Ask Paul: September 20 (Premium)

Happy Friday! Let’s kick off the final weekend of summer a little early with another great set of reader questions.

Start me up

Christian-Gaeng asks:

Do you have any information about what Microsoft plans to do with the start menu in the future? The problem with interacting with the start menu has always been that it is basically always closed. I also think the latest idea of showing information about mobile phone notifications is nonsense.

I agree with you about the bizarre coming Start integration with Phone Link, which couldn’t look more ridiculous or poorly designed. But then, this is emblematic of the problems with the Windows 11 Start menu or, even more generally, of the weird focus on minutiae we see now from this team throughout Windows 11.

This is a tough one for me. On the one hand, I respect Microsoft looking at Windows several years ago now, acknowledging some complexities, inconsistencies, and functional redundancies, and then deciding to simplify the experience, knowing that in doing so it will upset many people. But simplification is one of those things that more makes sense in theory than in the real world. And on the other hand, it can lead to functional regressions. The goal should be two-fold: Make the right design decisions and give users a way to get back that one missing feature they need.

And that was the real failure of the initial version of Windows 11: It failed on both counts. This UI was/is simple, yes, but also overly simplistic, and it didn’t/doesn’t consider how people really use it. The two major sections, Pinned and Recommended, are really just vectors for ads (Crapware preinstalls in the former, “suggested” apps now in the latter). And Microsoft has even burdened the other parts of this UI with what are essentially ads, as seen in the “Account manager” functionality under your profile picture that, among other things, hides actual UI for signing out and locking the screen under additional steps, and the Search highlights screen that appears when you choose Search, which is nothing but Bing- and MSN-fueled content with underlying ads and tracking.

Another issue, and this one is a golden oldie, is that we all use Windows differently, and thus we have different expectations. We’ve not seen any hard data for this from Microsoft in many years, but there’s a big divide here. Some percentage of people (I’m in this camp), pin their most often used apps to the Taskbar and treat Start as a place for launching secondary apps, either by clicking a shortcut icon or with Search. (This is like the Home screen and All apps screen on Android, or Home screen and Library on iOS.) Others go to the Start menu first. Others may do other things, who can say?

The Start menu doesn’t ideally help anyone. You can’t fully customize it by, say, removing Recommended and only using Pinned, or by going directly to All apps. In short, it forces you to adapt to it, and to work with its limitations, all of which are self-inflicted (by Microsoft).

Getting off my high horse, none of this impacts me directly all that much. As noted, I launch the apps I use most often from the Taskbar, and as a keyboard shortcut and keyboard user in general, I will tap Start and type the first few letters of an app name when I need to use that to start a secondary app. (I also use keyboard shortcuts like WINKEY + E to launch File Explorer or WINKEY + I for Settings, and so on.) As such, I barely customize Start anymore (something I gave up when it became obvious that spending time orienting live tiles in Windows 10 was pointless). And barely see it.

One common criticism here amounts to “why does Microsoft waste time/effort on something like this when Windows 11 needs more serious work elsewhere?” But the people who can do the Start menu work in question aren’t exactly sophisticated, and they’re in capable of doing work that matters. Even in the confines of their stupid little world, I’m sure they’re confined by Microsoft’s corporate needs/aims, which is what leads to the enshittification in the first place. So it kind of doesn’t matter. They can keep dicking around with these little things, and will. But … whatever. There’s no version of the Start menu that would inspire me to use it more anyway.

Coding AI, the final frontier

spacecamel asks:

With all of the talk that AI is going to take all of the programmers jobs, have you tried ChatGPT to see if through only prompts can get a notepad app for windows that is sufficiently close to your great work? I think some of the details that you have with functionality would be a little too difficult to produce but I think a layman could get something about 75% of your work.

I’ve not used AI for anything coding related yet, but I will. And I will do so in the context of that .NETpad modernization project, the first part of which is about to wrap up with a bit on fine-tuning the UI and cleaning up the code. That’s gotten me thinking about something that’s been in the back of my mind for a while now, which is Copilot for GitHub, and how it might help with optimizing the code in various places. I am very curious about this.

