I've been trying to reduce my reliance on all things Google, and one of the big ones is password management. I've tried several times to make the jump, but every time I start researching options I'm overwhelmed by the selection. There are a lot of popular options out there, and I really don't have the time/energy to endure a misstep. So without a clear idea of which manager will check all of my boxes, I end up bailing on the process and keep using chrome's built in option.

So to start, here's what I like about Chrome:

• Automatically offers to store passwords without extra clicks
• Autofills automatically where it can, and gives me an easy choice when it can't
• Works everywhere I need passwords. (basically everywhere I browse the internet since chrome works everywhere)
• Minimal overhead. This is hard to beat since Chrome just includes it, so I'm fine with a little extra setup if necessary.

I used to use keepass portable on a thumb drive (I want to say circa ~2009ish), but it became really inconvenient as my usage shifted more to mobile devices.

I see this as a first step to also reducing my reliance on Chrome so I can start to consider other browsers. Right now I feel locked in to Google's ecosystem, but I know I can break it up if I don't get too bogged down by choice. Much appreciate any help. :)

Playing with Llama 7B and 13B, I found that the 13B model was capable of doing a simple task, rewriting titles in sentence case for Tildes submissions. The 7B model doesn't appear capable of the same task, out of the box.

I heard about Android's new AICore available on a couple of new devices. But it sounds like Gemini Nano, which runs on-device, can only handle 2B or 3B parameters.

Is this size of model useful for real tasks? Does it only become useful after training on a specific domain? I'm a novice and wanting to learn a little bit about it. On-device AI is an appealing concept to me.

For my job I need to copy a person's name out of a field and split it into first and last names. I have a AHK script that can do this but my version only works with simple, one-word, non-hyphenated names. I tried prompting Gemini as below.

There is a line of text consisting of one or more words, a comma, then one or more words. write an autohotkey 1.1 script to select everything before the comma, store it to a variable called firstname, then select everything after the comma and store it to a variable called lastname. thanks

(the firstname and lastname variables should be swapped as the last name would come before the comma, that was my error)

Here's the AutoHotkey 1.1 script to achieve what you described:

; This script splits a text line into first and last name based on a comma

#NoEnv  ; Recommended for performance and compatibility.
SendMode Input  ; Simulates keystrokes for reliable input.

; Function to split the text
Func SplitText(_text, firstName, lastName) {
  StringSplit, parts, _text, `,`
  if (parts[1]) {
    firstName := parts[1] ; Assign first name
    lastName := SubStr(_text, 1 + StrLen(parts[1])+1) ; Extract last name after comma (including space)
  } else {
    firstName := _text ; Handle cases without a comma
    lastName := ""
  }
}

; Hotkey to trigger the script (replace F2 with your desired key)
F2::
  Send, ^c ; Copy the selected text to clipboard
  clipText := ClipboardAll  ; Read the clipboard content
  SplitText(clipText, firstName, lastName) ; Call the function to split the text
  MsgBox, First Name: %firstName%`nLast Name: %lastName%

  ; Optional: Clear clipboard to avoid unintended pasting (comment out if not desired)
  ; Clipboard := 
Return

Explanation:

#NoEnv and SendMode Input: These lines are included for performance and reliable input simulation.
Func SplitText(_text, firstName, lastName): This function takes the text (_text) as input and assigns the split parts to variables firstName and lastName.
StringSplit, parts, _text, ,:** Splits the text based on the comma (,`) delimiter.
if (parts[1]):** Checks if there's a second part (i.e., text after the comma).
If true, the first part is assigned to firstName and the text after the comma (including the space) is extracted using SubStr and assigned to lastName.
else:** If there's no comma, the entire text is assigned to firstName and lastName is set to an empty string.
F2::: This line defines a hotkey. When F2 is pressed, the following actions occur:
Send, ^c:** Copies the selected text to the clipboard.
clipText := ClipboardAll:** Reads the copied text from the clipboard and stores it in the variable clipText.
SplitText(clipText, firstName, lastName):** Calls the function to split the copied text into first and last names.
MsgBox:** Displays a message box showing the extracted first name and last name.
Optional: Clear Clipboard: The commented line Clipboard := clears the clipboard content after processing. Uncomment this line if you want to avoid accidentally pasting the copied text elsewhere.

Edit: To answer questions, and provide info I probably should have from the beginning: I'm a medical transcriptionist using AHK with my job's software (Emdat Inscribe) and a word expander (Instant Text). Inscribe seems to be browser-based but functions as a standalone app. I already have dozens of AHK 1.1 macros; my .ahk file has over 2000 lines, although I'm sure it's terribly written code. This is also why I'm disinclined to switch to AHK 2.0 since I have no desire to rewrite everything unless absolutely necessary. The first part of this macro is a bunch of keypresses to jump to the relevant patient info field, then this part would be used to store the names appropriately. I already have hotkeys to use the variables as needed and most macros are limited with #ifwinactive to Inscribe.

I am and have been on a bit of a quest to make my computing devices suck less.

Over the last few years I have migrated all of my laptops and desktops over to various Linux flavors. My experience with this process is that each flavor has their own quirks that need to be ironed out, but after getting things running there is little in the way of maintenance. Things kind of just work nowadays.

