A nerdy father of two, a husband of a beautiful and understanding wife, an engineer who loves anime and tinkering with PCs and games, but most of all, loves God.
Today’s world runs on digital computers, but there was a time when people relied on their analog siblings. Instead of electrical signals, mechanical computers utilize complex systems of gears, weights, and other physical implements to perform tasks. As recently as the 1960s, accounting devices like the Soviet Union’s Ascota 170 could even perform square roots, but some of the earliest mechanical computers may even trace all the way back to the famous Antikythera mechanism. Discovered in an ancient shipwreck near Crete in 1901, experts now believe ancient Greeks built the device around 100 BCE to calculate astronomical positions.
Mechanical computers may not oversee today’s automated tasks, but it’s still possible to build your own using literal children’s toys. That said, one YouTuber’s ongoing DIY mechanical computer project looks so dizzyingly complex that—for the time being—most people will likely stick with their smartphone’s calculator app.
The inventor who goes by Shadowman39 on YouTube is a longtime Knex aficionado. His previous projects built from the connectible, plastic building pieces include everything from coin sorters, to gigantic motion machines, to even a functioning Skee Ball cabinet. More recently, he opted to design and construct a device capable of performing basic calculations like addition and subtraction. Although he’s planning on expanding its abilities, the YouTuber recently uploaded the first look at his Knex 8-bit mechanical computer—specifically its “very simple” arithmetic logic unit, or ALU.
“Very simple,” in this case, refers to a roughly 3.5-foot-tall contraption built using thousands of Knex pieces, wheels, and conveyor belts. All those parts are divided up into eight columns, each one responsible for a single bit of binary data represented as a 0 or 1. The 8-bit system is capable of handling operations using the numbers 0-255, although Shadowman39 notes (using some rather dense mathematical reasoning) that it can also work with the number range of -128 to 127.
While there is an electrical line to power initial movement, the actual calculations themselves take place through mechanical means. Depending on the user’s input, the Knex computer will use its rack and pinions to initiate cranks capable of rotating 360 degrees while anchored by a pair of weights and counterweights. These actions compound across the eight columns until a calculation is completed.
“It looks like a mess but I promise it’s just organized chaos,” the inventor says at one point.
You might just need to take his word for it on that front. In any case, the YouTuber promises additional videos are coming that will delve into how data and program storage will work in its RAM and ROM. It may not find its way onto desktops, but the Knex project highlights the uniquely tactile and ingenious designs required to construct even the most rudimentary of mechanical computers.
Geeks love to get a little more power out of their hardware by digging deep in its guts, either literally or with software. A little gentle overclocking is usually safe, and in fact it’s expected to a certain degree by manufacturers. But an unorthodox method for laptops with discrete Nvidia RTX graphics cards is extremely powerful…and notably dangerous.
It’s called a “shunt mod,” and it’s essentially a way to get a graphics card to run at power input way beyond what it’s rated to run at. Shunt modding is a complex, physical process that needs active soldering on the GPU itself, but if done correctly, it can boost the power going into the card by a huge margin. For some of the most powerful RTX laptop cards, this can boost the power input from a spec maximum of 150-175 watts to up to 250 watts, with a similar delta in terms of performance.
How and why is this possible? It’s complicated. Some of the laptops with RTX 4090, 5080, and 5080 discrete GPUs are running the same physical graphics chips as seen on the desktop cards, which are rated for much higher wattage inputs. Because of the “binning” process for when these chips are manufactured, the most perfectly produced chips go into the most powerful desktop cards, while the ones that didn’t quite win the “silicon lottery” might be used for laptops. In short, the chips should — should — be able to handle much more watt input than they typically get from a more restrictive, battery-saving laptop configuration.
GizmoSlipTech, an expert in the minutia of desktop and laptop GPUs, has aggregated some data from both professional and enthusiast users who’ve attempted this complex process to boost the energy going into RTX 5080 and 5090-equipped laptops (via VideoCardz.com). The results are impressive. One test of an RTX 5090 laptop showed a whopping 41 percent increase in the 3DMark Steel Nomad benchmark when the card was modified to run at 250 watts. In-game results are less dramatic, but more than 20 percent improvement in Cyberpunk 2077 is nothing to sneeze at.
