Some years ago OpenAI released the first version of ChatGPT to the public. They made bold claims about their artificial intelligence and many people were impressed. For me, the product was quite handy and impressive from a technical perspective, but mostly a useful toy. Small coding tasks, text corrections or summaries of given information were most of the tasks I handed over. Whether it was ChatGPT, Google Gemini, Perplexity or Kagi assistant, the interaction with each of them was limited to an input field and attachments. More complex tasks I had to cut down into small chunks and then process them step by step. AI wasn’t able to get the big picture of a project.

Don’t get me wrong, even this was useful and made it possible for me to tackle projects, I would’ve needed far more time otherwise. But this approach on AI wasn’t intelligent at all. Just a big algorithm nobody was able to understand anymore, trained with most content of the world wide web, slicing a question in pieces and calculated word for word the mathematically most fitting response. The answer of an LLM is just the statistically most likely combination of words.

Considering, that correct facts and data are the majority of content in the WWW it works rather well, especially for commonplace stuff. But for unknown topics, current news and more complex contexts, this approach wasn’t only useless, but even dangerous. LLMs don’t know, if their response is correct. It’s just the best response the LLM can give and it will give it, always. In the meantime, there are ways to stop them giving an answer just to answer. The confidence threshold needed for an answer to be valid was adjusted higher.

But even then I wasn’t able to follow the hype around LLMs. I saw the potential, but I wasn’t as enthusiastic about the state of AI until mid 2025 as many people were. And even today, I assume around 95% of people only use the simple LLMs in the browser, celebrating the progress and thinking this is artificial intelligence.

Claude AI shows me the way

While these models and services were just a rock to hammer in nails for me - better than no tool - Claude AI has surprised me the last weeks. It’s not a nail gun, but a solid hammer. A hammer that I use shockingly regularly and that supports my personal and work tasks a lot. I’m not only more productive, I also can do stuff I couldn’t do before.

Complex coding projects? No problem. Claude isn’t limited to edit just one file, it could work through whole project folders. Multi-level analysis? Easy. File sorting and modification? Claude can do this.

To be honest, I didn’t expect how fast Claude Code and Claude desktop convinced me. The agentic approach, that allows for multi-level and multi-step actions, is my game-changer to make LLMs useful. This doesn’t make them real artificial intelligences, but really helpful tools. Like an untrained assistant that does the monkey work that you always should verify. But in the end it saves more time than needed for verifying.

Actual examples

As Product Owner I have access to our code and I have a basic understanding of our code. But the ERP code is complex, nested, divided and full of dependencies. To understand an “easy” function in a workflow I have to crawl for hours through the code. Or I just let Claude Code crawl through it, connected with our ticket system Jira and knowledge base Confluence. In just a few minutes, Claude checked the code, stories and documentation, then created a Confluence page with the answers to my questions. With basic IT knowledge the answers can be understood, but there are also enough details to go into discussion with a real developer.

People who know Confluence will also know, that the built-in search is no masterpiece. Even with Rovo, the newly implemented AI assistant by Atlassian, the search results couldn’t satisfy me. Claude, on the other hand, searches through Confluence, follows links to Jira and external sites and provides a great summary to topics. Not perfect, not without mistakes, but mostly solid and always reasonable and verifiable.

But also in private I use Claude Code a lot nowadays. I have implemented it as a sidebar in Obsidian. When you check my older blogposts you will find, that I had some troubles with nearly all notebook applications. My thinking is too fast and erratic to write more than some bullet points. So my notes were short, chaotic and therefore not helpful afterwards. But how would it be with a tool, that could take my chaos and convert it into structured, readable notes? Send “Rework the structure and wording of this note” to Claude and 30 seconds later I have a structured note that will be handy in future.

With this blogpost Claude has supported me too. Not for the writing itself, but for the structuring of the post. I write a note with the topics and ideas for a blogpost and let Claude bring structure to it. If you only knew how often I had to rewrite or rework blogposts in the past just because their structure was chaotic. Additionally I let Claude proofread my blogposts to improve spelling, syntax and forgotten words. But I typed this whole blogpost myself.

Hopes and worries

My experiences with Claude in the last weeks were an emotional rollercoaster. On one hand I’m extremely impressed by the capabilities and independence of it, especially with complex topics and tasks. On the other hand, this AI has reached a level that really could make some jobs redundant. Especially jobs that are based solely on data processing and presentation seem to be in danger. Consultation in the local electronics store may already be better done by AI than by a bored student. But also everything around research and summarisation. Why do I need a Controller who creates and updates reports into an Excel template? Let AI query and visualise the data itself. Always the newest numbers, always the numbers I need.

