Ten years ago, if it didn't understand what I meant, it told me so after 1-2 seconds.
Now, it'll show a loading indicator for 5-6 seconds and then do nothing at all... or do something entirely unrelated to my request (eg responding to "hey siri, how much is fourteen kilograms in pounds" by playing a song from my music library).
> or do something entirely unrelated to my request (eg responding to "hey siri, how much is fourteen kilograms in pounds" by playing a song from my music library
My personal favourite is Siri responding to a request to open the garage door, a request it had successfully fielded hundreds of times before, by placing a call to the Tanzanian embassy. (I've never been to Tanzania. If I have a connection to it, it's unknown to me. The best I can come up with is Zanzibar sort of sounds like garage door.)
I'm amazed more AI tools don't have reality checks as part of the command flow. If you take a UX-first perspective on AI - which Apple very much should - there's going to be x% failures to interpret correctly, causing some unintended and undesirable action. A reasonable way to handle these failure cases is to have a post-interpretation reality check.
This could be personalized, 'does this user do this kind of thing?' which checks history of user actions for anything similar. Or it could be generic, 'is this the type of thing a typical user does?'
In both cases, if it's unfamiliar you have a few options: try to interpret it again (maybe with a better model), raise a prompt with the user ('do you want to do x?'), or if it's highly unfamiliar, auto cancel the command and say sorry.
Apple shot themselves in the foot in the late 2010s by switching to deep learning methods and making things slower and worse, with the spell checker being the worst example.
I don't have anything in my music library, both Siri and Alexa (via a Spotify account I don't have) have responded to "${room I'm in} on" with their versions of "I can't find ${room name} in your music collection".
You're not wrong... but any space where Amazon, of all companies, has a shot at being the "most trustworthy player" is one I'm going to avoid where I can.
The killer feature of JXL is that most websites already have a whole bunch of images in JPEG format, and converting those to JXL shrinks them by about 30% without introducing any new artifacts.
Trusting some random VC-backed SAAS not to sell my data is (to me) as mad as trusting that the tide won't come in - it would be astonishing if they _didn't_ sell my data.
My bank has both commercial & cultural reasons not to sell my ID & transaction history. They might still do it anyways, but it's at least plausible that they wouldn't, if only due to the harm to their reputation if it ends up in the papers.
I'm a founder in this space (Fulfilled - posted above). Here's the reality:
You're right that incentives matter. But selling your data would be idiotic for us, same reason it would be for your bank in that trust is the entire business model.
If we want to monetize insights from aggregated data, we'd do it in-house and offer you better products. Example: Why sell your mortgage readiness data to some broker when we could source competitive mortgage offers and present them directly to you? Keep you in our ecosystem, add value to your experience, and build a revenue stream that doesn't destroy the core product.
The wealth space is crowded. Companies that burn user trust get exposed fast and die faster. The only sustainable path is treating your data like it belongs to you and not us. Any company here who doesn't get that is building on quicksand and I'd be very surprised to hear any of the larger players engaging in those practices but maybe I'm naive.
Either way, it's why we're a Fiduciary and that blankets the entire product suite.
"...Selling your data would be idiotic...same reason it would be for your bank [to sell your data] in that trust is the entire business model." I'm afraid that ship has sailed and taken your data with it.
I really don't think you read this article beyond the headline because that's not what it's about or implying...literally in the slightest.
That article is about JPMorgan being able to charge Plaid or other providers for the middleman access. They used to be operating almost for free, now Plaid has to pay for access the same way companies like mine pay Plaid.
So you don't think Plaid sells customer data? And by extension charging for customer data requests by Plaid and other aggregators isn't in fact charging for it?
So you didn't read the article and now you're just throwing out statements without validation? Great, if JPMorgan is selling your data then that's their decision and that's beyond the scope of this convo. We work with partners (Plaid, Snaptrade) who explicitly state that they DO NOT sell user data, and we maintain the same principles:
Here is the quote if you're too lazy to read this one too:
Does Plaid sell my financial data for advertising or marketing purposes?
No, we do not sell your financial data to third parties for marketing or advertising purposes.
Plaid only shares your data to power the services and products that you choose or to protect you and the Plaid network from fraud.
Plaid was founded on the principle that you have a right to your financial information and we are focused on providing products that allow you to safely and conveniently access your data and harness the power of Plaid’s secure financial network.
As Plaid develops more products and services, you may ask Plaid to share your information in ways that benefit you and that you control.
Plaid still rubs me the wrong way. Not selling to 3rd parties is great. But, everyone uses it, so that's still a lot of people getting data I don't necessarily want them to have. If I want to link a bank account to a credit card account in order to pay my bill, there's zero reason for that credit card company to have access to my bank transaction data. I still do the ACH deposit verification method where I can in order to avoid Plaid. I'd love more granular controls here or an audit log of what was pulled in.
SimpleFIN¹ looks compelling. Actual Budget can use that and it seems to work more like a privacy-oriented Plaid. But, now you need to trust a much smaller player. Really, I wish this were all standardized with strict privacy requirements.