But before you reach for a screwdriver, you might want to take a second to think it through. We’re talking about some pretty deep hardware modifications here, the kind of thing that shouldn’t be attempted by anyone without some pretty impressive, base-level electronics projects under their belt. This goes way beyond the basic desktop assembly or even laptop repair skills. Cracking open your laptop to access the discrete graphics card might void the warranty on its own, but the full process requires taking a soldering iron to capacitors and then modifying the firmware to make it run hotter.
In short, doing this might melt your gaming laptop. And if it does, you can kiss any kind of warranty support goodbye. In fact even if something completely unrelated to your GPU should fail, I imagine that any technician who spots this modification will instantly declare the laptop out of warranty due to user modification.
And these are some of the most powerful, and most expensive, gaming laptops available right now. Performing a shunt mod is essentially gambling with thousands of dollars of hardware for a performance boost. A nice boost, but in my opinion, one that’s not worth the risk even if you have the technical skills to achieve it.
Still, the fact that this modification is possible — and apparently gaining popularity among the bravest users — indicates that these discrete laptop cards might have a lot more power available than they’re currently utilizing. In fact, some laptop manufacturers have quietly performed similar mods to their gaming models without telling anyone (and then had to issue a patch to undo it). GizmoSlipTech recommends that Nvidia increase the power limits to some laptops, utilizing the latest thermal management tech, to deliver more performance.
It’s possible to eject a graphics card while you’re using your computer, if it’s an external model or some other non-standard setup. But if you’re using a desktop with a GPU installed on the motherboard, or a laptop with discrete graphics within the case, it’s generally a bad idea. Which is why some Windows 11 users were alarmed when their computers asked if they wanted to eject their graphics cards.
This appears to be a Windows bug that adds graphics cards to a list of devices that can be ejected in the “Safely remove hardware and media” tool in the notification center. It’s that thing you’re supposed to use every time you unplug a flash drive, which I’ve never seen anyone use in the real world, ever. Some Reddit users, seeing the option to “eject” the very powerful, very expensive graphics card soldered into the guts of their laptop, were understandably confused.
This bug has been showing up for years, apparently as an unintended UI expression of the option to disable some discrete laptop graphics and switch to integrated graphics to save power. (Note that “disabling” and “ejecting” are very different terms, and probably shouldn’t be conflated for regular PC users.) Windows Central notes that there are registry hacks available to turn off this alarming and apparently unintentional message.
While hardly the worst user-facing bug for Windows, I can see how this would be concerning even if you’re fairly tech-savvy. “I’m afraid I might accidentally eject it one day and god knows what’s gonna happen to my laptop,” says one Reddit user on the Windows 11 sub. If you’re seeing something similar, and you don’t want to dive into a registry hack (which I think is very admirable caution), maybe just leave it alone.
A major zero-day security vulnerability in Microsoft’s widely used SharePoint server software has been exploited by hackers, causing chaos within businesses and government agencies, multipleoutlets have reported. Microsoft announced that it had released a new security patch "to mitigate active attacks targeting on-premises [and not online] servers," but the breach has already effected universities, energy companies, federal and state agencies and telecommunications firms.
The SharePoint flaw is a serious one, allowing hackers to access file systems and internal configurations or even execute code, to completely take over systems. The flaw could put more than 10,000 companies at risk, Cybersecurity company Censys told The Washington Post. "It’s a dream for ransomeware operators, and a lot of attackers are going to be working this weekend as well." Google’s Threat Intelligence Group added that the flaw allows "persistent, unauthenticated access that can bypass future patching."
The US Cybersecurity and Infrastucture Security agency (CISA) said that any servers affected by the exploit should be disconnected from the internet until a full patch arrives. It added that the impact of the attacks is still being probed.
The vulnerability was first spotted by Eye Security, which said the flaw allows hackers to access SharePoint servers and steal keys in order to impersonate users or services. "Because SharePoint often connects to core services like Outlook, Teams, and OneDrive, a breach can quickly lead to data theft, password harvesting, and lateral movement across the network," Eye Security wrote in a blog post.
The FBI is aware of the attack and is working closely with government and private sector partners. It’s not immediately clear which groups are behind the zero-day hacks. In any case, the attack is liable to put Microsoft under the microscope again. A 2023 breach of Exchange Online mailboxes led the White House’s Cyber Safety Review Board to declare that Microsoft’s security culture was "inadequate."
