If the instructions were actually specific, e.g. Put a blackberry in its right eye socket, then yes, most humans would know what that meant. But the instructions were not that specific: in the right eye socket
168,483 karma · joined June 14, 2011
Formerly: teaching at Stanford and programming+newsing at ProPublica.
Blog: http://www.danwin.com
Twitter: https://twitter.com/dancow
[ my public key: https://keybase.io/dnguyen; my proof: https://keybase.io/dnguyen/sigs/69iS-DQno44YqjgaowUD_Zze660k3fojXZmHob-bgxA ]
https://www.npr.org/transcripts/nx-s1-5601059
A great listen and thought provoking all the way through, but the part most specifically relevant to HN:
> GONZALEZ: Preston Thorpe is the one making six figures, which, by the way, is double what the corrections officers who guard him make. And it's been a game changer for Preston. He said it's hard enough to get a job when you have a criminal record, let alone while you're still inside.
THORPE: And now, I feel like my life has a purpose. Like, there's no situation right now that would cause me to do something where I would risk losing, like, my job, my computer.
GONZALEZ: Preston is 33 years old. And he told Susan Sharon that he's always been a computer guy, a computer nerd, he said, since he was 13 years old. It's kind of what got him in trouble later in life.
SHARON: He talked to me about buying drugs on the dark web and selling them. And I think the second time, he was convicted because he had a powerful synthetic opioid, much more deadly than fentanyl, capable of killing lots of people.
GONZALEZ: Preston is about nine years into his 20-ish-year sentence. He used to be in a different prison, in a different state, and he says he got in a lot of trouble there, so much so that they transferred him to Maine. Like, we need this person out of our custody. And when he got to Maine, Preston started seeing possibilities-- school, picking up coding again. And he did super well, so well, no issues. And eventually, he got a remote job as a lead principal engineer for this nonprofit that pushes for education in prisons. And because Preston had a laptop, you know, in his cell all day and all night that he could use for certain approved things, Preston started contributing to this big open-source coding project.
Basically, this company was going to attempt to rewrite this database called SQLite in Preston's favorite programming language.
One of the agents named in the subpoena appears to have previously worked on child exploitation cases years ago:
https://www.supremecourt.gov/DocketPDF/22/22-6039/245948/202...
The most prominent and deep-pocketed promoters of this tech — e.g. Musk and Altman — are constantly making this analogy.
This story is about Christine Hunsicker, former CEO/founder of CaaStle, a “clothing as a service” company that purported to be a hybrid of Shopify and Rent the Runway
I first glimpsed it while doing a touristy bike tour my first day in Barcelona and it’s hard to convey how surreal it is, like a video game’s final level glitching into its tutorial stage.
Relates to this main story about the contents of Epstein’s yahoo email account (which is less technical)
https://www.bloomberg.com/features/2025-jeffrey-epstein-emai...
Don’t often see mainstream articles refer to Python libraries by name:
> Whenever emails included cryptographic “DomainKeys Identified Mail” signatures, or DKIM, in their headers, we used Python’s dkimpy library to verify that signatures matched their messages. A valid signature match is strong evidence that the message is genuine and unaltered, but a failed DKIM check is not, by itself, evidence of fabrication. Our analysis of emails from the years when DKIM was widely seen in the inbox (2013-2022) validated roughly 45% of the messages, rising from about 20% in 2013 to nearly two-thirds by 2020 as DKIM adoption broadened. For these messages we also checked DKIM-based DMARC identifier alignment—whether the DKIM d= domain aligned with the visible “From:” domain—and 55% of messages were aligned. We manually inspected the emails that weren’t aligned and found all matched to known mailing list providers and other similar services.
You're not uneducated, but this is a common and fundamental misunderstanding of how racial inequity can afflict computational systems, and the source of the problem is not (usually) something as explicit as "the creators are Nazis".
For example, early face-detection/recognition cameras and software in Western countries often had a hard time detecting the eyes on East Asian faces [0], denying East Asians and other people with "non-normal" eyes streamlined experiences for whatever automated approval system they were beholden to. It's self-evident that accurately detecting a higher variety of eye shapes would require more training complexity and cost. If you were a Western operator, would it be racist for you to accept the tradeoff for cheaper face detection capability if it meant inconveniencing a minority of your overall userbase?
Well, thanks to global market realities, we didn't have to debate that for very long, as any hardware/software maker putting out products inherently hostile to 25% of the world's population (who make up the racial majority in the fastest growing economies) weren't going to last long in the 21st century. But you can easily imagine an alternate timeline in which Western media isn't dominant, and China & Japan dominate the face-detection camera/tech industry. Would it be racist if their products had high rates of false negatives for anyone who had too fair of skin or hair color? Of course it would be.
Being auto-rejected as "not normal" isn't as "racist" as being lynched, obviously. But as such AI-powered systems and algorithms have increasing control in the bureaucracies and workflows of our day to day lives, I don't think you can say that "racist output", in the form of certain races enjoying superior treatment versus others, is a trivial concern.
[0] https://www.cnn.com/2016/12/07/asia/new-zealand-passport-rob...
This story is about former Texas Gov. Rick Perry becoming a vocal advocate for “using the powerful psychotropic drug ibogaine as a potential treatment for brain trauma, addiction and even cognitive decline”
> Within ~3 minutes of the crash, the Model S packaged sensor video, CAN‑bus, EDR, and other streams into a single “snapshot_collision_airbag-deployment.tar” file and pushed it to Tesla’s server, then deleted its local copy.
Putting aside the legal implications wrt evidence, etc — what is the ostensible justification for this functionality? To a layperson, it's as bizarre as designing a plane's black box to go ahead and delete data if it somehow makes a successful upload to the FAA cloud server. Why add complexity that reduces redundancy in this case?
That was helpful context, thank you for highlighting it. From the NYT article, it sounds like the NYT’s pursuit of a feature piece last year [0] (also worth reading) spurred Science to revisit making a retraction.
> The internet and scientific critics largely moved on, and so did the journal. While some researchers called for the paper’s retraction, Science instead published technical critiques of the finding. Then last year, Science’s stance shifted. A reporter contacted Science for a New York Times article about the legacy of the #arseniclife affair.
> That inquiry “convinced us that this saga wasn’t over, that unless we wanted to keep talking about it forever, we probably ought to do some things to try to wind it down,” said Holden Thorp, editor in chief of Science since 2019. “And so that’s when I started talking to the authors about retracting.”
[0] https://www.nytimes.com/2025/02/11/science/arseniclife-felis...
Last 30 days: 6%
2025 so far: 5.7%
2024: 4.5%
edit: analytics.usa.gov includes iOS and Android in its operating systems breakdown — e.g. Windows has a 32% share vs OP's 63%. Assuming most of Linux users are on desktop, it could be the case that Linux share in desktop users is a bit higher than 6%
Regardless of the program’s actual risk, it doesn’t seem that the government is fully aware of the program’s very existence. The article quotes the former CIO of the Pentagon as being surprised:
> John Sherman, who was chief information officer for the Department of Defense during the Biden administration, said he was surprised and concerned to learn of ProPublica’s findings. “I probably should have known about this,” he said. He told the news organization that the situation warrants a “thorough review by DISA, Cyber Command and other stakeholders that are involved in this.”
> The OPD didn’t share information directly with the federal agencies. Rather, other California police departments searched Oakland’s system on behalf of federal counterparts more than 200 times — providing reasons such as “FBI investigation” for the searches — which appears to mirror a strategy first reported by 404 Media, in which federal agencies that don’t have contracts with Flock turn to local police for backdoor access.