"Traditional Value Proposition: “I know how to do this because I’ve done it many times before.”
New Reality: “AI systems know how to do this because they’ve analyzed millions of examples and can apply that knowledge instantly.”
The professionals who built their careers on accumulated knowledge—the standard practices, established relationships, and procedural expertise—are discovering that their competitive advantage is evaporating. When AI can access decades of industry knowledge in milliseconds and apply it with mathematical precision, experiential knowledge alone becomes insufficient."
https://corser.substack.com/p/the-ai-revolution-isnt-replacing
1 person likes this
Well... no, not really. Depending on what kind of AI application you're talking about. Technical tools like have been developing in audio and video processing applications, structural design, chromakey replacement, point cloud generation, multicamera controlling and AI-assisted match-moving, etc. - these are technically repetitive and dull, sometimes mathematically exhausting exercises. So far as the hypothetical cinematography application - to the extent it's technical, it could apply but the traditional term, "machine learning" is much more accurate and appropriate and "AI" which is a marketing term. And these tools are useful but they're not actually creative in themselves. The only established tool the article talks about is script analysis, and IMO they are an absolute joke, with no actual application except possibly to justify the value of a script on paper, to someone with no interest in reading the script themselves. The AI on those systems being pioneered a decade ago and still without any real success by Slated.com - and now Largo.ai, brandedstreams.com, and others we deal with through our studio. Consider trying to predict the success of the next passenger car, by comparing it's proposed engineering schematics to the final engineering schematics of previous vehicles. If you don't see how crazy that is, then I can't help you. Yet that is exactly what script analysis software tries to do. On the other hand, deep research with some LLMs are useful - so long as (a) they have actual access to useful data and the ability to process statistical models and databases (most don't) and (b) the researcher already has some deep knowledge in they area being researched. Because while it's true that the machine can accumulate data, it has no way of analyzing it or even making it useful unless the people who are programming it know the actual area they are coding for - which unfortunately is rare - then the results are almost be definition mediocre, and often entirely erroneous and unreliable. Relying on the AI without a deep knowledge of the area you are relying on it for, is very dangerous and guarantees you will have serious problems at some point. Someone on another post observed that AI tools are helping to make those of us who are superior creators even better and those of us who are not even worse. I agree with that observation.
1 person likes this
You'd have to define what a superior creator is.
1 person likes this
Willem Elzenga Someone who creates something good and original rather than trite and mediocre.
1 person likes this
I think I did both. Nevertheless thank you for your insight and thorough reply.