I have been looking into getting something like (but not necessarily) LineageOS on my phone. As I am looking into this and reading forums on the subject, it seems like a perpetual arms race between application developers and application users. One puts in a way to check for root, then there are root hiders, then there are root hider checkers, then there are root hiders that you build with custom names, etc.

I want my device to not suck.

I don’t want to be going in and fighting with my banking applications every time there is an update. I am totally willing to fight a painful setup once.

Is a custom phone OS something that is essentially only viable to use if you are driven by spite? Am I reading too much into the struggles that are posted in various forums? I am looking for any input for anyone that has used a custom OS short or long term.

I doubt I'll find any new ideas, but maybe someone here has one. I'm running out of places to turn, with no solution.

I have been in a fight with Fitbit support for a few weeks now over their push to migrate everyone to a Google-linked Fitbit account. I'm pretty sure what I've found is a rare edge case of a bug, or rather, an unaccounted for set of conditions when trying to migrate.

A long time ago, 2017 I think, I created a Google-linked Fitbit Account (via oauth, "Login with Google"). I used it briefly and then stopped, and completely forgot about its existence.

In 2022, I got a Pixel Watch and created a new/second Fitbit Account with a different email address, as I did not remember I had a Fitbit account already and I wanted to use a masked email address through my personal domain.

With the push to migrate all Fitbit accounts to Google accounts, I decided to try to do so last month. When I attempted to migrate my second account to a Google account, I got an error that I already had an account registered under my Google account. So I logged in to that old 2017 Google account and initiated its deletion. It told it me would take 30 days to delete it, so I waited 31 days.

31 days later, I tried to migrate my second account to a Google account. When I try to do so, I get an error:

Can't use Fitbit with this Google Account
This could be because you're using a Google Workspace account, or because your account is supervised.

My account is neither a child account, nor a Workspace account, it's a standard (adult) account I've had for something like 16 years.

So I tried to see if my old Google account was perhaps not deleted after all. I tried to log in to my old account via oauth (Sign In > Continue with Google), and I get a different error:

Sign in again to continue
Since you deleted Fitbit from your Google Account, you’ll need to sign in again as a new user.

I suspect that what happens when I deleted my old/original Google account is that it wasn't actually deleted, but made inactive with some "deleted" flag, but the account hasn't been purged. As a result, I'm unable to migrate my new account to the same email address I used for my old Google account.

Reaching out to Fitbit, they continually put the blame on Google for reporting my account as a workspace/supervised account, and the only solution they'll offer me is "You should create a new Google account". Google has also been unable to help, but that doesn't surprise me, as I don't think it's an issue on their end. Requests to Fitbit to escalate my case to a higher tier of support and/or someone from some type of database team have been stonewalled, and I think that Fitbit support has now just stopped responding to me entirely.

Does anyone have any idea where I could turn?

I opened up the Google Earth app on my phone wanting to browse random beautiful and interesting places, but it doesn't seem to have a good way to do it. (Maybe this is a limitation on the mobile app, and the desktop app is better?)

For example, if I try searching for "wetlands", it only shows me a list of maybe 10 places near my current location.

I tried searching the web too, but I mostly get "listicles" like "10 amazing places on Google Earth" but they're practically unreadable with the webpage being covered with ads.

I found Earth View Gallery https://blog.google/products/earth/most-stunning-images-from-google-earth/ and it's nice but it's just a gallery of images with no contextual information (at least when viewing on mobile). It shows me beautiful pictures but no information about where the picture is from... I would have expected it to link to the spot on Google Earth.

Anyway if you happen to have some recommended places for me to check out on Google Earth I'd love to see them! I like:

• wetland-type habitats like marshes, pond systems, mangrove swamps
• beautiful natural scenery in general
• abandoned and/or ancient architecture

But open to any interesting recommendations in general.

Edit: It does seem to be slightly better on desktop. Searching on Google Earth works better, and Earth View Gallery does link to the location on Google Earth. I wish Google Earth had like, a built-in community feature for me to check out other people's public projects.

I have a Pixel 4a which has just reached end of support for Android. However, I love this phone and the only hardware issue is that the battery doesn't last me scrolling social media all day, so I am not looking to upgrade to a newer handheld just yet.

I've been looking into Graphene OS and Lineage OS as perhaps alternatives I should consider, at least just so I can keep getting security updates. However, looking through GOS, they say that their 4a build is an "extended support" build different from the main OS which is described as a "stopgap" before upgrading phones. LOS says it's supported but through an automatically generated page which doesn't leave me with much confidence about the attention and stability of the build on my particular phone.

I'm asking y'all's opinion on whether I should even bother. Security upgrades are important, but my phone is a secondary device at best, one which I always use with the same apps and websites and honestly not really that much of a security risk. Watch hubris get me.

I've always been a bit of a privacy enthousiast. Have had everything blocked that Google and by extension YouTube wants to scrape off you. This means I've also blocked my view history.

Recently YouTube started giving out a warning on the homepage that you have blocked your view history, that you can change it in your privacy settings and that it helps them serve you better content. What it also means is that your homepage is just one big popup to guilt trip you into sharing your data. The homepage won't show any suggested content anymore.

While it is in their interest to do so and since they are a company wanting to make money it is understandable. Nevertheless it seems harsh from going to see content that you might like to only seeing a big warning sign right now.

What are you experiences with this?