On November 7, 2024, my debut album as Kometenschmidt was released. That was like half a year ago, and I’m only now writing about it. Funny enough, I’ve had plenty of spare time recently, but perhaps I needed to let it all sink in.

When I was younger, I claimed to dislike hip-hop. Truthfully, I’ve liked the music since watching Beat Street. But growing up in the 1990s, my tastes gravitated toward techno and electronic music. Growing up in a small village and going to school in a quiet town in Germany, I had no connection to the ‘thug life’ or the crime-filled stories told in rap songs. My classmates didn’t either, but they still carried themselves with a ‘too-cool-for-everything’ attitude, mimicking the hustler persona. It felt out of place to me, so I never fully connected with hip-hop culture back then. Hip-hop became background noise—always there, but never something I really engaged with.

That changed when I started university in the late ’90s. I moved into a student residence and into the new age of the internet. My desk was crammed with two Technics 1210 turntables, a Numark DJ mixer, speakers, and the bulky CRT monitor of my computer. There was no space for books or study materials, but I didn’t mind. My dorm admin installed a network card in my computer, and for just 25 Deutsche Marks a year, I had constant internet access at a blazing 1Mbit/s. Napster, the DJs at our student club, and my music-obsessed friends opened doors to decades of music I’d never fully explored—hip-hop included.

Over the years, I experimented with making all kinds of music, losing my focus on electronic music but discovering new worlds. While some of my friends found success by sticking to one style, I wandered. I began creating hip-hop beats and instrumentals, though I always felt they lacked something: great rapped lyrics.

Fast-forward to my 40s, when I finally released my first hip-hop EP. Without a rapping counterpart, I focused on instrumental hip-hop, still longing to make “the real stuff.” That’s when I stumbled upon Looperman, a site offering free acapellas and samples. It was fun to experiment, and though the quality of most samples was mediocre, one connection stood out. A rapper whose acapellas I’d used reached out to collaborate. We planned to make a full album together.

But as John Lennon famously said, “Life is what happens to you while you’re busy making other plans.” Life pulled us in different directions, and our album stalled. I was frustrated and eager to finish, but John (the rapper, not Lennon) was busy. Then, earlier this year, a friend introduced me to AI tools for music production. I was blown away. Suddenly, I had access to unlimited, high-quality samples without worrying about copyright issues. It felt like a dream come true.

This new creative boost helped me complete the album by May. I wanted to release it immediately, but I waited for John’s feedback on his contributions—he’d written and rapped lyrics for half the tracks. His response was bittersweet: he liked the album but wanted to re-record some parts. So, I waited again. Meanwhile, I fine-tuned the details and polished the mastering. By September, it was finally ready to upload, and on November 7, it went live across all audio platforms.

The journey to this album was long, but every step felt worthwhile. While I’m taking a short break now, I’m not stopping here.

The album is called Kalsarikännit, a Finnish word that means “getting drunk at home alone in your underwear with no intention of leaving the house.” The name came to me during the COVID-19 pandemic, but given how the world is heading, it feels like a solid plan for 2025.

I’m deeply grateful to John and all the other artists who contributed to this album. Without AI, this project would have turned out very differently. AI tools helped me generate samples, write and rap lyrics (like the track Love Is Stronger Than Hate), and mix and master the songs.

After 30 years of making music, I haven’t become a hip-hopper. But I’ve stayed curious—and that’s what keeps me going.

Mrs. Davis was an Amazon suggestion for me. Usually, I ignore these, but the synopsis intrigued me: a nun fighting an AI. Having grown up in a Christian household—and as a fan of Warrior Nun (and still active church member)—I like those fantasy stories derived from the stories told in the Bible. I wouldn’t say I am deeply religious, but rather a hopeless romantic who believes there must be something like karma (I know, not a Christian concept, so let’s call it God…) or life after death. Yet, the science enthusiast in me knows there’s no evidence for such things.

