Despite the clickbaity title (in the original it EVEN has tabloid caps), actually a very interesting long form interview with Karen Hao, an MIT alumnus acquainted with several OpenAI execs since university, whose recently published Empire of AI[0] offers an informed, critical look at the AI explosion in Big Tech.
For example, I think this has the best exposition of OpenAI corporate history and Altman's part in it that I've encountered.
The interviewer, while annoyingly keen to inject his own political slant, at least asks some decent questions and knows how to sit quiet listening to the answers.
I've given this the (awful, clickbaity) title that it's under on YouTube, but it's actually a very interesting long form interview with Karen Hao, an MIT alumnus acquainted with several OpenAI execs since university, whose recently published Empire of AI[0] offers an informed, critical look at the AI explosion in Big Tech.
The channel is a Leftist UK media outlet with a capital L and the interviewer a self-professed Marxist, but if you can ignore or fight your way through the left spin there's plenty of informative and well presented narrative details about the history of OpenAI, its ambitions and methods and those of its CEO.
I've no inclination to click through to the video, but there are undoubtedly many youtube channels that use AI generated scripts spoken by AI generated voices in order to try and make a quick buck, following the advice of many "make-money-fast" videos on youtube, a substantial portion of which, no doubt, are themselves read by AI voices from AI generated scripts.
Whenever I encounter one in my recommends I use the "do not recommend from this channel" option, but I'm clearly fighting a losing battle.
Just one of the many reasons I think AI generated content should be required by law to be marked as such, and passing it off as human produced should be proscribed. When we drown in a sea of slop at least it will be clearly labeled.
Genuine question: does your bank not have a website you can use for that? (Wondering how your bank enables online for desktop/laptop users and/or folks without their app.)
(ETA:) My banks probably have apps, but I only use them via their websites, even on my phone (in a desperate pinch).
The problem is they usually allow 2FA through their app. I am not sure if this is the case for OP, but it is for most online banks. Luckily my bank still provides a physical authentication device.
I'm always amazed how all the banks I've used in Australia require either their own proprietary app or some bizarre closed source unvetted 3rd party TOTP implementation instead of just allowing TOTP.
Another solution is Conty[0] which is a download-and-run containerised Steam based on Arch; I use it to run games on Slackware without multilibs, and it's worked flawlessly so far.
However, it seems Arch are also dropping multilibs as a dependency of Wine, and moving to WoW64[1], with "reduced performance for 32-bit applications that use OpenGL directly".
What this implies for Steam on Arch (and hence for Conty) I'm not sure, though as of May some Proton versions have a PROTON_USE_WOW64 env var according to [2], so maybe multilibs can already be avoided running Steam natively anyhow.
> Humans being manipulated is not, in and of itself, a hint that things are going wrong.
Well, if this is meant to counter the abstract's claim that:
> Modern smartphones are designed to manipulate the attention and behaviour of users in ways that further the interests of the corporations that built them. In this they are importantly different from resources typically associated with the extended mind—such as notebooks, Scrabble racks and maps—which are not designed to manipulate or exploit users
I think it falls a little short of refuting a claim that in this particular case things are going somewhat wrong.
I admit I'm an edge case, but last year I installed Slackware on my main desktop specifically to escape the clutches of systemd, and by god it feels great to have a working /etc/rc.d again!
I do use containerised solutions for a few things (e.g. conty[0] for Steam, and a couple of rocm containers for messing with ML things).
KDE is the default desktop and does everything I need. I haven't used gnome for a decade anyhow.
We're supposed to think "oh it's an LLM, well, that's ok then"? A question we'll be asking more frequently as time goes on, I suspect.