With creative jobs, I don’t see a real problem. Current AI isn’t creating anything and they don’t understand anything. A good developer will be supported by agentic AI and most likely more efficient, but current AI will have troubles creating sustainable complex software. Also my work as Product Owner won’t be taken away by AI, but it will take over the monkey work. At least for now.

But maybe we are soon at a point where brain work will be done fully by AI, just a small group of people will manage them and give them the tasks. What will be left over? Manual labour and fundamental science. For the first, I’m too left-handed, for the latter, too dumb. We’re living in thrilling times.

Hardware & quality

But let’s start with the specifications:

• Type: Handheld Gaming PC
• Display: 7 inch IPS 144Hz FullHD (1920x1080 pixels) display with VRR (Variable Refresh Rate)
• Processor: AMD Ryzen™ AI Z2 Extreme
• RAM: 24GB LPDDR5X (soldered)
• Hard disk: 1TB PCIe 4.0 SSD (M.2 2280, changeable)
• Bluetooth 5.2, WiFi 6E
• Operating system: Windows 11 with FSE (Full Screen Experience)

But technical specifications aren’t that meaningful in the end. The processor was developed specifically for use in gaming handhelds and can’t be compared easily with desktop or notebook CPUs. But it has more power than its predecessor, for sure. However, it’s more an evolution instead of a revolution. The 24 gigs of RAM seems to be plenty, and it is more than enough for FullHD, even when it has to be shared between CPU and GPU. The swappable solid state drive is nice, in my opinion. I haven’t replaced the stock one yet, but for some, 1 terabyte won’t be enough. But what got me curious the most beforehand was the modified Windows 11, which should be optimized for gaming handhelds by providing a new Full Screen Experience (FSE). Does it work?

The handling of the device is excellent for me. I’ve read complaints about the ABXY buttons being too loud. I find them crisp and think they will become a bit quieter or more worn in over time. All other buttons and sticks work flawlessly for me. However, there is one flaw in the build quality. On the right grip, the housing doesn’t fit perfectly; there is a minimal gap. Since you hold the grip firmly in your hand, the housing parts move and creak slightly. At this price, this shouldn’t happen. But it doesn’t affect either the functionality or the ergonomics.

One last point though: The device is perfectly balanced and doesn’t tip in any direction, but the display is angled. When you place the device on its back, the display is tilted slightly backward. When I had the device in my hands, this was actually quite nice. But it still confuses me to this day, since you’re used to a flat display from the Nintendo Switch and other portable consoles. I just need to get used to this.

Software: Windows 11 with FSE

The FSE is the first iteration from Microsoft to provide a user interface tailored for gaming handhelds. Instead of booting into the usual Windows 11 desktop, the ASUS ROG Xbox Ally X loads the FSE optimised for control with touch and controller input. Additionally, the lack of the desktop allows Microsoft to get rid of some background processes and with this free up resources for gaming.
When starting the device, the FSE welcomes the user with a start page containing a lot of tiles showing the last played games at the top, but then mainly adverts to buy games in the Xbox Store or subscribe to the Xbox Game Pass. No surprise considering the name of the device, but for me the page is absolutely useless as I buy and own all my games on Steam. I also don’t plan to change this in the future, as well as I don’t plan to subscribe to the Game Pass. What puzzled me was that I couldn’t switch to any other page like the library as my main page. Every time I pressed the Xbox button or closed a game, I got presented this for me useless homepage.
With the menu on the left I can access the library. This shows all installed games, even from other sources like Steam and other launchers, at least when the launchers were installed over FSE. I’ve not tested installing other launchers that can’t be installed in FSE. But there is a big drawback, at least with FSE and Steam. FSE shows the Desktop shortcut icon on the games tile instead of a cover. This looks really ugly and you have to install the addon SteamGridDB to change the look of the tiles.

Another annoyance that shows that Microsoft still has to go a long way are the update mechanisms. The plural is by choice, as there are multiple places where the user must do updates. In the FSE, ASUS ArmoryCrate, the Windows Store and Windows Update itself. For other games, you also have to monitor their launchers. This is a mess and needlessly complicated, especially as there are no consistent notifications about new updates. Here Microsoft and ASUS must find a way to improve this. Nobody likes to regularly check four or more places for new updates.

