Article note: I'm really excited for the due-process implications.
It's been a "secret list so you can't have standing because you can't prove you are on the list" legal catch-22 since the early-2000s mass psychosis took over the US, now we can finally start picking the dumb apart.
Article note: ...That is some wild shit, and it's from 2004.
I do and teach digital design, and I've now skimmed through the dissertation twice, and I'm only half sure I understand the generalities of how it works.
Article note: I always find the tension between "Reproducible builds and immutable systems are good for security" and "Heterogeneity is good for security" interesting.
Both make whole classes of attack _dramatically_ harder, but the methods are often mutually exclusive.
Article note: Performance prediction/analysis on modern systems with dynamic behavior is wild shit. It's surprising how often "Don't do anything unexpected" is the best heuristic because of caching and branch prediction.
Article note: Ohh, yeah. The "cheap" SCSI emulators are continuing to get more performant and aren't getting more expensive, this is good news, and timely.
The giant Micropolis HDD in my HP9000/735 is looking dead, my Sparcstation10 hates all CD drives, and my SCSI2SD is mounted inside my PowerMac 6100/66 DOS at the moment, so I've been investigating options.
BlueSCSI is an open source, open hardware, and open design SCSI solution for vintage computers.
The original version 1.x devices use a “Blue Pill” microcontroller board based on ST chips. Due to the chip shortages, clone ST chips have often been used.
Today the team announced BlueSCSI v2!
BlueSCSI v2 is based on the Raspberry Pi Pico microcontroller and a fork of ZuluSCSI’s SCSI2SD code
It is open source, open hardware, and open design.
BlueSCSI v2 targets the Raspberry Pi Pico (not the Pi) which uses the RP2040 microcontroller. Like before we’re building on the shoulders of those who came before us – namely this code base is based on the ZuluSCSI’s SCSI2SD. This is a joint effort between Eric(nulleric) the maintainer of BlueSCSI and Jacob(Androda) a core BlueSCSI developer and maintainer of the F4 BlueSCSI fork.
We’ve added our “special sauce” to the hardware and software. Hardware for BlueSCSI v2 will be released under the same Creative Commons Non-Commercial license as BlueSCSI v1 – we believe in open hardware and this will allow you to build the device yourself if you like.
The original BlueSCSI isn’t dead! We will continue to support and port features to the F1/F4 versions of BlueSCSI as we can. Not all features can be ported back, and speeds are more limited on the earlier models.
See the product page and GitHub for all the details. Efforts like this are what open source is about.
Article note: I was joking about this inevitability with various people in the days before it appeared.
Of _course_ it makes it easier to subvert the "writing formular papers about nothing" style of assignment.
Academics have been doing chatterbot fake papers for _decades_ with things like SCIgen, this just lowered the bar to entry.
Students have been writing papers so halfassed they're incoherent assemblages of phrases, or mechanically paraphrased and/or straight-up plagiarized from various sources, or similar behavior forever.
And of _course_ turnitin is building a detector to sell to scared institutions.
This is all business as usual.
Turnitin, best known for its anti-plagiarism software used by tens of thousands of universities and schools around the world, is building a tool to detect text generated by AI. The Register reports: Turnitin has been quietly building the software for years ever since the release of GPT-3, Annie Chechitelli, chief product officer, told The Register. The rush to give educators the capability to identify text written by humans and computers has become more intense with the launch of its more powerful successor, ChatGPT. As AI continues to progress, universities and schools need to be able to protect academic integrity now more than ever. "Speed matters. We're hearing from teachers just give us something," Chechitelli said. Turnitin hopes to launch its software in the first half of this year. "It's going to be pretty basic detection at first, and then we'll throw out subsequent quick releases that will create a workflow that's more actionable for teachers." The plan is to make the prototype free for its existing customers as the company collects data and user feedback. "At the beginning, we really just want to help the industry and help educators get their legs under them and feel more confident. And to get as much usage as we can early on; that's important to make a successful tool. Later on, we'll determine how we're going to productize it," she said.