However, I decided to give the show a try, and after the first 20 minutes, I was left wondering, “Watzefack am I watching?” Even after the first episode, I still had no clue where it was going. But the ending compelled me to watch the next episode. It wasn’t until at least episode 4 that the storylines began to untangle, and the bigger picture became more and more visible.

Ben Travers of IndieWire wrote, “Mrs. Davis is made to entertain, first and foremost, and it’s a highly enjoyable experience whether you engage with its deeper issues or not. Betty Gilpin is delightful. The larger ensemble matches her irresistible energy. The visuals are colorful and vast. The scripts are tight, the momentum unwavering, and the sheer joy palpable.”

Rolling Stone‘s Alan Sepinwall said, “Mrs. Davis featured some of the crazier plot twists you will ever see, yet managed to be genuinely moving and thoughtful even in the middle of all its ridiculousness.”

I couldn’t agree more with these reviews. The way Mrs. Davis is told proves that creative storytelling can elevate a story, and the idea behind Mrs. Davis isn’t bad at all.

I had never heard of the show before Amazon suggested it, but it truly surprised me. I would recommend it any time, even if you’re not particularly interested in Christian history.

Occasionally, the browser market introduces interesting new concepts. Recently, a colleague introduced me to the Arc browser. At first, I didn’t see any features that made Arc superior to Safari. I was initially put off by the interface, with everything—tabs, address bar, etc.—positioned in a sidebar instead of at the top. However, I quickly forgot about the browser after the first day. A few days later, I read something about a new überbrowser on my iPhone and decided to give it another try. The app icon of Arc looked familiar, and I was instantly intrigued by its AI-supported features (Arc Max) and the presentation of search results.

With Arc, you type your query into the search bar and click on “Browse for me.” Arc then searches the internet and creates a live website with a summary of your search topic. This approach allows you to find what you’re looking for within seconds, without wading through ad-overkilled websites with SEO text and clickbait headlines. And if you browse the usual “Google” way, a mouse-over on the links will give you a short page summary, the so-called 5 second preview. As a marketeer, I wonder if this does count as a click? However, as a user I found this rather comfortable.

The clean, distraction-free interface is also impressive—just the content and the controls you need. The tab organization is just next level and a killer feature for everyone who has always dozens of tabs open. AI and ChatGPT integration

Arc has many well-designed features, but I won’t go into detail here. You can see for yourself on their website: https://arc.net/.

The one big minus for me is the lack of native support for Apple’s Keychain. Fortunately, I use a third-party password manager, so this isn’t a major issue. Since Arc is based on Chrome, the extensions I typically use work just fine. Although I haven’t set Arc as my default browser yet, I’m very close to doing so. But I can try the integrated “try Arc as default for a week” feature.

Okay, irony off. Irony is a thing AI won’t get anyway. Meta Generative AI is coming soon, but who was waiting for it? I consider myself an AI heavy user, I’m curious to see what AI can do next, only to scare the shit out of myself. But imagine Instagram or Facebook with generative AI — this takes shit to a different level.

I was already concerned when Spotify announced it was training its AI with artists’ music because it seems like a logical step going from AI-generated algorithms to AI-generated music. This makes life even harder for original content creators, as it undermines the creative process and becomes another source for a flood of fake news. If you feel masochistic enough to read the comment sections on social media these days, you’ll get the feeling that this will find fertile ground. And the way Meta rolls out GenAI is so shady, I would rather not contribute to that.

It’s not like, “Hey, we wanna roll-out Generative AI; are you interested in training it with your content?” No, it’s hidden somewhere in the Meta-Verse. And you cannot just say no to your content being used for training AI; “you have to submit a bloody good reason for objecting for us to approve your request”, as Threads user Matt Navarra points out. You have to fill out a form and explain, why do you don’t want to make Meta the happiest place on virtual earth?

