I’ve been using Claude to build native versions of a couple of apps and what was once unthinkable (maintaining multiple code bases) is now fairly trivial. And Electron/Tauri implementations are high quality.
I’m not sure platforms like Maui are necessary anymore.
I did note the comment “if you don’t want Liquid Glass” as a direct response to GenAI native development.
I just don't think people like having something shoved down their throats. The dedicated Copilot button on keyboards and adding Copilot shortcuts all over the OS (and automatic popups/ads) was far too far.
I think OS level integrations that are opt-in, not opt-out, may even be popular. But they have to be done carefully and tastefully.
I have the same feeling about any kind of integration. We're moving away from Google because we simply do not want to have this kind of forced relationship with products and/or services. It either fits and we'll pick it or it does not and then we don't. We won't pay for things we do not intend to use. And we don't want exposure to products that may constitute a security or a privacy risk.
Email and document collaboration are the big ones, email is probably going to be the easier one of those two, documents much harder because we have a pretty specific workflow that is tied closely to how google docs works. But the decision has been made and I don't care if it is going to cause us to have to work a bit slower or different, this is just unacceptable.
The whole Gemini thing is just a massive embarrassment for Google I really can't follow their thinking, you'd think that after the Google '+' debacle that they would have learned their lesson not to cannibalize your old products to launch a new one.
I actually think Gemini Pro is great and I don't have a problem paying for it, but I don't want its tendrils in Drive and Gmail or anywhere else, it actively damages the product experience there. Everywhere they've tried to integrate LLMs, it generally provides an experience that's inferior to just chatting with Gemini.
The closest to useful it's been is in the GCP console, but it seems to decide at random to forget context, and it might just be Gemini Flash with minimal thinking, which tends to mean it's just repeating things it's already said.
Around new years my company had to replace my windows laptop because windows update has been broken for a few months on my machine. They had a replacement windows laptop ready but I asked them to provide a MacBook instead. This is first time in my two decades of career that I specifically asked for a MacBook.
Funnily enough, there's a bug that's affecting all MacBook users in my company (does not wake after lid down overnight). Apparently the culprit is windows defender installed in the MacBooks. Corporate, you know...
And if the surprise is unpleasant you can disable it by turning off memories and holidays in the settings of the photo app. Not so easy to escape Copilot on Windows.
That's not nearly comparable tho. I don't care it's watching my photo as long as it doesn't annoy me when I want to watch them. Copilot is everywhere, you got to actively avoid it like the plague it is.
Even the linked blog post indicates that that is not the case. Windows has Copilot buttons on practically every built in application, a taskbar icon, and a dedicated physical keyboard key that people commonly accidentally hit (contractually required for OEMs to provide). They also actively promote Copilot in the OS (particularly Home Edition with nothing disabled e.g. "Tips," Notification Spam, Recommendations, etc).
Nobody can predict what Apple will do tomorrow, but as of today, they aren't really pushing Siri/Apple intelligence really hard particularly after initial setup. None of most of the above for example.
I have Pro Edition and for me Copilot only added two icons. One in Notepad and another one in Paint. I ignore both. There's also the Copilot app that I didn't even know I have installed.
I don't know what happens with Home Edition, but I though the pushback was mainly from Insider Preview?
You want to take a look at Microsoft office, my bad Microsoft copilot 365...
You can't even select a cell on notepad without a freaking copilot button pooping up every single time. Same on word, that's maddening !
You could argue that windows isn't Microsoft copilot 365, but then, why do people even use windows ? It's always because of the office, my bad, copilot 365 suite.
You can also get rid of both of them very easily with O&O Shutup 10++ (or any of many other GUIs or scripts designed for the same purpose of decrapifying Windows). I toggled off Copilot and Onedrive and haven't seen either in all the years I've been using Windows 11.
I'm not GP, so I can't comment on where their line is, but for me the difference between Copilot and Apple Intelligence is that I can turn off the latter and never see anything about it again. Copilot, on the other hand, is everywhere and it's almost all universally buggy garbage, even when it's disabled.