A great feature, on the other hand, is the game bar. It opens after a simple button press as a layer over the game and provides access to the most important settings. But also other tools are available there, like screenshot, screen capture or AI support for games. The AI support wasn’t that great in my tests. Its tips were rather generic. Even with access to the game, there were no situational hints or help.

I got angry over the standby. Usually the device should go into sleep mode with a short press on the power button. Another press then should wake it up again and you can continue where you left off. The “should” in the sentence before is intentional. Most of the time it works, but sometimes it doesn’t. Then the display goes off, but the device won’t switch into standby and continues to drain power. In other cases it won’t wake up again and you have to press the button multiple times to wake it up, or you have to force restart it as multiple presses don’t help either. So you always have to save before putting it in standby.

On the other hand, the biggest advantage of Windows is that you don’t have to worry whether your game runs. All games run on Windows.
In total I would say that the system isn’t finished. With games only on Steam, it isn’t a satisfying experience. With some tinkering it is possible to switch to Steam Big Picture instead of Microsoft’s FSE, but somehow the performance of the UI wasn’t as smooth as I expected.

Software: SteamOS as an Alternative

As I use Linux on my private PC and macOS at work, I have no dependency on Windows. So I decided to give a Linux-based alternative a try. My first thought was Bazzite, which has an official port for the Xbox Ally X. But then I read the news that SteamOS 3.8 Beta also added support for the device, and I gave it a try. Installation was without any problems and I could even update to version 3.9 Beta.

The standby with SteamOS is completely different, it just works like I expect it to. Press the power button and the device goes into standby, press the button again and it wakes up in less than three seconds - exactly where you left off.
But it’s not a seamless experience. After installation you have to configure some stuff to get all functions and keys working, but when done, most stuff is running far better than on Windows. Just two exceptions at the moment: The RGB lights around the sticks can’t be configured and you can’t update the BIOS from SteamOS. The RGB lights are no drawback for me, I deactivated them after unboxing and do not intend to activate these annoying lights. And for the updates I had to create a Windows-to-go USB stick that I boot into once every month to do the updates. Not perfect and rather slow, but better than wasting space for a dual boot Windows.
What will likely never be available are the AI functions like AI upscaling utilising the built-in NPU. These features are in development for Windows, but I don’t expect them to be ported to Linux.

But overall I prefer the Linux experience over the Windows experience.

Gaming experience & performance

The gaming experience itself is mostly the same on both systems. The performance depends on the game and on the system. It’s no surprise that the device can’t run the newest AAA games with maximum details in 1080p with 144 FPS. It’s often necessary to adjust the game’s settings and try out what works best. Maybe the performance is better when lowering details or even the resolution to 900p or lower. Cyberpunk 2077 for example I play in 1080p with Steam Deck preset and reach somewhere between 35-45 FPS reliably. At first, this was confusing for me as a PC gamer. Gaming below 60 FPS is possible? I assume thanks to VRR it feels surprisingly fluid and smooth.

While my PC’s old CPU struggles to push out more than 40 FPS in Dune: Awakening, the Xbox Ally X is able to do 50-60 FPS, with FSR3 frame generation even 70-80 FPS. On lowest settings of course. Tony Hawk 1+2 on the other hand achieves 60 FPS plus with high settings, as well as Need for Speed Unbound. I even have the choice to lower the settings a bit and reduce the power limit of the console. This increases the battery life by a good amount. An extreme example is Hades that I can run in 1080p with maximum settings at 7 Watts. When lowering the settings, even 4 Watts is possible, extending the battery life to over 7 hours.

With Cyberpunk 2077 I have to crank up the power limit to 25 Watt to get stable 30 FPS plus. Going above up to 35 Watt showed no big increase in maximum FPS for me, just the 1% lows were better.
And with some other games, it’s clearly visible that they weren’t developed for a small screen. In The Witcher 3 the text is really small and blurry. I can read it, but it’s not the best experience.

Conclusion

The ASUS ROG Xbox Ally X is a gaming device to its core, and it shows. But it does this great. The ergonomics are superb and superior to the Nintendo Switch. The performance is impressive for the size and I’m totally happy with the screen. Another big plus for me is the flexibility with the operating system, so I could install and use Linux.

Competitive shooters aren’t the games I would play or recommend on an Xbox Ally X. For singleplayer games, on the other hand, it’s a great device and my chance to work off my Pile of Shame. It’s fast to stow, fast to get out and anytime ready. No searching for an external controller, no searching for a place to put my notebook and mouse on, no searching for internet to stream the games. Just power on and play.