Also, even if you trust the startup's current management and investors... a lot of those virtuous promises aren't enforceable in bankruptcy, and they may be in some legal jeopardy if they safely dispose of those "assets" before they are liquidated into the hands of someone without scruples.
I trust that the VC-Backed SaaS will be beholden to not wanting to get cut off from the bank(s), meaning not selling that data. If they sell aggregate anonymized transaction data, I'm also not sure I give a shit.
At a certain point paranoia gives way to practicality.
My threat model is one well placed technical employee who knows a buyer that will pay fuck you money. Judas can work at any organization and frequently does.
Yeah I was taken a back as well, "sell gold? To whom??" To the people buying gold, is not reasonable to have to pinpoint who exactly that might be given there is not shortage.
That is because of both of your biases (which is understandable) and me, not specifying why I am asking.
"it would be astonishing if they _didn't_ sell my data."
If the "who" was answered with "... to Kim Jong Un, glorious leader of the DPRK" you might raise a brow. That is probably not what they meant.
Maybe the author meant "... to the legal buyer of the new company, as part of the content of the company"? Maybe they meant "... to anyone willing to pay a price, legally or illegally"?
I wondered what scenario the author had in mind. It's a reasonable question. To me, the implications are quite different. You might of course disagree.
Obviously to anyone willing to pay a price. Have you ever looked at those cookie banners? "In order to give you the best experience, we share your data with our 759 partners..."
I’m not sure why you read their question as being either surprised or disbelieving?
I’ve worked in marketing enough to know that all sorts of less-interesting data gets bought and sold, yet alone financial data, but even if limiting the context to “for marketing purposes” I wouldn’t know, and would be interested in being told, which companies actually are buying that sort of financial information.
And as the user has now explained in a comment further down the thread, they weren’t limiting their curiosity to just marketing profiling so there’s an even wider scope to their question.
But hey, easier to just call them naive than to give them the benefit of the doubt and engage constructively.
MS _can_ do that, but only with new APIs (or break backwards compatibility). Wine only needs to keep up once folks actually _use_ the new stuff… which generally requires that it be useful.
Or MS does deals with developers causing them to use the new APIs. I still haven't forgotten when they killed off the Linux version of Unreal Tournament 3. Don't for a second forget they are assholes.
Plus if it does happen, folks need to laern a bunch of new hostile stuff, given how linux is taking off, why not just move to treating linux as the first class platform.
> why not just move to treating linux as the first class platform
This is where the argument goes back to Win32 is the most stable API in Linux land. There isn't a thing such as the Linux API so that would have to be invented first. Try running an application that was built for Ubuntu 16.04 LTS on Ubuntu 24.04 LTS. Good luck with that. Don't get me wrong, I primarily use and love Linux, but reality is quite complicated.
stability has critical mass. When something is relied on by a small group of agile nerds, we tend not to worry about how fast we move or what is broken in the process. Once we have large organisations relying on a thing, we get LTS versions of OS's etc.
The exact same is true here. If large enough volumes of folks start using these projects and contribute to them in a meaningful way, then we end up with less noisy updates as things continue to receive input from a large portion of the population and updates begin more closely resembling some sort of moving average rather than a high variance process around that moving average. If not less noisy updates, then at least some fork that may be many commits behind but at least when it does update things in a breaking way, it comes with a version change and plenty of warning.
MacOS apps are primarily bound to versioned toolchains and SDK's and not to the OS version. If you are not using newer features, your app will run just fine. Any compatibility breaks are published.
> Try running an application that was built for Ubuntu 16.04 LTS on Ubuntu 24.04 LTS. Good luck with that.
Yea, this is a really bad state of affairs for software distribution, and I feel like Linux has always been like this. The culture of always having source for everything perhaps contributes to the mess: "Oh the user can just recompile" attitude.
Wild to me that the first mention of COVID is this far down the page.
Most people have been infected at least a couple of times may this point, and at this point it’s very well documented to cause lasting cognitive decline.
This thesis contradicts the chart though. Why would older people be much less affected and the generation 70+ even show a negative trend if these people were far more likely to experience a more severe disease progression? You would expect them to be hit at least as hard (if not harder) as young people from those long term memory effects. The trend for the youngest age group also starts well before 2019.
First when you have combination of factors, this can happen.
Second, old people were more likely to die on covid. Kids were getting covid too, just not dying and long term covid consequences were observed in then. It can easily be that where old person died, young ended up with long term consequence.
There is no reason to assume the effect would be uniform accross generations.
-----
Either way, cell phone obsession and "rewiring of biology" claims are wven further from anything shown in the article. They are both purely what HN and the blogger want it to be.
I don't see your argument. Are you suggesting that covid pruned old people with weak memory to the point that it improved their memory on average? Because that is the only conclusion of your argument combined with the data. And that's not just completely unfounded, it's a pretty wild violation of Occam's razor.
Now, it'll show a loading indicator for 5-6 seconds and then do nothing at all... or do something entirely unrelated to my request (eg responding to "hey siri, how much is fourteen kilograms in pounds" by playing a song from my music library).