This article originally appeared on Engadget at https://ift.tt/kjXWmEL
Imagine an electric track car that blasts to 60 mph in 1.38 seconds, generates 2,000 kg of downforce at zero speed, and carries a $1.13 million price tag. That’s the McMurtry Speirling. Every stab at the throttle plants you firmly in your seat like Maverick’s Tomcat. Few machines on earth deliver this kind of visceral thrill — and the Speirling isn’t just for pro drivers.
Although, budget-friendly? Only if you consider a small South Pacific island budget-friendly.
Performance and Drivability Insights
The McMurtry Speirling detonates off the line, eclipsing top-tier EV hypercars. It rockets from 0–60 mph in 1.38 seconds, thanks to 1,000 hp and a 1,000 kg curb weight — an unrivaled power-to-weight ratio . By comparison, the Tesla Model S Plaid takes 2.0 seconds and weighs 4,766 lb.
Steering feels razor-sharp. The rack-and-pinion setup relays every surface detail without twitchiness. Suspension grips aggressively through pitch and roll, then soaks up track bumps with race-car poise. Fan-powered downforce pushes cornering g-loads past 3Gs, yet transitions stay smooth and predictable.
The Speirling’s cabin serves a single driver. A carbon-fiber monocoque and closed cockpit offer motorsport-grade safety. You get an adjustable steering column and pedals — but no infotainment screen, just critical lap data.
Expect a 60 kWh pack built around Taiwanese cell maker, Molicel. It uses Molicel’s P50B cylindrical cells with, one of the first silicon-carbon anode EV batteries on the market that has every chance of being the next big thing. This Molicel pack recharges in 20 minutes at 600 kW and delivers roughly 25 minutes of full-tilt lapping.
On public roads, aggressive regen and the lightweight design yield about 50 MPGe. That 50 MPGe beats the fuel economy of mainstream hybrids like the 2025 Toyota Prius Eco at 56 mpg combined, or the 2025 Honda Insight at 52 mpg combined.
Unlike these small hybrids, though, noise does climb past 120 dB when fans spin up, so ear protection earns its keep. Storage and comfort take a back seat to performance, and the $1.1 million sticker guarantees exclusivity.
Silicon-Anode Battery Tech
Using silicon anodes boosts energy density up to 40% over graphite and cuts charge times in half. There is even some industry talk of 90-second 0-100% EV charging. Molicel deploys US-made Group14’s SCC55® material under license, pairing Taiwan’s cell-assembly expertise with advanced silicon chemistry.
Verdict: Daily Grind Meets Enthusiast Thrill
The McMurtry Speirling feels like sprinting alongside supercars — but leaving them in the dust. It won’t haul groceries or connect to Bluetooth, but it delivers fan-driven grip and lightning reflexes. You trade creature comforts and cargo space for pure, unfiltered performance.
This car is incredible. Its speed is out of this world. But the battery tech is where we need to be watching. Consider this almost hypersonic EV as the runway model for future EV batteries. Getting this silicon battery tech out to a larger market solves energy density and therefore range and charging anxiety, and would spark a new age for EVs.
For the enthusiast who lives for tactile feedback, track precision, the world flying past at breakneck speed, and the world’s first silicon-carbon battery EV, the Speirling stands alone.
It used to be that the most electrified thing about your car was the static shock you got stepping out in the winter. Now? It’s a different game entirely. Mild hybrids are everywhere in 2025. From German sport sedans to budget crossovers, manufacturers are slapping 48-volt systems onto everything with four wheels and calling it progress. And while that belt-driven boost might give you sharper throttle response or a smug feeling about fuel economy, there’s one thing nobody really talks about: what happens when the battery kicks the bucket?
Lynk &? Co?
Unlike plug-in hybrids (PHEVs) or fully electric vehicles (EVs), mild hybrids don’t rely on their battery packs to drive the wheels. They’re support acts. Smaller batteries, smaller motors, simpler setups. Which, in theory, should mean cheaper repairs. And they do — sort of. As of mid-2025, the average cost of a replacement MHEV battery in the U.S. sits around $1,500, according to research conducted by Jalopnik.