With a price tag of 899 Euros the ASUS ROG Xbox Ally X certainly is a luxury investment. For the same money I could buy a stronger gaming laptop or four years of GeForce Now. But these options won’t be as easy and comfortable as the XAX. For me, the investment was worth it.

At the end of 2022, I started this blog after freezing my old one. Back then I decided to use Ghost as a CMS.
I also wrote a blog post explaining my decision: Now It’s Kernic.

Three years later, my enthusiasm for Ghost CMS has mostly vanished. I still think it is a good CMS, but its focus doesn’t fit my needs. The newsletter feature sounds nice, but my type of blog isn’t made for newsletters. Also, the technical foundation may be modern, and even with Ghost being open source, it is aimed at providing hosted Ghost(Pro) services. For self-hosting, the system is unnecessarily complex.
Until now I haven’t been able to get the scheduling function working, and ActivityPub never worked.

So when I rented a new server on Black Friday with the goal of modernizing the systems, it was the perfect time to search for an alternative. The first alternative that came to mind was WordPress. With the change, I hoped to get two new features: ActivityPub and multiple languages. Both were possible with WordPress, and so I migrated my blog and pushed it live… for two days. Then I was so annoyed that I reactivated Ghost again and searched for a new plan.

You’re now seeing the result. There are a lot of changes behind the scenes, but also for users.

The most obvious one is the design. It’s simpler, no fancy features. Pure HTML, CSS, and a little JavaScript. But no external frameworks, fonts, or other resources. If there is no embedded content, your data isn’t sent anywhere. This is achieved with several other changes. The newsletter is gone. I had four subscribers and below a 50% open rate. Three comments in three years didn’t justify a comment system, so it’s gone too. This site is now prebuilt with Hugo, and the server just serves static files to you. Without any interactive elements, there is no need to store any information about you or in your browser. Cookies gone.

Some changes you might not notice immediately were implemented too. The main change certainly is that my blog is now multilingual. The main language is English, but every page and post is available in German too. On one hand, I increase my audience with this, but on the other hand, I want to improve my English. All previous blog posts are now available in English, not translated by me but by Kagi Translate.
Additionally, I extended the machine-generated audio versions of my blog posts. With Ghost I had to generate the audio version by hand and upload it manually to the post. As I automated the publishing of blog posts now with Hugo and some Python scripts, I also added the automatic generation of audio versions for the posts-in both languages. And when it was already automated, I had audio versions made for all previously published blog posts too. But what is a blog with audio? Right, a podcast. So now there is also a podcast feed in addition to the normal RSS feed. Feel free to listen to my blog posts, even inside your podcast app, instead of reading them.

That was it for now. Any feedback? Hit me on Mastodon.

It was one of those evenings when I didn’t fall asleep within 10 seconds. Instead, a question shot into my head:

How does the Stargate address system actually work?

I remembered that a Stargate address, at least within the Milky Way, consists of seven symbols. Six symbols for the destination, one for the point of origin. There are 36 symbols available.

Wait? If there are 36 symbols and one of them must be for the origin, how can there be more than 36 Stargates?

False Assumption

Without a solution to the problem, I fell asleep. The next morning, however, the problem was still running through my head and I started researching on the internet.

I quickly realized my mistake. I had assumed that the Stargates in the Milky Way are all identical. But apparently only 35 symbols are identical and the 36th symbol differs from gate to gate. Because it stands for the planet.

This post is intentionally kept short. It captures a thought of mine without diving deep into the topic.
What do you think of this type of post?

Motivated by the political situation in the USA, I reviewed my IT landscape and discovered more services from there than I’d like. I made a quick decision to use services from the EU when possible. Here, not everything is perfect either, but it’s significantly better than on the other side of the Atlantic.

But how do you find alternatives from Europe? It’s not that difficult, as several people have had the same thought and so Constantin Graf has set up a website showing European alternatives:
https://european-alternatives.eu/

Taking the First Step

Even though there are alternatives for almost everything, not everything can be migrated easily and quickly. But with music streaming, the providers are largely interchangeable. So far, I’ve been using YouTube Music. This had two advantages for me:

• YouTube Music provides access to all audio content from YouTube. This means you also have access to many remixes and mashups that aren’t available from other providers.
• With YouTube Premium, not only is YouTube Music included, but YouTube also becomes ad-free.

