Archive.org now has a page with “the raw analog waveform and the reconstructed digital tape image (analog.tap), read at the Computer History Museum’s Shustek Research Archives on 19 December 2025 by Al Kossow using a modified tape reader and analyzed with Len Shustek’s readtape tool.” A Berlin-based retrocomputing enthusiast has created a page with the contents of the tape ready for bootstrapping, “including a tar file of the filesystem,” and instructions on dumping an RK05 disk image from tape to disk (and what to do next).

Research professor Rob Ricci at the University of Utah’s school of computing posted pictures and video of the tape-reading process, along with several updates. (“So far some of our folks think they have found Hunt The Wumpus and the C code for a Snobol interpreter.”)
University researcher Mike Hibler noted the code predates the famous comment “You are not expected to understand this” — and found part of the C compiler with a copyright of 1972.

The version of Unix recovered seems to have some (but not all) of the commands that later appeared in Unix v5, according to discussion on social media. “UNIX wasn’t versioned as we know it today,” explains University of Utah PhD student Thalia Archibald, who researched early Unix history (including the tape) and also worked on its upload. “In the early days, when you wanted to cut a tape, you’d ask Ken if it was a good day — whether the system was relatively bug-free — and copy off the research machine… I’ve been saying It’s probably V5 minus a tiny bit, which turned out to be quite true.”

Read more of this story at Slashdot.

One developer tells MIT Technology Review that AI tools weaken the coding instincts he used to have. And beyond that, “It’s just not fun sitting there with my work being done for me.”

But is AI making coders faster? “After speaking to more than 30 developers, technology executives, analysts, and researchers, MIT Technology Review found that the picture is not as straightforward as it might seem…”

For some developers on the front lines, initial enthusiasm is waning as they bump up against the technology’s limitations. And as a growing body of research suggests that the claimed productivity gains may be illusory, some are questioning whether the emperor is wearing any clothes…. Data from the developer analytics firm GitClear shows that most engineers are producing roughly 10% more durable code — code that isn’t deleted or rewritten within weeks — since 2022, likely thanks to AI. But that gain has come with sharp declines in several measures of code quality. Stack Overflow’s survey also found trust and positive sentiment toward AI tools falling significantly for the first time. And most provocatively, a July study by the nonprofit research organization Model Evaluation & Threat Research (METR) showed that while experienced developers believed AI made them 20% faster, objective tests showed they were actually 19% slower…

Developers interviewed by MIT Technology Review generally agree on where AI tools excel: producing “boilerplate code” (reusable chunks of code repeated in multiple places with little modification), writing tests, fixing bugs, and explaining unfamiliar code to new developers. Several noted that AI helps overcome the “blank page problem” by offering an imperfect first stab to get a developer’s creative juices flowing. It can also let nontechnical colleagues quickly prototype software features, easing the load on already overworked engineers. These tasks can be tedious, and developers are typically glad to hand them off. But they represent only a small part of an experienced engineer’s workload. For the more complex problems where engineers really earn their bread, many developers told MIT Technology Review, the tools face significant hurdles…

The models also just get things wrong. Like all LLMs, coding models are prone to “hallucinating” — it’s an issue built into how they work. But because the code they output looks so polished, errors can be difficult to detect, says James Liu, director of software engineering at the advertising technology company Mediaocean. Put all these flaws together, and using these tools can feel a lot like pulling a lever on a one-armed bandit. “Some projects you get a 20x improvement in terms of speed or efficiency,” says Liu. “On other things, it just falls flat on its face, and you spend all this time trying to coax it into granting you the wish that you wanted and it’s just not going to…” There are also more specific security concerns, she says. Researchers have discovered a worrying class of hallucinations where models reference nonexistent software packages in their code. Attackers can exploit this by creating packages with those names that harbor vulnerabilities, which the model or developer may then unwittingly incorporate into software.

Other key points from the article:

LLMs can only hold limited amounts of information in context windows, so “they struggle to parse large code bases and are prone to forgetting what they’re doing on longer tasks.”