If you don’t know any reasons, here are mine:

• I prefer user-generated content because it’s a social network, not a socAIl network
• AI-generated content is a degradation of the human, creative creation process
• AI cannot recognize art and satire, which significantly increases the risk of fake news and manipulation, I don’t want to contribute to that
• AI-generated content is most likely to be favored by Meta and its algorithm, and is thus at the expense of user-generated content
• Meta is already abusing its market power to promote paid content and “recommendations” over the content of the people I follow and the content I really want to see

These were just a few points I included. You also have to provide your email address to get an opt-out mail with a verification code. According to Meta, the objection request is reviewed in accordance with relevant data protection laws. I guess that depends on the country you live in because you have to choose that from a drop-down menu, too. If your request is honored, it will be applied going forward. So far, so good. Sound like a lot of effort, doesn’t it? But think again: this would still be too easy. In the next paragraph, Meta acknowledges the following:

“We may still process information about you to develop and improve Al at Meta, even if you object or don’t use our Products and services. For example, this could happen if you or your information:

• Appear anywhere in an image shared on our Products or services by someone who uses them
• Are mentioned in posts or captions that someone else shares on our Products and service”

“ANYWHERE… on our products or services”… that is like a big eff you, we’ll use your data anyway. In Germany, it’s almost impossible to communicate without WhatsApp. I have like ten different and more secure messengers on my phone, but with whom am I communicating if no one is using it?

I really don’t like where this is going, and I think it’s time to leave Meta eventually. Unfortunately, I need that for work, and people are too lazy to switch to better messenger. Maybe it requires a hard cut here to educate the people around me?

While I’m still amazed by the possibilities of music AI (such as udio, Suno, Riffusion, etc.) these days, the new album by The Lemon Twigs is out, and I’ve never heard of them before.

The 12-track release envelops me in a strange plaid of melancholy, leaving me torn. On one hand, I’ve always loved it when old music influences young or new artists to create something both innovative and familiar (Adam Green, The Black Keys, The Checks, The Heavy, The Kills, etc.). On the other hand, AI is poised to soon take over, potentially creating music in similar veins, alongside the obvious genres like lo-fi and melodic techno house. This is why I feel doubly melancholic—the music itself and the uncertain times when an AI could produce an album like this.

This melancholy also encapsulates the feeling I had while listening to the album: I enjoyed the sound from the moment I pressed play until the last song, but none of the tracks stuck with me. The album notably lacks a standout hit. I believe the album could grow on me after a few more listens, but there isn’t one track that immediately stands out.

Nevertheless, it’s enjoyable to hear some retro sounds, and perhaps my opinion will change with time. However, after the initial listen, I would rate it a 3.5 out of 5.

Prompt engineering has nothing to do with engineering or coding; it’s more about communication. AI is not a mind reader; you have to tell it what you want. No matter how clear the picture is in your mind, there’s always room for interpretation. Anyone who has ever written a briefing for an agency or a freelancer, or has left the hairdresser with a fucked-up haircut, knows how difficult this can be.

Many AI tools have a chat-like interface, and you should use it like that. Communicate! You can refine your result iteratively, which can be frustrating, or you can give as much information as possible upfront. According to RISEN, you start with R.

Role:

Define the role the AI has to fulfill. If you need a copy text for advertising, you can write something like: “You’re a professional copywriter.” Next, you write what needs to be done.

Instructions:

To stick with my example of the copywriter, instructions could be: “You need to write a text for an advertisement.” Next, provide the necessary information to create the content such as:

• Who is the target group?
• What is the key message of the text?
• What product or service is advertised?
• What is the benefit or USP of it?
• What is the concept or design of the advertisement?
• Who is the brand that advertises? •
• What is the brand’s voice?
• etc.

To get a dynamic text, you should define the structure of it.

Steps:

A copywriter usually starts with a good headline to grab the reader’s attention and interest. Next, they would write a short intro to the text, which later features the USP or 2-3 benefits to generate desire in the readers. Lastly, you need a call to action.