Nevertheless, YouTube is owned by Alphabet (Google), part of one of the world’s largest tech giants - headquartered in the USA. Additionally, the company has ingratiated itself with Donald Trump, putting profit over morals¹.

Selecting an Alternative

There are currently four alternatives: Spotify, Qobuz, Deezer, and Soundcloud. The title of this post already reveals my choice: Deezer. But why?

Soundcloud isn’t a music streaming service in the way I’m looking for. There’s mainly user-uploaded music, major music labels aren’t represented there. So many well-known artists I still want to listen to are missing.

Spotify is certainly the best-known music streaming service on the market and the pioneer. But Spotify is backed by many investors from the USA, the service still isn’t financially sustainable², and doesn’t pay artists well.

That leaves Qobuz and Deezer. Both services are from France. Qobuz focuses more on high-res music and has a smaller library, Deezer is more oriented toward Spotify but, unlike the market leader, also offers music in higher quality.

In the end, I chose Deezer, mainly because of the larger library and lower price.

My Impression

Basically, music streaming services work quite similarly. There are personal and public playlists, you can search for artists, and there are recommendations. The Deezer app on iPhone works flawlessly, and the Windows application doesn’t work at all. When I try to move the window, the application crashes silently. But I simply use the website, which works without problems.

The quiz function is entertaining for in-between. You can guess music titles with friends and see who knows more. Certainly not a main feature, but implemented amusingly. Deezer also offers audiobooks and podcasts. The former aren’t as comfortable to listen to since there are no bookmarks. And I haven’t been able to get excited about podcasts so far.

All in all, Deezer is solid. The music library leaves little to be desired, apart from the remixes. The apps are designed for music and do that well. Much is also available in high-res, and lyrics are integrated into the applications.

Sources

¹ https://time.com/7218225/google-maps-gulf-of-america-search-impeach-trump-calendar-holidays/
² https://www.techloy.com/despite-adding-over-8-million-premium-subscribers-spotify-ended-q2-2025-at-a-loss/

What was extremely innovative with pure gesture control on the Jolla smartphone with Sailfish OS or webOS is now standard. Sometimes still not implemented as well as in Sailfish OS, but intuitive enough. And if we’re honest, the differences in look and feel between Android and iOS aren’t that big anymore. iOS still has the edge for me, especially in terms of consistency, but the concepts are very similar.

And that’s where the problem lies for me. Home screen with apps and widgets, a quick launch bar at the bottom, a status bar at the top. Pull down quick settings on the right, notifications on the left. On Android, maybe an app overview, that’s it. But even after 16 years of smartphones, I don’t feel like this concept suits me.

A few examples? I hardly ever use the app icons. I tap on the search, type in 1-3 letters, and select the app. At most, I open a few frequently used apps via the icon on the home screen. But that’s only because I have notifications there. These are my second annoyance. Why are they hidden? I’d prefer them as a list on the home screen itself.

Why? I manage my notifications very precisely. Which apps are even allowed to send me notifications? Which notifications should they send me? Sometimes even when. And then they’re just hidden behind a swipe. But notifications are the reason I pull my smartphone out of my pocket. I want that exact information, that’s why it’s active. And then interact with it directly. Well, hidden… And unfortunately for me: out of sight, out of mind.

I’m not entirely sure, but I think with Windows Mobile, you had notifications on the main screen. Well, there were hardly any apps that sent notifications back then. But emails, appointments, and messages were definitely visible right there. I miss that.

At first, I didn’t have time for troubleshooting. Although the reconnection happened more than ten times a day, it was so fast that we usually didn’t notice it. But someone who did notice was Deutsche Telekom. Three days after the problems started, I received a text message. Deutsche Telekom had noticed that we were having a problem with our connection. They wanted to help us and said we should call the hotline.

To be honest, I was pleasantly surprised: Proactive communication with a customer who’s having problems—where do you even find that anymore? However, the timing was inconvenient: we were leaving for vacation three days later, and I didn’t want to head to Italy with a half-finished troubleshooting session. So I ignored the text and spent the next three weeks lying on the beach.

The problem didn’t magically disappear

Three weeks later, I was curious. Maybe it was an external issue that had been resolved? No, unfortunately, it wasn’t. On the contrary, the problem had actually gotten worse and was now becoming annoying. I was getting kicked out of meetings while working from home, GeForce Now would get interrupted, VoIP calls would drop, and so on.

