Tuskie

@briankrebs@infosec.exchange @parismarx@mastodon.online As someone that has read the magazine for decades, it’s a good magazine for a lot of things but their perennial weakness is treating the far right — social, economic or political — with a sympathetic lens & underplaying their sincerity and destructiveness. It’s something they have no qualms airing about when it comes to anyone remotely mildly left of center (see their recent pieces on “Gen Z socialism”, and the horrors of figures like Madami, and the associated framing).

@briankrebs@infosec.exchange @parismarx@mastodon.online That Zanny immediately backtracks and apologizes for journalists writ large, despite Musk just throwing FUD out there, and not making any specific allegation (because he’s an idiot, doesn’t know anything, and just plays the typical far right playbook of attacking the messenger/media), speaks volumes. Meanwhile, Musk is directly responsible for killing nearly a million souls per most recent figures (see Kristoff, etc). It’s exhausting.

@jeffjarvis@mastodon.social Every time you read an op-ed from these places, always do a cursory background check on who’s writing these single ply editorials: “Lyman Stone is a demographer and the director of the Pronatalism Initiative at the Institute for Family Studies.” (He’s also a former missionary) Sane people don’t expect an infinite boom/line go up scenario for most things in life. Eventually, what goes up, must come down.

@carnage4life@mas.to Saw that earlier and just laughed. Seeing Côte d'Ivoire jammed into the far southeastern side of the continent and *landlocked* was just brilliant. That, on top of the fact that the countries are just made up shapes. Like Uganda looks like the size of Rwanda or something. I can’t even imagine what the prompt was but it had to be borderline deliberately malicious or insanely stupid or both. Link to the photo: https://www.reuters.com/resizer/v2/ODGUE6HF7JJ67G7S6RSK2Z5FB4.jpg?auth=60773c121a944f2d0a766dc323e18b15dd4282a853cc60e9cfe6835bd9a2792a&width=640&quality=80

@lauren@mastodon.laurenweinstein.org Probably worth noting that Schmidt is facing allegations of sexual abuse, spying, and corporate retaliation by his ex-girlfriend (https://www.latimes.com/business/story/2026-03-06/former-google-chiefs-spying-sex-assault-lawsuit-sent-to-arbitration) so his support for AI is probably the least of my concerns regarding the man. And this isn’t the first or only allegations of him being a gross creep either.

@carnage4life@mas.to An issue some of us have raised repeatedly with higher ups. You would think people that lived through "the grow the user base at all costs by making it free/heavily subsidized, then drive prices thru the roof" model would be aware of the dangers, but apparently higher ups also suffer from limited context windows.

@Migueldeicaza@mastodon.social @paul@tapbots.social Hm, to run a decent foundation model locally or to buy 15 years' worth of the most expensive Claude plan running on cutting edge servers in a data center. Tough choice.

@BasicAppleGuy@mastodon.social I feel like this sentiment nailed it, because whatever they’re doing now seems less discernible and useful: https://mastodon.social/@marioguzman/115890983018671432

@carnage4life@mas.to Even sadder to realize that there will probably be another figure shortly afterwards that we will be adding to this list, because when given the choice between modest improvements and stability or total chaos and regression, this country will generally favor the latter and repeatedly.

@ki @tek As others have noted, it doesn’t, but it is something that has been well documented, including in Karen’s Hao’s excellent book, Empire of AI (https://en.wikipedia.org/wiki/Empire_of_AI) These platforms leveraged RLHF, which often relied on English speaking, economically advantageous populaces with good connectivity to refine the feedback (Venezuelans were also used, but more so for image tagging/categorization).

@carnage4life@mas.to So we may need to supplement genAI usage with other forms of ML/AI. But if we take GitLabs’ statement at face value, you can extrapolate things out further and optimize: just take humans mostly out of the loop. Why have them direct AI at all? Why not have AI direct AI (as some are already doing/trying to do)? Just tell the directing AI agent what to do, and let its virtual software team do the rest and determine how to build things in machine optimized form.

@carnage4life@mas.to My gut feeling is that we’re going to end up swinging the pendulum back a bit, but it’s not going to be pretty in the short to medium term and it won’t revert all the way back either. There’s a lot of financial incentive, marketing, and myopic C-suite types that may overconfidently reduce headcount, find themselves in trouble, and then have to walk things back a bit.

@caseynewton@mastodon.social I guess my headline counterargument would be something like, "How Airbnb and decades of cheap credit led to the ongoing fertility crisis*" *half a dozen other factors will be added as footnotes

@carnage4life@mas.to I think it’s far too early for people to make pronouncements as to what the future of software will be — we’re figuring it out in real time. AI only knows how what & how to write because it trained on data sets, almost entirely originally created by actual people over decades. If that ceases, then AI will have to know how to create new architectures for new problem spaces — but it doesn’t actually “know” anything; it just makes statistical guesses based on what came before.

@caseynewton@mastodon.social Paywalled article, so can't read the full thing but my understanding is that the article points to economic insecurity (e.g., housing, job security, etc) as a factor and I would happily argue, that that would be the larger primary reason than just smartphones. That may influence things on the margins, but high speed connectivity has been availble in other forms (computers!) and young people constantly change how they interact with each other.

@paul@tapbots.social Just sounds like a win-win scenario -- they get extra compute and Musk gets a PR win for his forthcoming absurd SpaceX/allThingsMusk IPO. And to that end, they get to feign interest in 'orbital AI compute capacity' to lend that DOA brainfart of an idea credibility. FWIW, I don't pretend anyone involved in this AI space is remotely ethical or moral. They're not; they just make occasional noises in that direction for PR reasons.