“While an LLM-generated response to a problem may work in isolation, software is made up of hundreds of interconnected modules. If these aren’t built with consideration for other parts of the software, it can quickly lead to a tangled, inconsistent code base that’s hard for humans to parse and, more important, to maintain.”

“Accumulating technical debt is inevitable in most projects, but AI tools make it much easier for time-pressured engineers to cut corners, says GitClear’s Harding. And GitClear’s data suggests this is happening at scale…”
“As models improve, the code they produce is becoming increasingly verbose and complex, says Tariq Shaukat, CEO of Sonar, which makes tools for checking code quality. This is driving down the number of obvious bugs and security vulnerabilities, he says, but at the cost of increasing the number of ‘code smells’ — harder-to-pinpoint flaws that lead to maintenance problems and technical debt.”

Yet the article cites a recent Stanford University study that found employment among software developers aged 22 to 25 dropped nearly 20% between 2022 and 2025, “coinciding with the rise of AI-powered coding tools.”

The story is part of MIT Technology Review’s new Hype Correction series of articles about AI.

Read more of this story at Slashdot.

New York governor Kathy Hochul signed legislation on Friday aimed at holding large AI developers accountable for the safety of their models. The RAISE Act establishes rules for greater transparency, requiring these companies to publish information about their safety protocols and report any incidents within 72 hours of their occurrence. It comes a few months after California adopted similar legislation. 

But, the penalties aren’t going to be nearly as steep as they were initially presented when the bill passed back in June. While that version included fines of up to $10 million dollars for a company’s first violation and up to $30 million for subsequent violations, according to Politico, Hochul’s version sets the fines at up to $1 million for the first violation, and $3 million for any violations after that. In addition to the new reporting rules, a new oversight office dedicated to AI safety and transparency is being born out of the RAISE Act. This office will be part of the Department of Financial Services, and issue annual reports on its assessment of large AI developers. 

Hochul signed two other pieces of AI legislation earlier in December that focused on the use of the technology in the entertainment industry. At the same time, President Trump has been pushing to curb states’ attempts at AI regulation, and signed an executive order this month calling for “a minimally burdensome national standard” instead.

This article originally appeared on Engadget at https://www.engadget.com/ai/governor-hochul-signs-new-yorks-ai-safety-act-220503930.html?src=rss

“Yet another distro is making the move to the KDE Plasma desktop,” writes Linux magazine.

“Parrot OS, a security-focused Linux distribution, is migrating from MATE to KDE Plasma, starting with version 7.0, now available in beta.”

Based on Debian 13, Parrot OS’s goal is a shift toward “modernization, focusing on clearing technical debt and future-proofing the system.” One big under-the-hood change is that the/tmpdirectory is now automatically mounted astmpfs(in RAM), as opposed to the physical drive. By making this change, Parrot OS enjoys improved performance and reduces wear on SSDs. This shift also means that all data in/tmpis lost during a reboot.

ParrotOS senior systems engineer Dario Camonita explains the change in a blog post, calling it “not only aesthetic, but also in terms of usability and greater consistency with our future goals…”

“While MATE will continue to be supported by us as long as upstream development continues, We have noticed and observed the continuous improvements made by the KDE team…”

And elsewhere Linux Magazine notes two other distros are embracing the desktop Enlightenment:

For years, Bodhi Linux was one of the very few distributions that used anything based on Enlightenment. That period of loneliness is officially over, withMX Mokshaand AV Linux 25. MX Moksha doesn’t replace the original MX Linux. Instead, it will serve as an “official spin” of the distribution…

The Enlightenment desktop (and subsequently Moksha) was developed with systemd in mind, so MX Moksha uses systemd. If you’re not a fan of systemd, MX Moksha is not for you. MX Moksha is lighter than MX Linux, so it will perform better on older machines. It also uses the Liquorix kernel for lower latency.