I had already mentally prepared myself for the call with Telekom. But first, I thought about what could be causing the problem. I quickly made the connection to moving the FRITZ!Box. The obvious assumption was that something went wrong during the move. First, I checked the router’s temperatures. Maybe it wasn’t getting enough air on top of the cabinet? No, that wasn’t it.

But the cable from the phone jack to the router was now hanging in the air, instead of lying relaxed behind the TV stand. The router was five meters away from the phone jack—which was the maximum length of the cable.

A shot in the dark, and a direct hit

The cable is old and came with the router. My guess was that the cable’s quality wasn’t particularly high, and accordingly, it didn’t react well to being stretched taut in the air.

So before I had a Telekom technician come over, I thought it was worth a try. I quickly ordered a longer cable on Amazon and made sure the quality was good. Instead of a cheap wire, I chose a cable made of pure copper. It was ten meters long, twice as long as before.

What can I say? The cable was delivered, I replaced the old one with the new one, and restarted the connection. There hasn’t been a single disconnect since. Sometimes, it’s just a cheap old cable. Even if they all look the same on the outside, a proper cable can make a significant difference. And it can save you a phone call with Deutsche Telekom.

GeForce Now Isn’t for On the Go After All

I travel a lot. Every week, I spend more than ten hours on the train. Plus one evening at my parents’ place, away from my PC. Despite all that, I still want to be able to game when I’m away from home.
A few months ago, I considered getting a gaming laptop but decided against it. Instead, I opted for Nvidia’s game streaming service, GeForce Now. It actually sounds like a fair deal. For a little over 20 euros a month, you can play many of your purchased games from anywhere in high quality—streaming makes it possible.

The “actually” has less to do with Nvidia and more to do with the internet itself. On the train, you have two options for getting online: via the mobile network or via Deutsche Bahn’s Wi-Fi. The mobile network is significantly faster, but also much more volatile. I find it strange that there’s practically no mobile reception, especially when entering and stopping at the Mannheim and Karlsruhe train stations.
The Wi-Fi is more reliable and rarely has a complete dead spot. But it’s nowhere near fast enough for GeForce Now.

But the internet isn’t reliable in other places either. I’m writing this post right now from a campsite in Southern Italy using my smartphone’s hotspot. There’s supposedly Wi-Fi here, but I haven’t managed to connect to it in a week. And the two or three out of four bars of 5G are probably just marketing. Top speed: 7 Mbit/s with a ping of just under 95 ms. Often worse. That makes GeForce Now little to no fun at all.

Local Sunshine Instead of Distant Clouds

The cloud experiment has failed, at least on the go. I also have to be honest, gaming on a MacBook, even with a mouse, just isn’t that great. I could, which I haven’t tried yet, bring a controller along. But then I still have to set up the laptop and connect to the internet. Plus, my games just don’t work on macOS.

So the alternative is to have a device that can run games locally. I don’t want to lug around another laptop, and it’s not optimized for gaming anyway. So, a gaming handheld. The choices are devices with Windows and devices with SteamOS. The latter is based on Linux and is supposedly more optimized. Windows, on the other hand, is more flexible. I don’t have a personal preference, especially since SteamOS or Bazzite can now be installed on many Windows devices.

Finding a device wasn’t actually that hard, or so I thought at first. Valve, MSI, Lenovo, and ASUS are the major players in the market. The MSI Claw and Lenovo Legion are too big for me, the Steam Deck and Legion Go are too weak, which left only the ASUS ROG Ally X, and I was on the verge of buying it. Then it occurred to me that all these devices are almost a year old, and I started looking for announcements of new generations.

Microsoft and ASUS Are Collaborating

Not much new on the market. MSI just released a new Claw with an Intel APU, and Lenovo announced a new one with a new AMD APU. But both are too big again. Valve is waiting and saying nothing at all. That leaves ASUS. And lo and behold: a new gaming handheld has been announced, with the new AMD AI Z2 Extreme. Not a revolution, but still 20–30% more performance for the same power consumption. It would be a shame not to wait for that, right?

I’m not sure, but I’m going to wait anyway. Why unsure? The new handheld from ASUS will be called the ASUS ROG Xbox Ally X and is the result of a collaboration with Microsoft. Xbox-style grips are supposed to make it more ergonomic, but also thicker. It will also get an Xbox interface, and the Windows on the device is optimized. That actually sounds good. Finally, an interface that fits a handheld, and more power for games. But how will it work with Steam? I haven’t used the Xbox Store at all so far and don’t plan to. I’d also be interested to know if I could install SteamOS or Bazzite on it if I wanted to.

