!ENTRIES!

(This was an article, originally posted on linkedin – but as this site is a protest on social media, I’m reposting it here!)

I’m a big fan of LLMs since the first GPT version dropped. It’s an amazing tool and I’ve said it before, built for testers – since by its nature it is a curiosity tool and this is the core gig of a software tester. However, I’m kind of amused on how it became the core of all technology companies as it is a tool and nothing more really. The parallel with my other love —photography, film in particular—is quite interesting. See, digital and automation tried to kill film; it kind of did, but film’s making a huge comeback. So let’s dig down!

I started my photography career with digital. The Nikon D50/D70 samples stunned me—the idea to have an instant high quality “magazine” like photograph, written on a small memory card was amazing – the idea that you can instantly correct unsavable exposure errors with RAW files was a photography sci-fi! Entirely new sensors build to capture at scale and quality! They took commercial photography to new heights, exploding with social media to the current point, where even an iPhone “outperforms” digital cameras (physics says bullshit, but that’s another rant). I’ve always loved B&W photography and the “film look.” I mastered B&W conversion, mentored by a favorite photographer. Still, old film photos—masterpieces or family snaps—had this special thing that makes Leica film cameras to currently cost a fortune. In 2016, a teammate tossed me a 1950s viewfinder camera from his basement. I’d shot film as a kid, but back then, it was point, shoot, and pray with the “great automation” of cheap cameras like Skina. Now, knowing photography, I decided to load that Ilford HP5 and see what it was all about. I shot randomly, and developed it myself with internet guides, expecting nothing. When I saw that negative though, something completely flipped. What I saw was both bad and perfect. Every flaw on that film suddenly looked like a unique tattoo, a photo build to last, and this was just on the surface – deep down was the real dopamine boost that you actually did it, you shot it yourself and developed it out of thin air – and it was good – with no Lightroom switches whatsoever. All those megapixels, razor-sharp, expensive optics, high-end photographic equipment became simply irrelevant for me at that moment. Film’s imperfection was perfect, just like human creativity’s messy soul. And…

AI’s sterile outputs can’t touch that.

It is and will become too perfect, too diluted by hype, an unfulfilled mainstream promise. And this is happening at scale – when is the last time an AI tool got you excited? It is GPT 4o for me. Remember the fiascos? Apple, Google, whole companies died trying making AI hardware – billions are already lost and the result is – we have AI slop everywhere – they even think that Hollywood would be replaced, while Netflix will produce AI video slop blockbusters with a few prompts. It sure was super exciting in the beginning; the image generation was amazing, the ability to code, to ask, to perform on completely new grounds, to learn in focused mode! But at the end of the day, you will never put an AI-generated image on your wall or be proud of your own work if it is entirely generated without your input. You didn’t steal code from Stack Overflow before LLMs – you stole the idea on how to solve your own problems and craft the solution that you need. In art school, all students learn the same thing and paint the same object or model, and yet there is never a duplication – even when the subject IS duplication. We perform with our own skill; everything we do is a combination of everything that has happened in our lives and we always admire human achievements. There is no universal path or algorithm, because creativity does not live in a computation. That is why LLMs are not even big enough to compare as invention to computers, they are simply not, they are just a useful or wasteful layer. The tech industry is not hyping because they will replace all development jobs, they hype because they are currently pouring billions into it. Have you opened your work application recently – like Jira or gmail – could the AI implementation in these apps can be any more sad? Finding Jira tickets with “AI” or have Gemini write your email – who would read this? Have you noticed that you are not finding any AI generated ads (except when the hype is that they were created using AI)? And ads are the moving force of the internet – creatives should have been long gone, together with the whole production industry.

Should you be concerned about your job? Hell yeah—if you can be outperformed by a prompt, yes, you’re in immediate danger. But remember? Everything in technology was build around people, because they are the ones giving the dollar from the first calculators or Palm PDAs. The game is big and it will become ugly – monetization will be the ultimate evil again. The only possible monetization with AI is again – you. No one will ever earn from LLMs if their customer is an AI agent, they will earn from the same profiling and human behavior analysis. I would argue that this is the techs ultimate implementation, knowing you better than you know yourself. So don’t worry, we are again the product, the promise of AI is the little toy that guy from the van gives you.

Film is not dead, you too – so get off this stupid site, find a camera in someone’s basement, load film and take a walk outside – it is free… for now. 🙂