I actually trust the Apple Intelligence, when off, doesn't exfiltrate my data.
Oh yeah and even when you turn off preferences/settings/features in Windows, they mysteriously come back later in one of the unilateral forced updates, against your wishes.
> So the issue isn't actually that it's baked into the OS, it's that you should have control over when it's used.
Baked into the OS implies that it's integral to its operation in a way that the two are fundamentally inseparable. Having a global off switch implies that's not true.
There are other irritating baked in aspects of the newest macos and other recent versions that are arguably less avoidable, like Tahoe's entire UI design, or the Settings app.
The real issue is copilot is implemented in their apps inconsistently. Very clear there’s little cross app planning. Apples solution is global and apps and hook into it or not. And if you turn it off apps done break.
Why would Apple Intelligence bother them? It's very unobtrusive and actually useful when it's visible. I literally don't notice it except when it's helpful.
I don’t think I’ve ever even noticed Siri/Apple intelligence on macOS. I’ve disabled it somehow (probably at install) and have not heard about it since
Have you even tried it? I'm a Mac user for 20+ years and I'm running Tahoe. Not once have I ever thought about Apple Intelligence. I don't even notice it. I think you have to switch it on.
Microsoft directors want Copilot so they can make the case to executive leadership that they're aligned with that vision. It's why even in this announcement, the admission that they've maybe taken the whole AI OS thing a bit too far is phrased positively for AI with "Integrating AI where it’s most meaningful, with craft and focus", so the skim reading exec or financial journalist can read it as "good, Windows is still integrating AI"
Or onedrive integrations and constant ‘backup your computer now’ popups which are _advertisements_ for onedrive, or Netflix, Spotify, or LinkedIn pre-installed and difficult to remove, or all of the above reinstalling during windows updates.
In fact, basically any feature added since Windows 10 is probably unwanted.
Comparing windows to an OS I don’t use isn’t a fair comparison unless my work machine stops being windows. I assume Apple are a slightly less variant of bad though
My iCloud is full. Every once in a while my iPhone nags me to upgrade for a few days in a row and I tell it no and it goes away for 6 months or so. My Mac has never once nagged me about iCloud storage.
I have been on a MacBook Pro exclusively for the past 3 years and I do not ever see anything about iCloud. I also never signed up so may be that is why?
I don’t care if it had the best UX of all apps on windows. I don’t want or need data scraping in the form of cloud storage.
Edit: but I am somewhat surprised that it’s qt and not the typical react electron bloat that Microsoft is slopping out. I’m sure it’s only a matter of time.
Yea, I've replaced Windows with Ubuntu on my pc and have just ordered an M5 Macbook Air.
Sure both have their quirks, but it's just wild how much Windows goes out of its way to be annoying. From a billion startup notifications to basic UI stuff to copilot and the list goes on.
The funny thing is there's plenty of things I think to myself, it would be nice if AI could do this, but instead it's all top-down "this is what we think you should need"
No, I don't need you to summarize a two sentence email. How about I move emails to folders and you start to learn the patterns? Or which alert emails I want to ignore? Or who asked me something last week and I forgot to respond? Or which emails I should look at first after a vacation? Etc.
"Hi, I'm Copilot! I can do lots of things. Do you want me to make interpretive art from your TPS report? How about writing a thousand meaningless words in a style you wouldn't be caught dead using? I can also make fake pictures of your company Initech where your logo says Imglarb!"
I put bazitte on my desktop and left it connected to the TV. It's the most seamless linux distro I've ever used. Been using it for a year now purely with an xbox controller.
some are, it depends, but i'd expect to lose access to those type. if it mattered and crossplay existed for the game i'd get a console if it was gunna be a big deal...
I switched to Mac around Vista and never looked back. For games, enlightenment is realizing the PC gaming tribalism is dumb and PlayStations are actually really nice. It's an appliance that plays games without giving you trouble, in a comfortable place instead of encouraging you to spend even more time at a desk.