That figure comes from a cross-section of OEM parts prices, and it’s about what you’d pay for a decent laptop, or roughly three months of insurance on a Range Rover you bought during a moment of weakness. But the range is broad. The battery for an Audi RS Q8 will cost you over $2,000, while the one in a BMW M340i clocks in just above $1,200. Mazda’s CX-70 hybrid? $1,980. The Mercedes C-Class? Try $2,116. Even the humble Volvo XC60 sneaks in at $1,159.
2025 Audi RS Q8
Audi
Of course, if you’re still under warranty, it’s not your problem. Most automakers offer 8-year/100,000-mile coverage on hybrid components. But what happens when you hit year nine? Or when a newer battery spec replaces yours and the old one gets quietly discontinued? This is where the mild hybrid’s simplicity comes back to bite.
The new Range Rover Sport HST introduces the brand’s first Ingenium 3.0-liter straight-six, and production mild hybrid system.
Land Rover
Today’s Tech, Tomorrow’s Orphan
The story of the original Honda Insight is a cautionary tale here. Sold from 1999 to 2006, the first-gen Insight was what we’d now call a mild hybrid. Fast-forward to today, and Honda no longer makes replacement batteries for it. You can still get a third-party pack for about $1,749, but that’s assuming someone’s still bothering to stock it. The hidden costs of EV ownership aren’t just about charging infrastructure and tax breaks — they’re also about long-term parts availability, something mild hybrids have largely escaped scrutiny for. Until now.
EVs might get the headlines, but as of 2025, hybrids still win on total cost of ownership over ten years. That includes fuel savings, maintenance, and yes, battery replacements. PHEV battery failures remain rare, too. So for now, mild hybrids offer a sweet spot. Reasonable savings. Manageable tech. And no need to panic about charging ports.
Still, things are changing fast. China’s pushing sodium-ion battery tech that could bring EV battery costs down to $10/kWh. If that happens, the cost gap between mild hybrids and proper EVs might shrink to nothing.
So if you’re buying a mild hybrid today, go in eyes open. That extra boost might feel clever now, but it could be a pain in the wallet come 2035. At which point you’ll either be driving something new, or trying to explain to your mechanic what an RS Q8 even was.
Soil excavated from the moon could be used to produce oxygen and methane, which could be used by lunar settlers for breathing and for rocket fuel.
This is the conclusion of a team of scientists from China who have found a one-step method of doing all this. Whether it is economically viable, however, is up for debate.
But the Chinese team thinks that it is. "The biggest surprise for us was the tangible success of this integrated approach," said team-member Lu Wang, who is a chemist from the Chinese University of Hong Kong, in a statement. "The one-step integration of lunar water extraction and photothermal carbon dioxide catalysis could enhance energy utilization efficiency and decrease the cost and complexity of infrastructure development."
They point out that studies have shown that transporting supplies from Earth to any future moon base would be expensive because the greater the mass of cargo, the harder a rocket has to work to launch into space. Studies have indicated that it would cost $83,000 to transport just one gallon of water from Earth to the moon, and yet each astronaut would be expected to drink 4 gallons of water per day.
Fortunately, the moon has plentiful water, although it is not automatically apparent. Brought to the moon by impacts of comets, asteroids and micrometeoroids, and even by the solar wind, water lurks in permanently shadowed craters at the lunar poles, trapped within minerals such as ilmenite.
Extracting the water for drinking is relatively easy and there are numerous technologies that describe how this can be done, including heating the regolith by focusing sunlight onto it. However, the Chinese team has been able to take this one step further.
"What’s novel here is the use of lunar soil as a catalyst to crack carbon dioxide molecules and combine them with extracted water to produce methane," Philip Metzger, a planetary physicist from the University of Central Florida, told Space.com. Metzger was not involved in the new research, but he is the co-founder of the NASA Kennedy Space Center’s ‘Swamp Works‘, a research lab for designing technologies for construction, manufacturing and mining on planetary (and lunar) surfaces.
Get the Space.com Newsletter
Breaking space news, the latest updates on rocket launches, skywatching events and more!
Methane would be more desirable than liquid hydrogen as a potential rocket fuel because it is easier to keep stable, thereby requiring less machinery and less cost to keep on the moon. Liquid methane, when mixed with oxygen as an oxidizer, is a potent rocket fuel. Many commercial companies such as China’s Landspace are already launching methane-powered rockets.
