When will such AI be possible?
OpenAI has released ChatGPT4. When do you think things like this will be possible? 5 years? 10 years? 20 years?
Let's say I'm a student and have five subjects. For each subject, I have handwritten notes, PowerPoint and PDF scripts, Excel spreadsheets, charts, learning videos, and possibly even audio recordings.
An AI that could take all of this as input and create engaging and well-explained learning videos in the most condensed form possible would be brilliant. It puts all the information into context, recognizes connections, and explains everything clearly within a framework, like a series with multiple episodes. Let's say 20 episodes, each lasting an hour, and creates these episodes with narration. It even creates 3D graphics, etc. It also creates graphics, etc., from my own scripts or those from the professor.
And the AI would then, based on all the input material, set meaningful tasks of varying difficulty.
(1 votes)
Loading...
Within the next 1-2 years. Maybe now if you have the right connections….
One must not forget the AI development is not linear! From a certain point that can occur at any time, the systems will quickly improve themselves and will also independently get new training data….
The people who talk here about 10-15 years do not yet know what is already possible for absolutely incredible and creepy things with the help of AI.
(Free) HQ video creation has been dramatically improved over the last 3-4 months. Sometimes they are no longer different from real videos!
It’s all speculation. In principle, such tasks are already feasible by today’s AIs, but I would not want to learn with it because of hallucinations. If you are optimistic, such problems are perhaps 5 years of history, but I think that something will last at least 10-15 years. As the technology will develop in the future, it remains to be seen, but the rapid growth of recent years is likely to slow down with time.
So:
As far as I know, the generation of videos for AI is very difficult. So at the moment not possible. I can explain it easily. Videos consist of 60 frames per second. A AI, if we leave the training outside, simply needs a lot of time and energy for processing. Currently there are only 5s videos that make “sense” possible.
I’m not saying it’s impossible. But in general it will not be possible in the next 5 years.
That looks different to Hollywood…
This is already possible today, but of course not for free