AV Linux has been released with the Xfce and LXDE desktops at different times and has only recently opted to make the switch to Enlightenment.

Read more of this story at Slashdot.

During a search for the Brown shoogin suspect, a law enforcement press conference included a request for “Ring camera footage from residents and businesses near Brown University,” according to local news reports.

But in the end it was Flock cameras according to an article in Gizmodo, after a Reddit poster described seeing “odd” behavior of someone who turned out to be the suspect:

The original Reddit poster, identified only as John in the affidavit, contacted police the next day and came in for an interview. He told them about his odd encounter with the suspect, noting that he was acting suspiciously by not having appropriate cold-weather clothes on when he saw him in a bathroom at Brown University. That was two hours before the shooting. After spotting him in the bathroom wearing a mask, John actually started following the suspect in what he called a “game of cat and mouse….” Police detectives showed John two images obtained through Flock, the company that’s built extensive surveillance infrastructure across the U.S. used by investigators, and he recognized the suspect’s vehicle, replying, “Holy shit. That might be it,” according to the affidavit. Police were able to track down the license plate of the rental car, which gave them a name, and within 24 hours, they had found Claudio Manuel Neves Valente dead in a storage facility in Salem, New Hampshire, where he reportedly rented a unit.
“We intend to continue using technology to make sure our law enforcement are empowered to do their jobs,” Flock’s safety CEO Garrett Langley wrote on X.com, pinning the post to the top of his feed.

Though ironically, hours before Providence Police Chief Oscar Perez credited Flock for helping to find the suspect, CNN was interviewing Flock’s safety CEO to discuss “his response to recent privacy concerns surrounding Flock’s technology.”

To Langley, the situation underscored the value and importance of Flock’s technology, despite mounting privacy concerns that have prompted some jurisdictions to cancel contracts with the company… Langley told me on Thursday that he was motivated to start Flock to keep Americans safer. His goal is to deter crime by convincing would-be criminals they’ll be caught… One of Flock’s cameras had recently spotted [the suspect’s] car, helping police pinpoint Valente’s location. Flock turned on additional AI capabilities that were not part of Providence Police’s contract with the company to assist in the hunt, a company spokesperson told CNN, including a feature that can identify the same vehicle based on its description even if its license plates have been changed.

The company has faced criticism from some privacy advocates and community groups who worry that its networks of cameras are collecting too much personal information from private citizens and could be misused. Both the Electronic Frontier Foundation and the American Civil Liberties Union have urged communities not to work with Flock.
“State legislatures and local governments around the nation need to enact strong, meaningful protections of our privacy and way of life against this kind of AI surveillance machinery,” ACLU Senior Policy Analyst Jay Stanley wrote in an August blog post. Flock also drew scrutiny in October when it announced a partnership with Amazon’s Ring doorbell camera system… [“Local officers using Flock Safety’s technology can now post a request directly in the Ring Neighbors app asking for help,” explains Flock’s blog post.]

Langley told me it was up to police to reassure communities that the cameras would be used responsibly… “If you don’t trust law enforcement to do their job, that’s actually what you’re concerned about, and I’m not going to help people get over that.” Langley added that Flock has built some guardrails into its technology, including audit trails that show when data was accessed. He pointed to a case in Georgia where that audit found a police chief using data from LPR cameras to stalk and harass people. The chief resigned and was arrested and charged in November…

More recently, the company rolled out a “drone as first responder” service — where law enforcement officers can dispatch a drone equipped with a camera, whose footage is similarly searchable via AI, to evaluate the scene of an emergency call before human officers arrive. Flock’s drone systems completed 10,000 flights in the third quarter of 2025 alone, according to the company… I asked what he’d tell communities already worried about surveillance from LPRs who might be wary of camera-equipped drones also flying overhead. He said cities can set their own limitations on drone usage, such as only using drones to respond to 911 calls or positioning the drones’ cameras on the horizon while flying until they reach the scene. He added that the drones fly at an elevation of 400 feet.

Read more of this story at Slashdot.