Turnitin's VP of AI, Eric Wang, said there are obvious patterns in AI writing that computers can detect. "Even though it feels human-like to us, [machines write using] a fundamentally different mechanism. It's picking the most probable word in the most probable location, and that's a very different way of constructing language [compared] to you and I," he told The Register. [...] ChatGPT, however, doesn't have this kind of flexibility and can only generate new words based on previous sentences, he explained. Turnitin's detector works by predicting what words AI is more likely to generate in a given text snippet. "It's very bland statistically. Humans don't tend to consistently use a high probability word in high probability places, but GPT-3 does so our detector really cues in on that," he said.
Wang said Turnitin's detector is based on the same architecture as GPT-3 and described it as a miniature version of the model. "We are in many ways I would [say] fighting fire with fire. There's a detector component attached to it instead of a generate component. So what it's doing is it's reading language in the exact same way GPT-3 reads language, but instead of spitting out more language, it gives us a prediction of whether we think this passage looks like [it's from] GPT-3." The company is still deciding how best to present its detector's results to teachers using the tool. "It's a difficult challenge. How do you tell an instructor in a small amount of space what they want to see?" Chechitelli said. They might want to see a percentage that shows how much of an essay seems to be AI-written, or they might want confidence levels showing whether the detector's prediction confidence is low, medium, or high to assess accuracy. "I think there is a major shift in the way we create content and the way we work," Wang added. "Certainly that extends to the way we learn. We need to be thinking long term about how we teach. How do we learn in a world where this technology exists? I think there is no putting the genie back in the bottle. Any tool that gives visibility to the use of these technologies is going to be valuable because those are the foundational building blocks of trust and transparency."
Article note: Hm. They had gathered a bunch of the people with interesting OS ideas from the last 30 years (Be influence via Travis Geiselbrecht and Brian Swetland, bunch of clear Plan9 and 90s research microkernel influences), but like many google projects it was in that weird "not pure research, not obviously leading to product" space.
Google is still reeling from the biggest layoff in company history last Friday. Earlier cost cuts over the past six months have resulted in several projects being shut down or deprioritized at Google, and it's hard to fire 12,000 people without some additional projects taking a hit. The New York Times has a report about which divisions are being hit the hardest, and a big one is Google's future OS development group, Fuchsia.
While the overall company cut 6 percent of its employees, the Times pointed out that Fuchsia saw an outsize 16 percent of the 400-person staff take a hit. While it's not clear what that means for the future of the division, the future of Fuchsia's division has never really been clear.
Fuchsia has been a continuous mystery inside Google since it first saw widespread press coverage in 2017. Google rarely officially talks about it, leaving mostly rumors and Github documentation for figuring out what's going on. The OS isn't a small project, though—it's not even based on Linux, opting instead to use a custom, in-house kernel, so Google really is building an entire OS from scratch. Google actually ships the OS today to consumers in its Nest smart displays, where it replaced the older Cast OS. The in-place operating system swap was completely invisible to consumers compared to the old OS, came with zero benefits, and was never officially announced or promoted. There's not much you can do with it on a locked-down smart display, so even after shipping, Fuchsia is still a mystery.
Article note: Heh. Pre-VM UNIX ported to microcontrollers that dramatically out-muscle the machines it was developed on. That's fun. I don't see exactly how they're handling the 16/32b matter (the parent was PDP-11 16 bit UNIX, these are 32bit micros).
I don't think I have any suitable F4 boards lying around or I'd be wasting the rest of the morning fiddling with it.
Article note: Speaking of Pascal. That's pretty neat. I'd be more hyped about a good bare-metal Pascal of that style targeting something a little smaller (stm32? esp32? riscv?), but it's still neat. Looks like it requires a hacked-up FreePascal/Lazarus system, and isn't self-hosting (like that little crosstalk Smalltalk-80 environment).