It’s strange – we have great new tools, but many folks are using them in such a sloppy way, that the tools get discredited. And this also tends to boil down that serious users get into trouble. You probably guess what I am talking about: AI tooling.
I am for sure nobody who jumps on the latest hype. So I resisted AI quite long for complex things. Until it was ready, which for me was around summer 2025. That was for coding. For some doc writing it was ready earlier (and I used it to cover my weak spots). Now, AI has evolved to also help with video and audio.
The past days I played with Google’s NotebookML. It always has been a great tool for hallucination-free AI processing, but it also has evolved into a great media generator. given a curated set of information, it can create great presentations but also explainer videos and even podcast-like audio. The quality is very high. I had several tries on deeply technical topics and the output is exceptionally well. I know what I am saying – I verified everything and I actually know everything by heart. I could not provide better content than what NotebookLM generated (find a sample at the bottom).
Actually, if I create that very same content, it would clearly be human. My non-native English skills would show clearly – accent, wording all the like. I know from the past that’s not a no-go, but it’s not a go-to either. Then story telling. Doing a good, coherent and not-too-long explanation takes good prep work. And finally creating the video, retrying-shots, editing … takes a lot of time or access to someone who knows all this be heart (that’s not me).
So, besides some experiments: not something I can do in a large volume. I need to engineer world-class software, consider architectures, see upcoming trends. Not much time left for the video thing. As such: no time – no video. That was the state up to now.
Then I checked back in into the state of AI video for tech topics. It’s really good, and much better than what I could do with manageable effort. With AI, I can create pretty good explainers and backgrounders. Stuff beginners can use to learn and understand. Still is some effort, but I can build on a large known-good basis, including the code itself. Remember: code is the ultimate truth – do not go for less!
I have pipeline for concepts implemented in code, generating the code, checks, changing it into doc. With AI, I can build on all of this.
But guess what? Now folks come up and think “Ah, this is AI generated, so it must be slop, so I will not use it”. Ironically, this is as lazy as it could be, judging content without even digging a bit in. As a side-note, I frequently see this pattern when people object rsyslog’s AI First project – obviously without reading even the anounement in full.
Well, laziness has its price. I think I will continue to create quality content with AI. Those that do not like it can happily ignore it. Those who dismiss it without looking closer can do so. But I know that many of them will turn to Search engines or AI Chat to obtain the answer. Well, then they will be able to actually get it. Because we at rsyslog know our way around AI and also provide the search engines and LLMs with the content they need to know. Ironically, this closes the loop.