So-called artificial intelligence (AI) is all the rage right now between your grandma asking ChatGPT how to code in Python or influencers making videos without having to hire extras, but one growing concern is where the power is going to come from for the data centers. The MIT Technology Review team did a deep dive on what the current situation is and whether AI is going to kill us all (with carbon emissions).

Probably of most interest to you, dear hacker, is how they came up with their numbers. With no agreed upon methods and different companies doing different types of processing there were a number of assumptions baked into their estimates. Given the lack of information for closed-source models, Open Source models were used as the benchmark for energy usage and extrapolated for the industry as a whole. Unsurprisingly, larger models have a larger energy usage footprint.

While data center power usage remained roughly the same from 2005 to 2017 as increases in efficiency offset the increase in online services, data centers doubled their energy consumption by 2023 from those earlier numbers. The power running into those data centers is 48% more carbon intensive than the US average already, and expected to rise as new data centers push for increased fossil fuel usage, like Meta in Louisiana or the X data center found to be using methane generators in violation of the Clean Air Act.

Technology Review did find “researchers estimate that if data centers cut their electricity use by roughly half for just a few hours during the year, it will allow utilities to handle some additional 76 gigawatts of new demand.” This would mean either reallocating requests to servers in other geographic regions or just slowing down responses for the 80-90 hours a year when the grid is at its highest loads.

If you’re interested in just where a lot of the US-based data centers are, check out this map from NREL. Still not sure how these LLMs even work? Here’s an explainer for you.

As cars increasingly become computers on wheels, the attack surface for digital malfeasance increases. [PCAutomotive] has shared its exploit for turning the 2020 Nissan Leaf into 1600 kg RC car. [PDF via Electrek]

Starting with some scavenged infotainment systems and wiring harnesses, the group built test benches able to tear into vulnerabilities in the system. An exploit was found in the infotainment system’s Bluetooth implementation, and they used this to gain access to the rest of the system. By jamming the 2.4 GHz spectrum, the attacker can nudge the driver to open the Bluetooth connection menu on the vehicle to see why their phone isn’t connecting. If this menu is open, pairing can be completed without further user interaction.

Once the attacker gains access, they can control many vehicle functions, such as steering, braking, windshield wipers, and mirrors. It also allows remote monitoring of the vehicle through GPS and recording audio in the cabin. The vulnerabilities were all disclosed to Nissan before public release, so be sure to keep your infotainment system up-to-date!

If this feels familiar, we featured a similar hack on Tesla infotainment systems. If you’d like to hack your Leaf for the better, we’ve also covered how to fix some of the vehicle’s charging flaws, but we can’t help you with the loss of app support for early models.

Spending time as wee hackers perusing the family atlas taught us an appreciation for a good map, and [Billy Roberts], a cartographer at NREL, has served up a doozy with a map of the data center infrastructure in the United States. [via LinkedIn]

Fiber optic lines, electrical transmission capacity, and the data centers themselves are all here. Each data center is a dot with its size indicating how power hungry it is and its approximate location relative to nearby metropolitan areas. Color coding of these dots also helps us understand if the data center is already in operation (yellow), under construction (orange), or proposed (white).

Also of interest to renewable energy nerds would be the presence of some high voltage DC transmission lines on the map which may be the future of electrical transmission. As the exact location of fiber optic lines and other data making up the map are either proprietary, sensitive, or both, the map is only available as a static image.

If you’re itching to learn more about maps, how about exploring why they don’t quite match reality, how to bring OpenStreetMap data into Minecraft, or see how the live map in a 1960s airliner worked.

As there is no cure for celiac disease, people must stick to a gluten free diet to remain symptom-free. While this has become easier in recent years, scientists have found some promising results in mice for disabling the disease. [via ScienceAlert]

Since celiac is an auto-immune disorder, finding ways to alter the immune response to gluten is one area of investigation to alleviate the symptoms of the disease. Using a so-called “inverse vaccine,” researchers “engineered regulatory T cells (eT_regs) modified to orthotopically express T cell receptors specific to gluten peptides could quiet gluten-reactive effector T cells.”

The reason these are called “inverse vaccines” is that, unlike a traditional vaccine that turns up the immune response to a given stimuli, this does the opposite. When the scientists tried the technique with transgenic mice, the mice exhibited resistance to the typical effects of the target gluten antigen and a related type on the digestive system. As with much research, there is still a lot of work to do, including testing resistance to other types of gluten and whether there are still long-term deleterious effects on true celiac digestive systems as the transgenic mice only had HLA-DQ2.5 reactivity.

If this sounds vaguely familiar, we covered “inverse vaccines” in more detail previously.

Sometimes it’s nice to have a widget to do a single task and avoid getting distracted by the supposed simplicity of doing it with an app on a smartphone. [Dina Amin] built a timer from an old flip clock to stay focused.

Starting with a disassembly of the flip clocks she found at a flea market with [Simone Giertz], [Amin] decided to change the twenty four hour mechanism to a twenty four minute one which was similar to the amount of time she was already using for several different practices. Since she’s an expert in animation, she planned on turning a set of CT scans into the animation that would play on the section that had previously been the minutes of the clock.