I also have free access to Gemini Advanced as part of a one-year Google One AI Premium upgrade I got for buying the Pixel 9 Pro XL. So I will experiment with that as well.

Related to this, I coincidentally came across an article in The Register last night that relates to this, which I found quite fascinating. (It’s the second part of a two-part series, the first of which is here.) It’s really about how human-made software will become more valuable once AI does everything for us, but the bit I cared about was how the author envisions the future unfolding: Human write all software code now, but no human can completely understand big software projects, like OSes (which was a problem as far back as the original NT, in which only two members of that team, Dave Cutler and Mark Lucovsky, allegedly had any real understanding of the entire code base), and this leads to us just adding code to these systems, not rewriting or fixing them. AI can help with that, and will. And it will write new code to replace old code, fixing issues, and working as a helper to humans. In time, though, AI will replace human coders, first with human oversight. And then it just becomes automatic, when the successes rack up. Just doing something like taking an existing code base and recasting it in a safer/better language will be tremendously useful (as compared to the manual process of updating small subsystems one at a time in the Linux or Windows kernel, as we do today, by hand). It’s a fascinating thing to think about.

Anyway, the AI tools we have today–ChatGPT, Gemini, Copilot for GitHub, or whatever else–are very early in that predicted cycle, and are just basic helpers now. But I still believe they can be useful, and that this type of thing is ideal for AI in its current state because this work can be grounded in finite and well-understood data sets (specific languages, frameworks, so on) to avoid hallucinations. And I am eager to get started on that to find out.

Thanos-ing web ads

wright_is asks:

Have you tried Safari in macOS 18 yet? Have your tried out the “Thanos-Snap effect” for removing unwanted content from a web page? (Remove distracting elements)

Yeah. It’s interesting. But it’s also a manual process. And on the sites I’ve tested it on–including this one, sadly, I wish this were different–it’s not clear that you’ll ever get ahead of the ads since they keep coming back in different forms. The Thanos-like animation is fun, but I think “real” ad blockers are still the better choice and that the two used together is perhaps ideal. That is, traditional ad- and tracker blockers work well, but they often miss with some things I find annoying, like video pop-ups in the corner. So using this to get rid of those too is nice.

For those not using an adblocker, it is a game changer for displaying web pages. Jason Schnell said on MacBreak Weekly that some sites have “normal” static ads under the video ads, so getting rid of the distracting video ad leaves behind a less distracting ad, a win-win, maybe if advertisers see this happening a lot, they will stop with the stupidity of video ads on a text page?

Obviously, people will try to use this feature to get rid of ads. But it’s really about removing any elements of a web page, including actual content. Which can have unintended consequences if you mis-click or whatever.

I could see Brave and Vivaldi doing something similar, but I doubt we will see Google Chrome or Microsoft Edge mirror this feature.

Same. This is a nice escalation of the ad- and tracker blocking features in those browsers, and exactly what Microsoft and Google do not want. Though Edge does at least have a nice Reader view for removing distractions, it’s just that it’s manual. This is another thing Safari does well: You can configure it to automatically use Reader mode with specific sites. I probably spend an inordinate amount of time trying to read without distractions. It’s one reason I use Pocket, even if I want to read an article immediately: Save it to Pocket, go to Pocket, read it there in a clean view.

On iOS it seems to be an automatic removal tool, you can’t hit a bad ad and watch it disintegrate off the side of the page, like it does on macOS.

Yeah. I think the auto-Reader view thing has been there for a while as well.

For whatever it’s worth, Jason Snell’s example video of this feature is classic to me because those freaking Callisto Project videos on Ars Technica have been driving me crazy for weeks. Nice coincidence there.

Also, I was just chatting with Laurent that I wished Apple had kept Safari on Windows. I know that Brave, especially, satisfies this role, but the world needs more truly minimalist browsers that are just private and secure by default. What puts Safari over the top compared to Brave is the design attention to detail. Brave is a blunt object in that regard, and I wish it had Apple’s design sense. I guess this is yet another potential reason to buy into the Apple ecosystem.

