Been using them for years. The price is reasonable too. It’s the only way I found to block ads everywhere on iOS (except the YT app, Mullvad’s Albania wireguard did that)
That makes a lot of sense. With the icon method: The user is in full control.
In our case, the CLI is part of a data sync system, so it does need a "special" once-per-hour check for new data (keeping local and mobile in sync).
I've actually been surprised how strict Apple is before it will happily let signed code run. But the average user (including me previously) doesn't know that. I think trust comes down to being clear up front.
Note to self: Add extra details/options to installer (WYSIWYG)
I moved to the Mac App Store version of OneDrive specifically because of the constant issues with the background processes it insists it must have running at all times.
The MAS version has only one of those and it works.
The issue is that someone can look over your shoulder and memorize your code, if you use a pin in public.
Some people will use the same pin for screen lock and BankID, and if someone gets their code and their phone they can cause a huge amount of trouble in a very short time.
Technically yes, but it seems like people's intuition for safe behaviour is worse across the board with modern technology, than it is with old technology.
Maybe it will improve by itself over time, but maybe not? There are a lot of things that humans generally suck at, and successful, human-friendly design must take that into account and work around it systematically.
Maybe phones/it/banking/payments/identity/authorization needs to be radically changed, in order to serve society well.
Passwords are "protected content" that may be hidden in a web form, but usernames are not. Your username won't be routinely hidden by any ordinary password manager.
You may have difficulty when the need arises to share a long string of random digits with another human, or in writing. This may not ordinarily be a thing if more LLMs and chatbots are doing customer service, but if it's a government service, one day we find ourselves at a field office speaking with a civil servant at a window that is armored with Plexiglas, and nobody has a pen to write with, or scratch paper.
One of my colleagues said he was so clever that all his "security question" answers were random hex strings. So I imitated him, and it was very easy with the KeePassXC password generator parameters to generate a hex string of 32 characters, and so "my first grade teacher" was 0xf7a83392b12b1029ac3be7e029acbe7e; my "favorite color" was 0xa83392ac3be7e029acbe7f7a83392b17, etc.
This turned comical because security questions are particularly designed to be asked and answered in human conversations. Over the phone, the pharmacy CSR was not amused, and I lost a few accounts thanks to this ultra-paranoiac attitude and activity.
My username doesn't fit in the web form. Instead of seeing "Symbiote" (which can be read and remembered immediately) someone sees "Sym-065c7d6c-e0b3-4d——" with half of the random characters not visible.
People in Denmark are expected to use this system to authenticate money transfers and to digitally sign documents (e.g. buy a house!).
My investment account allows up to 5,000,000DKK (€670,000) to be transferred with this digital signature.
Interesting. I've actually never enabled biometrics on it just as a precaution, and it's a different passcode than the phone, plus I use a random set of chars as the username like many others do.
I do wonder if there are some experts that are more likely to be hit. So if the normal optimised setup runs in 12GiB an you have 4GiB extra to spare, you could say “promote the most used X experts to this stable (old gen in GC parlance) region and don’t swap it out.
Maybe you could even do something like profiling and remember over multiple sessions (per project/workspace) what the most used agents are and load those up before hand.
I have this exact problem with 4.8 and Fable. Sometimes I can barely understand what it’s saying. I’m no english first-language speaker, but I don’t consider myself bad at English either, and it’s gotten increasingly hard to understand Claude’s claims and explanations.
Take it to sonnet 5 or gpt and ask it to explain this to a layman. If you still don’t get it ask it for the why it matters or the how it relates.
You can also ask fable/4.8 to do it but I find it helps to keep the working model surrounded by the complexity rather than drawing it out. Simplifying text is something that takes relatively low effort in comparison to technical tasks. Sometimes I use Gemini, deepseek, grok, and recently meta just to see if they have any added perspectives, sometimes they do. Meta is really good at turning a technical mess into a story that paints a picture in my head.
I hope it's actually talking to itself at higher bandwidth, and hope this is because of training that succeeds better given quality inter-agent communication.
For those who care to read everything rather than walk away, Fable can be extraordinarily dense. I suspect they'll pull the promo before I learn to read it.
Still, I feel I must read rather than correct it, as results are that much better if I let it do this "with" itself: orchestrator to agents back to orchestrator.
reply