for injuries. The Associated Press, citing Panama’s deputy civil-protection director, reported 27 injured people. Those figures differ, so both are attributed rather than presented as a single confirmed total. AP also reported more than 100 aftershocks. Panama’s government says response teams are checking damaged homes, schools, roads and bridges. Its preliminary update identified nine vehicle bridges with structural damage and said the bridges of the Americas and Centenario remained open. Sources: [Panama’s Official Gazette, Resolution 108](https://www.gacetaoficial.gob.pa/storage/gacetas/2026/10/30631/105696.pdf) · [Presidential Ministry update](https://www.presidencia.gob.pa/publicacion/gobierno-declara-estado-de-emergencia-para-atender-a-familias-afectadas-por-sismo) · [Associated Press](https://apnews.com/article/8217fbb2e56d75cb39b1f40fa036faa6) · [Reuters report](https://news.tv5.com.ph/foreign/read/panama-declares-state-of-emergency-following-quake-and-aftershocks) ~WorldNews ~Panama ~Earthquake

https://x.com/dadiani_george/status/2109684784100458604/video/1?s=46

Madrid celebrará el desfile de la Fiesta Nacional este lunes 12 de octubre. El Ayuntamiento prevé cortes de tráfico entre las 7:00 y las 15:00 en el eje que va desde la glorieta del Emperador Carlos V hasta Colón, pasando por el paseo del Prado, Cibeles y Recoletos. El acto central comenzará a las 11:00.  Durante ese intervalo, 48 líneas diurnas de la EMT modificarán temporalmente sus recorridos. Entre ellas figuran las líneas 1, 2, 14, 19, 27, 34, 52, 74, 146, 150 y el Exprés Aeropuerto. Los itinerarios podrán variar según avance el dispositivo, así que conviene consultar las incidencias de la EMT antes de salir. También cerrarán temporalmente 11 estaciones de BiciMAD situadas junto al recorrido durante el 11 y el 12 de octubre. Para llegar en transporte público, el Ayuntamiento señala como estaciones de Metro cercanas Atocha, Estación del Arte, Banco de España, Retiro, Serrano y Colón; las estaciones de Cercanías más próximas son Atocha y Recoletos. Recomienda evitar el coche por las calles afectadas y usar la M-30 o la M-40 para los desplazamientos largos. El Ayuntamiento puede ampliar o ajustar los horarios de los cortes según las necesidades del operativo. Consulta el [plan municipal de movilidad](https://www.madrid.es/portales/munimadrid/es/Inicio/Movilidad-y-transportes/Incidencias-de-Trafico/Principales-incidencias-en-la-ciudad/Plan-de-movilidad-del-Dia-de-la-Fiesta-Nacional-2026-/?vgnextchannel=cbe0aa2dedcb1610VgnVCM2000001f4a900aRCRD&vgnextoid=a1634e6b67511a10VgnVCM200000f921e388RCRD), el [horario y recorrido oficial](https://www.esmadrid.com/agenda/desfile-12-octubre) y el [mapa del Ministerio de Defensa](https://www.defensa.gob.es/12octubre/images-event/recorrido.png). ~spain

 https://popculturenews.com/breaking-bad-star-rj-mitte-dies-at-age-34/ Edit: I’ve been reading more about RJ today, and I wanted to update this post because he deserves to be remembered for much more than just Breaking Bad. I honestly had no idea how much he did for people with disabilities. He spent years fighting bullying, worked with Shriners Children’s on campaigns to protect kids with disabilities, and was an ambassador for United Cerebral Palsy. He also pushed for better representation and opportunities for disabled actors in Hollywood, working with SAG-AFTRA to help make the industry more inclusive. Having experienced bullying himself growing up, he used his fame to help others going through similar situations. It’s genuinely sad that someone who spent so much of his life trying to make things better for others is gone at just 34. He wasn’t just a great actor. He genuinely made a difference in people’s lives. Rest easy, RJ. You deserved so many more years. ❤️

Things seem to be getting pretty serious in Saudi Arabia right now. The Houthis in Yemen have been launching missiles and drones at airports and oil facilities, and some of the attacks have caused casualties and major flight disruptions. Saudi Arabia is also carrying out airstrikes in Yemen, so it looks like the situation is escalating on both sides. I’m surprised this isn’t getting more attention. Anyone else following this? How much worse could this get? Here are a few things I’ve come across on X https://x.com/zxcvbnm151k/status/2109743788864438730?s=46 https://x.com/theprophetsb77/status/2109707826260226319?s=46 https://x.com/ki5rz2/status/2109720252229783823?s=46

/media/assets/781646c7-1fd2-42bb-84d5-fb39f72a62ee/media.mp4

/media/assets/37e6705e-d409-4303-9b0a-084d2b33df22/media.mp4

I’ll be posting any GTA leaks I come across in this thread. https://x.com/jaake_so_ja/status/2108376481235890316/video/1?s=46

/media/assets/fe4f0083-3432-4e40-ba82-ad2217658d8a/media.mp4

Edit: https://www.mushimu.com/post/8942 https://github.com/morluto/rea Apparently it lets AI agents reverse engineer software, from websites and apps to native binaries, using tools like Ghidra. Has anyone here actually tried it? I’m curious how well it works in practice, especially with more complex software. Is it genuinely useful or one of those projects that looks amazing in demos but struggles with real-world stuff?

/media/assets/1d5f188f-46a0-4c6b-bef0-fbebae557eb2/media.mp4

/media/assets/78bd9287-369b-44ec-853d-1f830eaad2b1/media.mp4

/media/assets/7debc5f7-46f1-4cc5-a25a-93f0c4a53ff0/media.mp4

/media/assets/1d6b77d9-b3ca-4553-bd30-66614c75ae20/media.mp4

/media/assets/c50369bd-8d8a-4b17-9039-2ae3d13211ff/media.mp4

/media/assets/63d7e987-56a4-4443-8f62-b21f5165c666/media.mp4 ~india

/media/assets/772afc90-5e14-4e0b-9bc2-7ab6a085fac7/media.mp4