If your interests lie entirely or mostly in the realm of AAA or AA games that are playable with a controller, then I completely agree.
However if your interests lie in indie games or games that require a keyboard and mouse interface (precision shooters, grand strategy games, RTS games, etc) then having a PC that can play games is completely necessary. (I say this as someone who runs linux btw, not a windows defender).
If integrated properly, something akin to copilot generating Mac shortcuts, with close supervision, copilot could be extremely powerful on the desktop. Now that Apple has licensed Gemini, I would expect that to come soon.
Gen AI has even more power at task generation than at content generation. Imagine running Photoshop or Final Cut Pro via prompts. People seem squeamish because so far the Copilot entrypoints have been encouraging tacky text & image content generation, like Clippy. But imo that’s the weakest and most sensitive application.
V1 is often not very good, for any new application.
I don't think macOS will liberate you from OS-level integration with AI. If you really cannot tolerate built-in AI, Linux and the BSDs are your only choice.
I continue to jump into these discussions because I feel like these upvoted posts completely miss what’s happening…
- guardrails are required to generate useful results from GenAI. This should include clear instructions on design patterns, testing depth, and iterative assessments.
- architecture decision records are one useful way to prevent GenAI from being overly positive.
- very large portions of code can be completely regenerated quickly when scope and requirements change. (skip debugging - just regenerate the whole thing with updated criteria)
- GenAI can write thorough functional and behavioral unit tests. This is no longer a weakness.
- You must suffer the questions and approvals. At no time can you let agents run for extended periods of time on progressive sets of work. You must watch what is generated. One thing that concerns me about the new 1mm context on Claude Code is many will double down on agent freedom. You can’t. You must watch the results and examine functionality regularly.
- No one should care about actual code ever again. It’s ephemeral. The role of software engineering is now molding features and requirements into functional results. Choosing Rust, C#, Java, or Typescript might matter depending on the domain, but then you stop caring and focus on measuring success.
My experience is rolled up in https://devarch.ai/ and I know I get productive and testable results using it everyday on multiple projects.
No one should care about actual code ever again. It’s ephemeral.
Caveat: it still works best in a codebase that is already good. So while any one line of code is ephemeral, how is the overall codebase trending? Towards a bramble, or towards a bonsai?
If the software is small and not mission critical, it doesn’t matter if it becomes a bramble, but not all software is like that.
I think it works great in codebases that are good, but I think it will degrade the quality of the codebase compared to what it was before.
A good codebase depends on the business context, but in my case its an agile one that can react to discovered business cases. I’ve written great typed helpers that practically allow me to have typed mongo operators for most cases. It makes all operations really smooth. AI keeps finding cretaive ways of avoiding my implementations and over time there are more edge cases, thin wrappers, lint ignore comments and other funny exceptions. Whilst I’m losing the guarantees I built...
> No one should care about actual code ever again. It’s ephemeral.
> very large portions of code can be completely regenerated quickly when scope and requirements change.
This is complete and utter nonsense coming from someone who isn't actually sticking around maintaining a product long enough in this manner to see the end result of this.
All of this advice sounds like it comes from experience instead of theoretical underpinning or reasoning from first principles. But this type of coding is barely a year old, so there's no way you could have enough experience to make these proclamations.
Based on what I can talk about from decades of experience and study:
No natural language specification or test suite is complete enough to allow you to regenerate very large swaths of code without changing thousands of observable behaviors that will be surfaced to users as churn, jank, and broken workflows. The code is the spec. Any spec detailed enough to allow 2 different teams (or 2 different models or prompts) to produce semantically equivalent output is going to be functionally equivalent to code. We as an industry have learned this lesson multiple times.
I'd bet $1,000 that there is no non-trivial commercial software in existence where you could randomly change 5% of the implementation while still keeping to the spec and it wouldn't result in a flood of bug reports.
