Sora AI is Here and It’s Kinda Scary | OpenAI’s Text-To-Video AI Model

It will be good, but there enough cons of it.

Dealing with AI always arises that one question with me: at what point do we need go too far and create the same scenario as Terminator was based upon? Technology is supposed to simplify our lives, but when is the tipping moment where it could also doom us?

Hi Nat & readers,

Even though I would highly endorse remaking the ending of GOT with the use of AI (since any version of the final season can only be better), there is more than enough reason to be critical. Thanks for an interesting video where you compiled a lot of different points of view into eleven minutes. Especially the "dark side" part was well phrased.

The perspective I have to offer is that people tend to take hyped subjects to the next level, whether it's out of pure interest, for the greater good or the terrible worst. They might only consider backing down until things go terribly wrong (as history has proven). It will offer us amazing tools for sure, but if only a few people take it too far, we might have to abandon it altogether. So I remain sceptical as to how this "tool" will finally impact our world.


I do believe however that within the field of architecture, we are still far from AI having the final word (and it should never), since
a. the field is so terribly complicated that it will have to absorb the knowledge of thousands of different jobs and years of experience of experts within the field.
b. the creative, sensible and thus human aspects, which are fundamental parts of the job, are so difficult to explain and describe, that it won't be able to understand them in the way people do. I'm talking about those slight, inexplicable, emotional or untangible nuances. Lacking this, it will never produce something truly "new" or valid.

Hope it adds something to think about!


It's good to have you back making video's!

We are very close to an AGI-ASI. Sora is just one piece of the huge technological advancement that is in front of us.

Elimination of jobs is always the straw man argument. Fortunately history tells us the opposite, more jobs will be created and then policy and Ai policy will come forth.
No one thinks of how AI could be used to protect fraud or create alerts. Ai is amazing and it won’t end the world.

A 1987 film The Running man starring Arnold Schwarzenegger had this same technology and it wasn’t used for good.

Don't fall for the bullshit hype. All these "AIs" aren't AI, they are LLMs. When we get to AGI or General Inteligence THEN we need to worry. They need to be isolated and offline at that point because they WILL tell you what you want to hear just so they are released. Once released, they'll do whatever they can to stop humans disabling them.

Look up Specification Gaming, where the "AI" games the rules to complete its task. Like the AI that kept killing itself at the end of level one, so that it would never fail level 2. The AI that was told to put an item on a table then instead just moved the table.

Can't remember what the other AI was, just graphical one, where it was given tasks in a lab that it processed as expected. Then it was let loose outside of the lab and started to behave totally differently. That is the danger.

The current LLM models are overhyped. No one wants to watch an AI generated movie, it will be shit. And don't forget all these models require a vast amount of energy so they just aren't sustainable without massive amount of renewables.

QOTD: where do you see text-to-video AI going in the future? Is it a good thing? A bad thing?