News
- Startup’s ribbon-based holographic tape promises 200TB per LTO cartridge
- The tech uses polymer film and $5 laser to write optical voxels
- Integrates into LTO systems with no upstream software or hardware changes
UK startup HoloMem is developing a holographic storage system aimed at replacing or supplementing LTO tape.
The company, founded by former Dyson engineer Charlie Gale, uses polymer ribbon cartridges written with $5 laser diodes. Each 100-meter cartridge could store up to 200TB in a write-once, read-many format.
The cartridges match LTO dimensions and work in existing tape libraries without changes to upstream software. Drives function as drop-in shelves, allowing libraries to operate in a hybrid LTO and HoloMem setup.
HO1OThe idea began at Dyson, where Gale helped create a holographic label system called HO1O. It embedded multiple QR codes in a single hologram, readable from different angles or light sources.
“What we originally did at HO1O for prototypes was to use a light-sensitive polymer material that you just exposed to laser light… it locks polymer change and retains that image,” Gale told Blocks & Files.
This concept evolved into multi-layer data storage using similar materials.
Unlike other optical approaches that use glass or ceramics, HoloMem writes data as holographic voxels into polymer film. The film uses a 16-micron thick polymer sheet laminated between PET layers, forming a 120-micron ribbon.
The prototype HoloDrive writes and reads holograms using a 3D-printed lens and a digital micromirror device.
“We are writing data pages of thousands of bits,” Gale said. Throughput hasn’t been disclosed, although it reportedly operates at LTO-9 speeds. The drive uses £30 circuit boards and modified LTO mechanics.
HoloMem has received £900,000 in UK innovation grants and is partnering with TechRe and QStar for field trials and integration testing. It holds patents for the optical engine, media design and volumetric storage method.
Blocks & Files reports: “We understand TechRe will deploy prototype Holodrives inside LTO libraries in its UK data centers to test out the product’s performance, reliability and robustness. HoloMem has written device firmware so that, we understand, it presents itself as a kind of LTO drive.”
Future capacity increases may come through multi-channel recording, using multiple light wavelengths to layer data. Each added channel could multiply storage with no hardware change.
You might also like- Only 2% of enterprises are highly ready for AI, report claims
- Fewer than one-third have deployed AI firewalls to date
- Another one in three could do with diversifying their AI models
Although more and more applications are getting AI overhauls, new F5 research had claimed only 2% of enterprises are highly ready for AI.
More than one in five (21%) fall into the low-readiness category, and while three-quarters (77%) are considered moderately ready, they continue to face security and governance hurdles.
This comes as one in four applications use AI, with many organizations splitting their AI usage across multiple models including paid models like GPT-4 and open-source models like Llama, Mistral and Gemma.
Enterprises aren't benefitting from the AI they have access toAlthough 71% of the State of AI Application Strategy Report respondents said they use AI to enhance security, F5 highlighted ongoing challenges with security and governance. Fewer than one in three (31%) have deployed AI firewalls, and only 24% perform continuous data labelling, potentially increasing risks.
Looking ahead, one in two (47%) say they plan on deploying AI firewalls in the next year. F5 also recommends that enterprises diversify AI models across paid and open-source opens, scale AI usage to operations, analytics and security, and deploy AI-specific protections like firewalls and data governance strategies.
At the moment, it's estimated that two-thirds (65%) use two or more paid models and at least one open-source model, demonstrating considerable room for improvement.
"As AI becomes core to business strategy, readiness requires more than experimentation—it demands security, scalability, and alignment," F5 CPO and CMO John Maddison explained.
The report highlights how enterprises that lack of maturity can stifle growth, introduce operational bottlenecks and present compliance challenges.
"AI is already transforming security operations, but without mature governance and purpose-built protections, enterprises risk amplifying threats," Maddison added.