Hibernation

gg1 asks:

Paul, the Windows Insider program has been very quiet lately. Excepting a small Windows 10 build, there have been no releases in September. The Dev channel is worse, with no releases for a full month. I suspect it’s somehow related to the impending widespread availability of Windows 24H2, but an impending release has not slowed down Dev builds before. What do you think may be happening here?

I think you’re correct. This is essentially the quiet before the storm of a Windows 11 feature update, and that things will get back to their normal chaotic nonsense once 24H2 ships in October. The Windows 10 stuff will always be minor, and will always involve something familiar from Windows 11. But I do expect to see more happening with Windows 11 once we get past this milestone.

Managing multiple phones

jrzoomer asks;

Paul how do you manage having both an iPhone and Google Pixel in your life? I assume only one of them has your primary number (I believe on Google Fi)? Is the other used as a data only device? Do you carry around both when out and about?

This is one of those “it depends” things.

I only have one primary phone at a time. I started using the iPhone 15 Pro Max again when I bought the MacBook Air back in March, so that derailed my plans to move back to the Pixel 8 Pro after reviewing the Galaxy S24 Ultra, as per my original plan. But since then, I got the Pixel 9 Pro XL, so I’m using that now. And my wife has upgraded to that S24 Ultra belatedly because I needed to finish some Phone Link work (for the updated book chapter and a Hands-On Windows episode, the latter of which I needed to re-record in July after we got back from Mexico).

And of course the iPhone 16 Pro Max is arriving today, adding a new wrinkle. My plan there is to get it configured with apps and data, but stick with the Pixel 9 Pro XL until my review is done. And then swap over to the iPhone for that review period. And then we’ll see.

Swapping phones has typically been easy, regardless of whether I was using a physical SIM or an eSIM, with the exception that T-Mobile prevented me from doing this myself when I was using that service. But I switched back to Google Fi, and all-eSIM now, and it’s easy again. So there are no worries there.

Traveling internationally adds an additional quirk, but the advent of low-cost international data SIMs makes that simpler. I use Nomad a lot, and have tested some other services like Airalo. I would get a U.S.-based data eSIM for the iPhone so I could use it out in the world more easily before I switch over, but we’re heading to Mexico in about 10 days, so I’ll probably just wait. But I’ve done this temporary eSIM thing several times for testing, and it can be inexpensive.

When we went to Berlin recently, I did get a Nomad eSIM as a backup, and then had to use it when I realized that I forgot my current Google Fi plan only works fully in Canada and Mexico, so that ended up working out nicely. My wife also got a Nomad eSIM for her phone while there because it was cheaper than using her Verizon data. For this coming Mexico trip, I’ll use the Fi eSIM in the iPhone because I’ll have switched by then. But I’ll get a Nomad eSIM for the Pixel too.

So the short answer is, yes, I only use one phone with phone/texting services (and data) and then use pay-as-you-go eSIMs as needed for secondary phones, swapping back and forth if/when needed. I don’t typically carry both. But if I’m reviewing a phone, I do that as needed. And when we’re in Mexico, I often ask Steph to carry a secondary phone in her purse if we’re doing something that may benefit from having it. But generally, one phone on me.

eSIM city

jrzoomer asks:

Paul what are your thoughts on eSIM vs SIM cards, which do you prefer?

I definitely prefer eSIMs because physical SIMs can be damaged or lost, and they’re difficult to handle with my large hands. As alluded to above, the ability to auto-add/provision and, conversely, disable eSIMs on a phone is wonderful, and it makes switching devices fairly seamless. When we were in Berlin and I realized Google Fi data wasn’t going to work, I switched to the Nomad eSIM for data in under 30 seconds and was up and running. That’s fantastic.

My only temporary issue with eSIMs is/was specific to my situation: Mexico carriers were slower to adopt eSIMs than was the case in the U.S. and so having a phone like recent iPhones that only supported eSIMs was problematic. But that’s corrected itself, and if I need to use a Mexico carrier line with a phone number and data, that’s easier now too. So I’m good with fully switching to eSIM now.

