Very funny, unique, and well written little scifi prompt, but clearly we’re nowhere near that level of technology. AI sucks and it’s only getting worse at the moment.
I work in tech and I don’t like what AI is turning into or how it’s being used. But saying it’s only getting worse is just wrong. Many hurdles in generation, comprehension have been overcome. Context windows are growing in orders of magnitude. Cycles with different logical approaches can be applied to the same prompt. I mean just topically you can look at the quality of deep fakes over the last 3 years.
We can hate it, but let’s not let that blind us to reality.
i have excellent news for you! we’re approaching what will probably be called the Third AI winter, which refers to periods when ai funding dries up after people realize the limitations of the current generation. the last two were around 1975 and 2000. if you expect to live 30 more years then you will probably see the Fourth AI Spring!
but to be clear, the height of AI technology by the end of the last AI winter in 2000 was, like, voice transcription software, and at the end of the first one in 1975 all they had was perceptrons (basic handwriting recognition)
TBH I don’t think we ever left the first AI Winter. The fundamental issues with their models and our models are virtually the same, we just made it wrong faster.
Completely agree. There are some things ai clearly struggles very hard with but the rate at which it is improving in just about every single facet is insane. I mean just take a look at this, this was unimaginable like literally months ago. That thing is 100% ai generated. Sound, visuals, script… everything is made from a prompt in a few minutes. And we’re still basically just at the infant stage. Just imagine what the hell these things will be doing in 20 or 50 or 100 years.
Very funny, unique, and well written little scifi prompt, but clearly we’re nowhere near that level of technology. AI sucks and it’s only getting worse at the moment.
I work in tech and I don’t like what AI is turning into or how it’s being used. But saying it’s only getting worse is just wrong. Many hurdles in generation, comprehension have been overcome. Context windows are growing in orders of magnitude. Cycles with different logical approaches can be applied to the same prompt. I mean just topically you can look at the quality of deep fakes over the last 3 years.
We can hate it, but let’s not let that blind us to reality.
i have excellent news for you! we’re approaching what will probably be called the Third AI winter, which refers to periods when ai funding dries up after people realize the limitations of the current generation. the last two were around 1975 and 2000. if you expect to live 30 more years then you will probably see the Fourth AI Spring!
but to be clear, the height of AI technology by the end of the last AI winter in 2000 was, like, voice transcription software, and at the end of the first one in 1975 all they had was perceptrons (basic handwriting recognition)
TBH I don’t think we ever left the first AI Winter. The fundamental issues with their models and our models are virtually the same, we just made it wrong faster.
This is definitely marketing driven.
Thank God! Too dystopian for me
You should read the Murderbot Diaries. If you find this funny and unique, you will like Murderbot.
How is AI getting worse?
Maybe some LLM models are getting worse, but we keep finding new ways to train models even more efficiently.
A year ago running a local LLM would have been a fairy tale, now we have things like Deepseek and AnythingLLM
Completely agree. There are some things ai clearly struggles very hard with but the rate at which it is improving in just about every single facet is insane. I mean just take a look at this, this was unimaginable like literally months ago. That thing is 100% ai generated. Sound, visuals, script… everything is made from a prompt in a few minutes. And we’re still basically just at the infant stage. Just imagine what the hell these things will be doing in 20 or 50 or 100 years.
It has never produced viable product and after sampling its own piss the new iterations are actually less and less accurate.
That’s a very interesting take