You might also like- Why not consider the best AI tools and best AI writers
- Make your workers more efficient with the best productivity tools
- Many SMBs say they can't get to grips with AI, need more training
- Hackers only need cheap hardware and basic skills to stop a moving freight train remotely
- The American Association of Railways dismissed the threat until federal pressure forced a response
- The system still isn’t fixed, and full updates won’t arrive until at least 2027
A critical flaw in the wireless systems used across US rail networks has remained unresolved for more than a decade, exposing trains to remote interference.
The vulnerability affects End-of-Train (EoT) devices, which relay data from the last carriage to the front of the train, forming a link with the Head-of-Train (HoT) module.
Although the issue was flagged in 2012, it was largely dismissed until federal intervention forced a response.
Ignored warnings and delayed responsesHardware security researcher Neils first identified the flaw in 2012, when software-defined radios (SDRs) began to proliferate.
The discovery revealed that these radios could easily mimic signals sent between the HoT and EoT units.
Since the system relies on a basic BCH checksum and lacks encryption, any device transmitting on the same frequency could inject false packets.
In a concerning twist, the HoT is capable of sending brake commands to the EoT, which means an attacker could stop a train remotely.
“This vulnerability is still not patched,” Neils stated on social media, revealing it took over a decade and a public advisory from the Cybersecurity and Infrastructure Security Agency (CISA) before meaningful action was taken.
The issue, now catalogued as CVE-2025-1727, allows for the disruption of U.S. trains with hardware costing under $500.
Neils's findings were met with skepticism by the American Association of Railways (AAR), which dismissed the vulnerability as merely “theoretical” back in 2012.
Attempts to demonstrate the flaw were thwarted due to the Federal Railway Authority's lack of a dedicated test track and the AAR denying access to operational sites.
Even after the Boston Review published the findings, the AAR publicly refuted them via a piece in Fortune.
By 2024, the AAR’s Director of Information Security continued to downplay the threat, arguing that the devices in question were approaching end-of-life and didn’t warrant urgent replacement.
It wasn’t until CISA issued a formal advisory that the AAR began outlining a fix. In April 2025, an update was announced, but full deployment is not expected until 2027.
The vulnerability stems from technology developed in the 1980s, when frequency restrictions reduced the risk of interference, but today’s widespread access to SDRs has altered the risk landscape dramatically.
“Turns out you can just hack any train in the USA and take control over the brakes,” Neils said, encapsulating the broader concern.
The ongoing delay and denial mean US trains are probably sitting on a keg of gunpowder that could lead to serious risks at any time.
Via TomsHardware
You might also like- These are the fastest SSDs you can buy right now
- Take a look at some of the best external hard drives available
- This popular WordPress plugin has been hijacked to push potential malware
- AWS unveils Kiro, an agentic AI way to code
- Kiro looks to help solve typically issues seen in "vibe coding"
- Kiro is in preview now, with three tiers set to be available
Amazon Web Services (AWS) has unveiled Kiro, an IDE which uses AI agents to streamline the development process.
Available now in preview, Kiro looks to cut down on potential issues with "vibe coding", the process where agents are being asked to create and build software with minimal human interaction.
As well as helping with coding, Kiro can also automatically create and update project plans and technical blueprints, solving one of the most troublesome issues for developers who are still getting to grips with the potential AI brings.
AWS KiroAnnouncing the launch, AWS said Kiro is looking to help transition from “vibe coding to viable code.”
It works by breaking down prompts into structured components, which can then be used to guide implementation and testing, as well as tracking any changes as the code evolves, ensuring no inconsistencies break through.
There's also Model Context Protocol (MCP) support for connecting specialized tools, steering rules to guide AI behavior across your project, and agentic chat for ad-hoc coding tasks.
Finally, it can also automatically check through code to make sure nothing is amiss, making sure developers can submit or launch code without fear of any problems.
Kiro looks, “to solve the fundamental challenges that make building software products so difficult — from ensuring design alignment across teams and resolving conflicting requirements, to eliminating tech debt, bringing rigor to code reviews, and preserving institutional knowledge when senior engineers leave," Nikhil Swaminathan, Kiro’s product lead, and Deepak Singh, Amazon’s vice president of developer experience and agents, said.