Chang’e-5 lunar soil sitting at the bottom of a photothermal reactor. (Image credit: Sun et al.)
The water-bearing ilmenite is also a useful catalyst for reacting the water with carbon dioxide to produce oxygen and methane, and the Chinese team have developed a one-step process for doing so. First, they heat the regolith to 392 degrees Fahrenheit (200 degrees Celsius) by focusing sunlight to release the water inside. Then, carbon dioxide such as that which could be breathed out by astronauts is added to the mix, causing the ilmenite to catalyze the reaction between the extracted water and the carbon dioxide. Researchers tested this process, known as photothermal catalysis, in the laboratory using a simulant based on samples of lunar regolith returned to Earth by China’s Chang’e 5 mission (the lunar samples are far too previous to destroy in such experiments, which is why a simulant is used instead).
While previous technologies have also been able to accomplish this, they required more steps and more machinery, and used a catalyst that would have to be transported up from Earth. This, the research team believe, makes their process more efficient and less expensive than the alternatives.
However, Metzger is not wholly confident that it will work. For one thing, lunar regolith is a proficient thermal insulator, so heating a sample all the way through would not be easy.
"The heat does not spread effectively deeper into the soil, and this greatly reduces the amount of water that can be produced in a given time," Metzger said. One option could be to ‘tumble’ the regolith, turning it over repeatedly so that the heat is more evenly applied, but this slows the extraction of water and increases the mechanical complexity of the process. In an environment where lunar dust gets into every nook and cranny, and where temperature fluctuations between night and day can be as great as 482 degrees Fahrenheit (250 Celsius), the risk of breakdown only increases as more moving parts enter the equation.
"It may be doable, but more maturation of the technology is needed to show that it is actually competitive," said Metzger.
Lunar soil samples collected by Chang’e 5 lunar probe is on display during a science exhibition marking the 10th Space Day of China at Shanghai World Expo Exhibition and Convention Center on April 27, 2025 in Shanghai, China. (Image credit: VCG/VCG via Getty Images)
There’s also a problem with the application of carbon dioxide, something recognized by both the Chinese team and Metzger. Specifically, there’s a question mark over whether astronauts could produce enough carbon dioxide through their normal exhalation. Metzger calculates that astronauts could only provide a tenth of the carbon dioxide required. Alternatively, carbon dioxide could be shuttled up from Earth, but this would rather defeat the purpose of the proposed technique, which was to develop a lot-cost means of obtaining water, oxygen and methane with resources largely already available on the moon.
However, in the long-run, perhaps shipping some materials up from Earth will be beneficial. Metzger points out a similar experiment that used an exotic granular catalyst – nickel-on-kieselguhr (kieselguhr is a kind of sedimentary rock) – rather than lunar regolith. Metzger suspects that a material specifically designed to be a catalyst, such as nickel-on-kieselguhr, would be more efficient than lunar regolith. Plus, although it would be expensive to transport from Earth, the nickel-on-kieselguhr can be re-used so you would only need to transport it to the moon once. In a cost-benefit analysis, in the long term it might be more efficient to do this instead.
Regardless, the research team has convincingly shown that using lunar regolith as a catalyst to produce fuel and water works. The next step is to show that the technology can be scaled up to sustain a base on the moon more efficiently than other techniques, and that it can operate in lunar conditions where the gravity is weaker, the temperature swings to large extremes, and there is intense radiation from space.
"I think these are highly interesting results and there may be additional applications to use lunar soil as a photocatalyst," said Metzger. "More work will be needed to show whether this concept can be economically competitive. I am skeptical, but all good ideas have their detractors and you can never really know until somebody does the work to prove it."
There is certainly no immediate rush for the technology. With NASA’s Artemis III mission, which aims to finally return astronauts to the surface of the moon in 2027 at the earliest, and funding made available for Artemis IV and V at some indeterminate time in the future, we’re not yet in a position to build a permanent lunar base.
However, the Artemis missions are the perfect opportunity to trial some of these technologies and will be greatly important for showing whether we really can live on the moon or not.
The research was published on July 16 in the journal Joule.
