Whether this is a breach or, as others have suggested, the LLM bullshitting you - for which I guess the politically correct term is still "hallucination" - it's bizarre to me that these tech companies who really should know better* continue the rushed and forced rollout of these "features" for use cases that are exactly the worst places to use them - where the information matters but a plausibly fluid prose kinda doesn't, and you're focused on something else meaning that factual errors and omissions aren't something you can check in the moment, but perhaps won't even think to check later because you have been led to believe you got the gist of the message while not able to check it
As someone with ADHD who could at times really benefit from a "personal assistant", the way tech companies are pushing these things sounds more like nightmare fuel than a godsend. And that's as someone who really does think there exist good use cases for these things, even as an increasingly vocal crowd not in the tech bubble is naysaying even that - also a predictable backlash to this godawful rollout
*Especially google. Like I watched them spend basically the entire time I was in grad school and many years after luring prominent professors in AI/ML with crazy salaries and their own labs
A rushed product pushed out before it's ready is a very small price to pay in the endless race for ever increasing profits. The tech companies who "really should know better" do know better. It just isn't something they care about or prioritize at all.
There is no issue with a company that someone's not gonna show up and say "Well of course they don't care about anything but profits" which in an environment where governments are systematically willing to backstop risky bets made with financial leverage, mostly means make whatever decisions align with the confidence or whims of the majority of an increasingly top-heavy pool of investment capital
I'm well aware and I agree. Systematically, this fundamentally means that no human being should extend any trust to any corporation in this economic environment, that no corporation can be trusted to build anything remotely resembling infrastructure (including any tool anyone will depend on for a period of time longer than one fiscal quarter). That no corporation can be trusted to keep any promise to anyone who is not a top-level investor
You are telling me, and I agree, that these large firms are dangerous animals that I can't reason with, that could decide to maul me at any time, and that my best bet is to limit my exposure to them as much as I can manage, moreso the larger they are. This means that it is in the interest of everyone who isn't themselves an oligarch to use every meager lever of power available to them to try to change this situation, up to and including killing these things when possible, because it is deeply difficult to avoid being under the power of these behemoths, and they do not and will never have your best interests in mind in any respect someone doesn't force them to
That's the logical conclusion of this thing people keep telling me is obvious as though that also means it's inevitable. But for some reason a lot of people who wanted to tell me how this all works seem either surprised or sometimes even hostile when I start talking about stuff like that
That said, the part they should "know better" about is that this use case makes no sense whatsoever for the technology as it actually works. I'd be really surprised if this kind of thing even positively impacted their stock valuation because it reeks of a flailing and incompetent vision for how to apply the technology. The most charitable I can be for a thing like this is to say "well they got caught up in the stupid hype and are just trying to throw an LLM into anything they can", which works for a lot of companies but not for one that's spent over a decade building world-class expertise and could be argued to have actually invented the specific underlying technology powering this hype wave. I get that the underlying motivation for any decision made by a company is that they think they will profit from it. This is tautological. I am saying that even from that perspective, this is conspicuously stupid in a way that only seems like it can inspire investor confidence if investors continue to buy hype that seems decreasingly tethered to reality
As someone with ADHD who could at times really benefit from a "personal assistant", the way tech companies are pushing these things sounds more like nightmare fuel than a godsend. And that's as someone who really does think there exist good use cases for these things, even as an increasingly vocal crowd not in the tech bubble is naysaying even that - also a predictable backlash to this godawful rollout
*Especially google. Like I watched them spend basically the entire time I was in grad school and many years after luring prominent professors in AI/ML with crazy salaries and their own labs