"Kiro is great at ‘vibe coding’ but goes way beyond that—Kiro’s strength is getting those prototypes into production systems with features such as specs and hooks."
For now, Kiro is free to use during the preview period, but it seems AWS is looking at introducing three pricing tiers: a free version with 50 agent interactions per month; a Pro tier at $19 per user per month with 1,000 interactions; and a Pro+ tier at $39 per user per month with 3,000 interactions.
"Kiro is really good at "vibe coding" but goes well beyond that," Amazon CEO Andy Jassy wrote in a post on X.
"While other AI coding assistants might help you prototype quickly, Kiro helps you take those prototypes all the way to production by following a mature, structured development process out of the box. This means developers can spend less time on boilerplate code and more time where it matters most – innovating and building solutions that customers will love.
You might also like- Understanding the vibe coding trend and considerations for developers
- We've rounded up the best laptop for programming options out there
- And these are the best business laptops around
- An employee used a very bad AI-generated image to advertise graphic designer jobs at Xbox
- The image shows a woman writing code that somehow appears on the back of a computer monitor, among other problems
- The ad is especially awkward as Microsoft recently completed laying off more than 9,000 people
A post on LinkedIn seeking graphic designers for Xbox is going viral for the irony of terrible AI-generated graphics. Principal Development Lead for Xbox Graphics, Mike Matsel, shared a post announcing the roles, accompanied by what at first glance appears to be an innocuous cartoon of a woman at a workstation typing code. Except the code is on the back of her monitor, and that's just the beginning of the issues with the image.
The fact that Microsoft concluded the latest of several rounds of layoffs, affecting a total of more than 9,000 people, including many in the Xbox division, just a few weeks ago, makes it even more awkward.
(Image credit: LinkedIn/Mike Matsel)The more you examine the image, the more obvious it becomes that it was (poorly) produced with AI. The computer is unconnected to anything, the desk sort of fades away into nothingness, and the shadows don't make sense. Plus, would Microsoft want a graphic of someone clearly using Apple headphones? Not to mention the fact that, in 2025, you're very unlikely to see someone with the corded iPhone headphones of nearly 20 years ago.
The image does at least sell the idea that Microsoft desperately needs graphic designers, or at least people who know when graphics are very wrong. The dozens of comments on the post emphasize just how annoying many people find the post. A lot are from developers and graphic designers who might otherwise be interested in the positions.
Awkward AIThe fact that this wasn’t just a bad image, but one that undermines the entire point of the job being advertised, is truly mind-boggling. It’s like handing out flyers for a bakery that uses clip art of a melting candle with "bread" written on the attached label.
It's so bizarrely bad that more than a few commenters wondered if it was on purpose. It might be a way to draw attention to the open positions, or, unlikely as this may be, a form of malicious compliance from someone instructed to use AI to announce the open jobs after their colleagues in those positions were recently let go. Or maybe it was the sharpest satire ever seen on LinkedIn.
Those are wildly unlikely theories, but it's telling that they aren't totally impossible. An ad symbolizing everything people are worried about, especially regarding the very artistic jobs being advertised, would be far too blatant to use in a joke. Still, apparently, that's just reality now.
The fact that Microsoft is currently investing billions of dollars in AI only adds to the dissonant reaction. Even if it wasn't formally approved by Microsoft, it still has their Xbox logo on it. Then again, even senior executives can faceplant when discussing and using AI.
Just last week, Executive Producer at Xbox Game Studios Publishing Matt Turnbull suggested that people recently let go could turn to AI chatbots to help get over their emotional distress and find new jobs. He took down the essay encouraging former employees to use AI tools to both find jobs and for "emotional clarity," eventually, but this graphic disaster remains visible to the public, as opposed to the code hiding behind the back of the monitor.
You might also like