End goal:

The goal of the text would be a catchy headline (or some proposals for it) and a copy text. Narrow: The last step is to narrow the task down with precise instructions. You can define the text length, must-haves, if you use a corporate glossary, or no-gos to avoid wrong words or expressions.

These are the five steps of RISEN that help to create awesome results with AI. As you can see, good results are not achieved with a single sentence. Sometimes you have to write a thousand words to create 300 good words. And when you’re done with prompting and the result still sucks, you can still send the briefing to a professional copywriter. This is just a basic example but think of the possibilities. And maybe you need to write a cover letter for a CV? Get it done with AI.

Of course, sometimes you don’t even have an idea what you’re looking for and just need inspiration or someone to talk to. That’s also a way to go because in the end, it’s all about communication.

Super Panavision 70 was a widescreen film format developed in the 1950s by Panavision. It offered a combination of technical advantages and aesthetic qualities that set it apart from other film technologies of the 1950s, contributing to its enduring legacy in the history of cinema.

Some took an AI to generate The Simpsons as real-life characters in that look, and the outcome is remarkable. It’s not the first time someone has generated real-life Simpsons images, but with that late ’50s/early ’60s colorful look, the video gets a neat little twist. And I think I can smell Barney Gumble through the video.

It won’t take long, I guess 2030 at the latest when we will see the first fully AI-generated movie. And I think the cast will be a mix of invented actors together with already dead actors, e.g., James Dean, Humphrey Bogart, or Marilyn Monroe.

Currently, all the AI-generated content is impressing as eff, but it’s also pretty concerning. AI can be a useful tool to boost your creativity, but in the wrong hands… well, you know how this will end.

As of today, AI is already generating playlists according to users’ tastes, and there are rumors that Spotify is pushing internal artists with exclusive deals into prominent playlists. Now, they are transitioning from algorithmically generated playlists to algorithmically generated music.

But how will users react to this? While big stars may not be directly affected, what about small and indie artists? Many people I know simply put on a playlist for background music without much concern for the artists. As a consumer myself (though I use Apple Music), I understand how easy it is to do so.

What are the alternatives then? Bandcamp could be one, but it needs to offer a streaming app to remain competitive. To achieve that, it requires a competitive pricing model for both users and artists. Additionally, artists need to have the rights to exclusively sell/stream their music on that platform. Furthermore, artists have to become their own managers—a task that, as an artist myself, I know can be burdensome. Artists want to focus on making music and performing, not on promotional activities, accounting, and fulfillment.

In my opinion, this is the time for artists like Taylor Swift to take a stand. If artists like Billy Caso were to leave Spotify, it might prompt some people to consider a change. But without stars on Spotify, people will start seeking alternatives.

Alternatively, Spotify might need to reconsider its approach.

But today, a good friend of mine told me about udio, and I have to admit I am deeply impressed. I asked the AI to generate a song in German (my native language and a tough one) about the Libertarian Party (FDP) and how they obstruct the climate and mobility transition with their fairy tale arguments of avoiding debts and relying on the self-regulation of the market, in the style of a famous German punk rock band.

First, udio recognized the band and derived the genre mix from it: punk, rock, ska, alternative. Then it generated two ideas, each 30 seconds long. From here, you can go different ways: extend the song (add sections before, after, intro, or outro), remix it, or publish it. I chose to extend the version I liked more, and again two versions were generated. After some time of listening, comparing, rehearing, and generating, I had a solid 2:30 song, with lyrics that make perfect sense 99% of the time, and I’d even say 5% of it is really good.

Afterward, you can decide to publish, remix, extend, or download the song, and even generate a video.

AI has come a long way, and it’s changing the world rapidly. I’m really curious to see where this is going.

Here’s the song I created (as mentioned before, it’s in German).

And here’s the link to it: https://www.udio.com/songs/d12anEL76bPCNavkY8ntap