So, a cuttting edge AI model turned out to be much cheaper and easier to produce than we thought. Weird reason to call something a "fad". Here's to hope nobody invents a way to produce much cheaper and faster cars, or this whole Car Fad will be over too.
From M-W dictionary:
Fad : a practice or interest followed for a time with exaggerated zeal : CRAZE
Books are also a fad, if "for a time" stretches across lifetimes; advanced technology that requires some infrastructure, dependent on a complex system, but not nearly as much as computers do, nevermind AI, and which is popular amongst a subset of people.
Ignore earth-systems collapse and the underlying technology that keeps these fads afloat will cease to function.
Computers are a cool diversion but not essential for human "survival and thrival", nor is our widespread embracing of the technology without consequences for life on earth.
I would far rather talk with other biased humans than the regurgitations of some biased amalgam-bot made with stolen data, even if it can act syncretically.
My bias is as a high-school science teacher who likes helping students gain understanding about the world, and hopefully wisdom, a sense of awe and responsibility, and their own sense of purpose.
Remember in the mid-90s when people talked about the internet being a fad? I think AI is a fad in the same way, which is to say not a fad at all no matter how much certain entrenched interests wish it were so.
> Remember in the mid-90s when people talked about the internet being a fad?
People keep bringing this up and it is pure bullshit.
I lived through the 90's and the Internet was never seen as a fad back then. It was hyped through the whole decade as the future, in many ways rightfully so. The hype was so strong that it culminated in the dotcom bubble early next decade.
You may think that AI skepticism is due to "entrenched interests" that want it to be a fad, I argue that AI hype is due to "entrenched interests" wishing that all overprimises are real.
I regularly use AI - Mosltly local models with either Ollama or Stable Diffusion.
I find it mildly useful in some specific scenarios, but very far from being comparable to the internet in terms of how ubiquitous and necessary it might become.
Well, it was a bubble. There was a major burst in the turn of the century that proved it, with many consequences.
A bubble just meant that the valuation of internet companies at the time were overinflated and detached from reality, not that the Internet as a technology was useless.
I think comparing AI to the internet in terms of usefulness is absolute wishful thinking thinking. The Internet was a major inflection point in the history of the world, maybe in the same magnitude of the advent of computers or the industrial revolution.
AI (and we should be clear that we are actually talking about Generative AI in this context) is an interesting tech, may be pretty useful in some contexts, but it is not in the same league of the previous examples.
Good luck with that. It's not your fault (hopefully). No really we all wanna pat your head, but not until we see results. Bring it fucker.
In the meantime you should expect everything to fall apart for reasons you're completely ignorant of and disconnected from. Maybe your fault? Who knows and who cares? hahahahahah
You can roll over your investments as we all do. You gotta think about number one. Act quickly. You're supposed to be smart money not dumb money...
The fad part is that LLMs don’t work. People keep mistaking simplistic demos for practical applications.
They don’t provide good summaries. They don’t help non-experts simulate expert work. They don’t provide reliable search results.
If someone promoted a calculator that gets 90% of the digits correct in its answers and 90% of the time those digits were in the right order, that would be a useless calculator.
I have not spoken with any AI fanboy who can substantiate his claims about the usefulness of LLMs. I have used Deepseek twice, now, and both times its results were unusable for engineering purposes but would have impressed tipsy people at a party.
I have heard credible reports that Co-pilot is helpful. And I routinely use ChatGPT for prototyping tools— but that makes it one more interesting tool, not a revolution.
I mean, just because something is a fad doesn't mean it's not _real_. "There was a fad for X in the 1980s" doesn't necessarily, or even usually, mean that X doesn't happen today, just that it is no longer such a big deal. You could call the late 19th century railway bubble a fad of sorts; railways obviously still exist and are very important, but speculative railway building is no longer a double-digit percentage of the global economy, say.
Like, in ten years it is likely that LLMs will be used for some things. It is less likely that people will be talking about spending literally trillions of dollars on LLM arms races, however.