Then there’s the price: €899.00 is announced or rumored. The price isn’t quite certain yet. The ASUS ROG Ally X is currently available online for €729.00, which is 19% cheaper, and its price will surely drop again with the release of the new device. So the extra performance is directly reflected in the price. At least one-to-one. But as I said, I’ll wait for the release, watch some reviews, and then I’ll decide. But it will almost certainly be a device from ASUS.

In any case, I’m looking forward to being able to game on the go, even when the internet isn’t so fast or stable.

My car is old and has been driven a lot. For me, a car has always been one thing above all: a utilitarian object for getting from A to B. So, I don’t wash it every Saturday, and while a dent is annoying, it’s just a side effect of using it. What’s important is that the car is practical, reliable, low-maintenance, and durable.

For 14 years, my BMW 118d has largely met these criteria. It offers a bit less storage space than I would have liked, and it lacked a trailer hitch (and a €1200 retrofit was too expensive for us). But that’s about it for the criticism. For almost 14 years, the vehicle has carried us over 450,000 kilometers around the world, mostly without any problems. Wear and tear parts, yes, but usually later than expected. Last year, the first major repair came due, costing nearly €1,000, which made us question if it was still worth it. But the defect wasn’t unusual, and the car was otherwise in good condition, at least technically.

Something with the Steering, or Whatever

Until four months ago, when the power steering suddenly failed. First sporadically, then permanently. Always with a smug “ding” and a generic steering wheel symbol on the display. Our trusted workshop took a look at the car and was somewhat at a loss. Yes, it’s the power steering. But they couldn’t tell if it was the servo motor or a control unit. Unfortunately, the diagnostic interface only showed a generic error message: problems with the power steering. They went to the trouble of checking all the contacts; they were fine. But is it the servo motor, the control unit for the servo motor, or one of the other four control units? The car simply wouldn’t give up that information.

But regardless of whether it’s the servo motor or the control unit, a repair including the spare part would cost at least €1,000, and that’s if they replace the right part on the first try. Otherwise, double that. A heap of money for a vehicle with 450,000 km and 14 years under its belt.

Why So Vague?

Fortunately, we can temporarily borrow my mother’s car. Also a 1-series BMW of a similar age, but with “only” 50,000 km and it’s a convertible. We’re just realizing on our camping trip that a convertible is not for us. And surprise: on the highway near Naples, it goes “ding” and a steering wheel symbol appears. The error description? “Something with the steering.” And my mother’s car even has a proper system with a display. Can’t it be more specific? Especially since the steering was working completely normally. Okay, the error was gone after restarting the car, but still:

Why so vague?

My CPU can tell me the voltage, the current draw, the temperature, and who knows what else for every single core. If there’s an error, I can see exactly where the deviation is. And the cars? There’s an error somewhere in the steering and all related systems. But I’m not telling you any more than that. Why not? Shouldn’t it be quite simple to know which part is causing the error? The error is being generated, after all! And couldn’t they just install a few more 2-cent sensors to detect problems earlier and report them more clearly? Servo motor needs more power to turn because it’s not well-lubricated? You could surely read and display that if you wanted to.

But I suspect two reasons speak against this for the car manufacturers: first, it would save people trips to the workshop, and second, it would make it clear to everyone how the vehicle is built - and how well.

Well, there’s nothing you can do. A new vehicle with even more technology and even less transparency. The lobby is doing its job well.

I define the cloud as services where you don’t have fixed resources and systems, but rather the service is offered completely independently of any hardware. Dropbox was one of the first major cloud service providers, but so were Google with its office suite and Amazon AWS with dynamically scaling resources. The actual system is managed by the provider, and you can use exactly the service you’ve subscribed to, whether it’s storage space, office software, or computing power.

In contrast, there are self-hosted solutions. You can’t quickly add resources, and the management of the entire system lies with the user. The technical expertise required here must be significantly greater, but the costs must also be considered. What performance do I need? Do I need it continuously, and if not, can I smooth out peaks and use weaker hardware instead? What software do I even use, and how do I secure it and the underlying system?

Advantages of the Cloud

You can see pretty quickly that for the end user, cloud solutions are very convenient. Download Dropbox or OneDrive, install it, log in, and select your folders. Just like that, your files are synchronized to the cloud and back again. Open Google Docs, log in, and start writing. It saves automatically, and nothing needs to be installed.