The advantage of prompting in a natural language is that the AI fills in the gaps for you. It does this by making thousands of small decisions when implementing your prompt. That's fine for one offs, and it's fine if you take the time to understand what those decisions are. You can't just let the LLM change all of those decision on a whim, which is the natural result of generating large swaths of code, ignoring it, and pretending it's ephemeral.
> - No one should care about actual code ever again. It’s ephemeral. The role of software engineering is now molding features and requirements into functional results. Choosing Rust, C#, Java, or Typescript might matter depending on the domain, but then you stop caring and focus on measuring success.
I think this has always been the case. "Bad programmers worry about the code. Good programmers worry about data structures and their relationships." Perhaps you mean that they shouldn't worry about structures & relationships either but I think that is a fools errand. Although to be fair neither of those need to be codified in the code itself, but ignore those at your own peril...
Data structures are still conversational items. I come from the DDD community and adamantly push back on data first architectures. Modules or Bounded Contexts reveal their relationships and data over time.
Perhaps you can think of the modules or "Bounded Contexts" as a type of data structure and the relationships between them. Idk. I don't have a particularly great view of DDD fwiw.
The post is about using LOC as a metric when making any sort of point about AI. Nowhere do I suggest someone shouldn't use it, nor that they should expect negative results if they opt to.
No one I’ve ever worked with in 40 years has ever seriously used loc as a measurement of progress or success. I honestly don’t know where this comes from.
That’s the odd psychosis here. Everyone knew loc was a terrible measure. But perhaps the instinctual pull was always there, and now that you can generate halfway coherent tens of thousands of loc in hours, our sensibilities are overwhelmed.
Yes, but it comes up in conversations of LLMs a lot. Thus, the rant in question. I think we are in agreement, or at least we lack disagreement, because that is the only stance I endeavored to take in the post.
I’m the Founder. Been working on this for a while. Finally was able to get the iOS app designed and implemented.
TestFlight link is on the site.
Mach9 is the Karate Kid of poker tournament training. You iterate through math, strategy, language, and run Monte Carlo tournament simulations against your own strategy (set of rules). Each test is against 449 AI bots with varying skills and personalities. Tests run instantly. Review your hands, get advice, see gaps, add rules, test again.
Over time you’ll learn poker tournament play just like learning Karate.
Satya Nadal will go down in history as the guy that killed Microsoft. The insane push to AI and copilot jammed in every app plus ads has done exactly what the OP states…
I will recommend that $599 MacBook every time now and power users invest in a MacBook Pro.
I was a loyal Windows user and now my own Surface Laptop 5 sits dark while I work on a Mac-Mini that was meant to be a side app dev machine.
Read the article. Satya brought the share price from $35 to $400, that wont kill Microsoft.
I guess what you’re trying to say is that it will kill Windows. But that wont happen since enormous percentage of businesses run Windows ecosystem.
Lets face it, Windows is in maintenance mode, pointless for MS to invest heavily in it since there is no threat for businesses switching to Linux or something else. MS devs primary maintenance job these days should just be scrabling MS Office API every 6 months or so to break Wine and other Linux non-emulators. Wine devs in constand rearrange deck chairs mode, while Win32+Office devs just add a new parameter to an API interface in their 6 month cyclic undocumented API breaking scheme.
You need a better Office than MS Office to break the cycle, and this will be a Web based office / collaboration tool. And guess where MS Azure and Web services fit in this brand new world.
Microsoft dominance aint going away in our lifetimes. Only non US government pressure may force other countries to switch to a flavour of Linux due to US sanctions. Only then can you see a visible migration from Windows. This is a decades long process.
He is definitely killing Windows though. Lack of quality of foundational products do not affect share prices until several years are gone. You will only see an effect in 2029-30, users slowly migrate away and license contracts are slowly dropped. I guess blame will fall on the next CEO and Nadella will be lionized as most successful CEO of Microsoft.
reply