Usability vs. features

TheJoeFin asks:

Edge (and Chrome) have been building more advanced tab/browser management for some time now. Features like:

• Workspaces
• Tab groups
• Split windows
• Vertical tabs
• Side bar

I’ve tried all of them but have only really found Workspaces to be useful and natural enough to use regularly. Have you adopted any of these in your regular workflow?

I was literally just thinking about this.

The short answer is no. But I do look at this stuff, of course, and think about how it might be useful.

One thing I do semi-randomly when I review laptops is use different web browsers, and with the HP OmniBook Ultra I’m currently using, I’ve been using Microsoft Edge for better or worse. I literally disabled the Sidebar (which feels like a new option, maybe I’m missing something) and the Split window feature, I hid the Workspaces button (which I always do, this doesn’t sync), and I never use vertical tabs or tab groups (I hide the Tab actions button, but that one seems to sync). In short, I turn this bloated thing into something more minimal from a UI/features perspective. (There’s more to this, actually. I disable that weird Edge-specific right-click menu, for example, and make many other configuration changes when using this browser.)

In How I Configure Brave, I described which toolbar buttons I hide in the browser I prefer to use, but at that time, there were a few oddities: You couldn’t just right-click Brave Rewards and choose “Hide,” you had to click it once then do that. And you couldn’t right-click to remove the Sidebar button: You had to configure that in settings for some reason. Since then, this is easier: You can right-click to remove the Sidebar button, and though there’s a new button for Leo (Brave’s AI), you can right-click to remove that too. So this already minimalist browser is very easy to strip down, which I really like.

I use a custom (minimalist) New tab page in every browser–Bonjourr, currently–which came up somewhere recently, though I can’t seem to find it. (I’m looking at Tabliss right now as well.) In short, I do everything possible to make whatever browser I’m using simpler and less busy. I don’t find any of the piled on stuff particularly useful. I have feeds, I read articles, I save some for later with Pocket. That’s pretty much it.

Please ignore that giant sucking sound

JustMe asks:

Laurent reported this week that LinkedIn is changing its terms of service to use user data for AI training by default. The user will have to opt out if they don’t want this to happen. While I can easily see Microsoft, Meta, and Google doing this, how widespread do you think this will become? Will we all be stuck in a world where part of signing up for any service online will be to necessarily opt of AI data usage (assuming that even bothers anyone but folks like your audience)? Will the normal user even care?

In 1999, Sun Microsystems CEO Scott McNealy said that consumer privacy issues were “a red herring” and that, “You have zero privacy anyway. Get over it.” This was controversial even then, and for all the obvious reasons. And it should be noted that he said this at an event launching Jini, a Java-based network device discovery service, and that he was perhaps trying to address a potential complaint proactively. (Or a then-recent complaint that Sun had received about privacy issues.)

I always liked McNealy for his plain-spoken honesty. He was a vocal critic of Microsoft and Bill Gates, especially, and his (and Sun’s) competitive retorts to Microsoft and its products were always amusing and usually correct. But I don’t like the idea of giving in to Big Tech on privacy or anything else, for the matter, especially these days. It’s an open secret that all major AI systems have stolen data by silently scrapping it from wherever it could.

The only silver linings I see to the LinkedIn news is that they’re at least not taking the data without telling customers, and they at least do let you opt out; I suspect many services are silently using this data now. Microsoft is big on opt-out as opposed to opt-in–defaults matter, etc.–so I guess that doesn’t surprise me. But anyone who uses LinkedIn will tell you that it’s a cesspool. It’s the only social media service I know of that allows businesses to pay it so they can deliver ads to other users via an internal messaging system that appear like normal messages. It’s a disgrace.

AI is useless unless it can be trained on data, preferably lots of it and even more preferably good data. So any entity with a source of good data–really, any entity with any data–will see this as a potential windfall. Trustworthy and non-trustworthy services will line up on either side of this debate, as they will. But LinkedIn is owned by Microsoft, which is one of 3 or 4 companies that will benefit the most from this data theft. So in that context, what LinkedIn is doing is arguably as close as it can be to a best-case scenario. Which stinks: We should have laws that prevent this and force these companies to make us opt-in and perhaps even offer financial incentives to do so, given how much money is at stake.