As much of the original clock’s components were damaged, and [Amin] didn’t have a chance to learn clockmaking from scratch in a week, she tried a few different drive mechanisms for the build. The drum from an air fryer timer driven with an electric motor fit the bill, but off enough from proper minutes that [Amin] switched from numerals to a yellow circle that fills in as it approaches the satisfying ding of completion.

If you want to see Simone’s Moon flip clock we’ve covered that project too.

We love clocks, but we especially love unusual timepieces that aren’t just about showing the hour of the day. [Simone Giertz] built a flip clock moon phase tracker for a friend.

While in Egypt for Cairo Maker Faire, [Giertz] and [dina Amin] found some old flip clocks at a flea market and had to have them. [Amin] mentioned wanting to make a moon phase tracker with one, and [Giertz] decided to try her hand at making her own version. A side quest in more comfortable flying is included with the price of admission, but the real focus is the process of figuring out how to replicate the flip clocks original mechanism in a different size and shape.

[Giertz] cut out 30 semi-circle flaps from polystyrene and then affixed vinyl cut-outs to the flaps. The instructions for the assembly suggest that this might not be the best way to do it, and that printing stickers to affix to the flaps might work better since the cut vinyl turned out pretty fiddly. We really like the part where she built a grid jig to determine the optimal placement of the beams to keep the flaps in the right position after a disheartening amount of difficulties doing it in a more manual way. Her approach of letting it rest for twenty minutes before coming back to it is something you might find helpful in your own projects.

Best of all, if you want to build your own, the files are available for the flip moon station on the Yetch website. You’ll have to come up with your own method to drive it though as that isn’t in the files from what we saw.

The Mac mini is the closest to an Apple-based SBC you can get, so it lends itself to unusual portable computers. [Scott Yu-Jan] is back to tackle a portable build using the latest and greatest M4 mini.

[Yu-Jan] walks us through his thought process of how to maximize the portability of the system without all that tedious mucking about with setting up a separate keyboard, monitor, and the mini while on the go. With the more complicated electronics, the monitor risked tipping the keyboard over when attached, particularly since [Yu-Jan] isn’t a fan of batteries for his portables.

By affixing the Mac mini to the side of the keyboard, it makes the whole thing easier to slip into a bag without being overly thick. We get a peek into his iterative process as well when he evaluates the build and decides that the closing of the lid wasn’t what he was hoping for. By adding some TPU rests for the monitor to rest on in the closed position, he says it’s really brought the whole project up a notch. We certainly have had our own projects where one little detail really moves it from sketchy to polished, and we appreciate when makers clue us in on where that happened for them.

You may recognize [Yu-Jan] from our previous coverage of his older portable all-in-one Mac mini and this luggable version where he explains why he doesn’t like laptops. If you like your computers more stationary, how about some G4 iMacs with the newer internals from an M-series mini?

If you’re like us and you’ve been wondering where those viral videos of single water drop chemical reactions are coming from, we may have an answer. [yu3375349136], a scientist from Guangdong, has been producing some high quality microchemistry videos that are worth a watch.

While some polyglots out there won’t be phased, we appreciate the captioning for Western audiences using the elemental symbols we all know and love in addition to the Simplified Chinese. Reactions featured are typically colorful, but simple with a limited number of reagents. Being able to watch diffusion of the chemicals through the water drop and the results in the center when more than one chemical is used are mesmerizing.

We do wish there was a bit more substance to the presentation, and we’re aware not all readers will be thrilled to point their devices to Douyin (known outside of China as TikTok) to view them, but we have to admit some of the reactions are beautiful.

If you’re interested in other science-meets-art projects, how about thermal camera landscapes of Iceland, and given the comments on some of these videos, how do you tell if it’s AI or real anyway?

While tools like CRISPR have blown the field of genome hacking wide open, being able to predict what will happen when you tinker with the code underlying the living things on our planet is still tricky. Researchers at Stanford hope their new Evo 2 DNA generative AI tool can help.

Trained on a dataset of over 100,000 organisms from bacteria to humans, the system can quickly determine what mutations contribute to certain diseases and what mutations are mostly harmless. An “area we are hopeful about is using Evo 2 for designing new genetic sequences with specific functions of interest.”

To that end, the system can also generate gene sequences from a starting prompt like any other LLM as well as cross-reference the results to see if the sequence already occurs in nature to aid in predicting what the sequence might do in real life. These synthetic sequences can then be made using CRISPR or similar techniques in the lab for testing. While the prospect of building our own Moya is exciting, we do wonder what possible negative consequences could come from this technology, despite the hand-wavy mention of not training the model on viruses to “to prevent Evo 2 from being used to create new or more dangerous diseases.”

We’ve got you covered if you need to get your own biohacking space setup for DNA gels or if you want to find out more about powering living computers using electricity. If you’re more curious about other interesting uses for machine learning, how about a dolphin translator or discovering better battery materials?