But even for professional users, it’s tempting. Need to quickly set up a small homepage? Open AWS, book web resources, and the web server is ready to go in minutes. If the site becomes more popular than expected, you just add more power—if that doesn’t happen automatically up to a set limit.

At work, I also see the advantages regularly, and I have a current example. Our OCR solution, which reads text from scanned documents, works well, but not as well as we’d like. Thanks to the cloud, we now rent GPU power from a provider with absolute flexibility and can test whether GPU and AI can improve recognition. If it gets better, but more power would be useful? Just book more power. If there’s no improvement? Then we just turn it off again.

Disadvantages of the Cloud

For Providers

So, cloud services can be simple and flexible. But at the end of the day, there are still physical servers behind the cloud. The real difference is the software on them. This software has to be written and maintained. The servers have to be purchased and operated. However, the risk of utilization now lies with the provider. Since clouds are supposed to be flexible, a certain amount of overcapacity must be available to react to customer requests at short notice. If there’s no demand, the provider pays for the capacity without generating revenue from it.

This is one of the reasons why mainly large and broadly positioned companies offer cloud services. Amazon itself needs a lot of power but is surely flexible in its internal allocation. They use unused capacity for their own purposes until someone needs it and pays for it. When capacity becomes free, it’s used for internal calculations. Small companies don’t have enough flexible tasks for such an approach.

For Business Customers

In the previous paragraph, you can see that providers take on a certain risk. But they don’t want to bear this risk alone and price it in accordingly. The management software and interfaces also have to be created and maintained. This is an additional expense that also has to be paid for. If you know your resource needs and software environment precisely, you don’t need this flexibility and will probably be better off with a classic solution.

In recent years, this topic has also been seen in very black-and-white terms. Cloud-only or on-premise-only were the options. Meanwhile, many have realized that a hybrid operation is also possible. “Cloud-first” would be something like that. New projects are first implemented and evaluated in the cloud. If they work and their demand can be clearly determined, they can be moved to their own hardware. Or they remain in the cloud if you want to stay flexible. You decide on a case-by-case basis.

For Private Customers

For private customers, the situation looks a bit different. Here, it’s less about pure performance and more about services. Providers make the cloud extremely appealing through its ease of use. When you start up a new PC, you’re immediately offered to back up your data to OneDrive. Backup your iPhone? Just use iCloud! Notes? Always synced with Notion, Evernote, and the like, and therefore always with you. Always with you and always with the provider. Many claim to encrypt the data, but you can hardly verify that. You trust that the provider will keep the data private while ensuring its security and availability.

Additionally, you are dependent on the provider. About 10 years ago, Evernote was the note-taking app on the market. Fast, lean, and with great features, thanks in part to native applications for every environment. However, maintaining these and keeping them at the same functional level was extremely difficult and costly for Evernote. So the company decided to build a universal web app instead, alienating many customers. The application was suddenly slow on all platforms, had fewer features, and plenty of bugs. The only problem was that migrating thousands of notes wasn’t so easy.

Wunderlist was shut down and became Microsoft To-Do, Amazon Photos discontinued its service, Google killed its RSS reader and many other services, Sonos is releasing new software with fewer features, the cloud for a security camera is being shut down, making the device unusable, and so on and so forth. The list of discontinued or significantly changed services is long, and the options for users are limited. Usually, it’s just accept it or switch.

Conclusion

Would it be convenient if my personal files were automatically synced to OneDrive? Pictures from my iPhone go directly to iCloud? And my notes are available everywhere? Of course, it is. But do you still have control over your data and services? Can you rely on these services? Unfortunately, no.

I myself use very few cloud services. For communication, I use Discord, iMessage, and WhatsApp. For Obsidian, I’m currently using the official sync service. However, the data is still stored locally and is fully encrypted. The plugins were all a bit unstable. That’s about it.

My data is on a local NAS with a RAID1 setup. Currently synchronized, so not as a backup. Some services run on my old laptop, while important services run on a rented server. It offers much more power than I need, but it runs 24/7 and has redundant connectivity. I still need a solution for backups. Hetzner seems to have good offers for storage here. Classic storage, not a “cloud.” There, I could back up my data incrementally and encrypted.

But you can see, clouds are simpler than self-managed solutions. Nevertheless, whenever possible, I don’t rely on the cloud but on a solution that is under my control.