Anyone that relates GenAI's importance to that of the Internet must not have been around when the WWW actually launched.
"Internet" was not the killer app, email and instant messaging were. Email was free through your ISP and didn't require more than a $9.99/mo Earthlink connection. The alternative was Fax / Telex etc that had zero network effect outside of businesses and required a dedicated phone line and hardware.
LLM based GenAI on the other hand has been around for long enough that we know that it's main use case is limited to helping schoolkids cheat on their essays and polluting the Internet with factory-farmed social media "content".
Anyone relating LLMs to school kids cheating hasn't seen programmers be 3x as fast using it and journalists churning out articles 3x as fast by focusing on what they do best (gather and sort facts) and leaving the tedious make-easily-readable text writing to a machine.
One can certainly have opinions about how some people use it and how they check the quality of what comes out, but as long as it's not used to make up facts but merely to do the primitive busy work, like machines are supposed to do, I don't see how that's not just as revolutionizing as the fax/telex comparison you are giving.
> LLM based GenAI on the other hand has been around for long enough that we know that it's main use case is limited to [...]
Sounds exactly like what Bill Gates said in the early days of the internet. I don't have the exact quote, but I'm sure typing half a sentence full of grammar and spelling errors into ChatGPT would give me the quote including a link to its source. I should got get it fast before that tool disappears when the hype is over and we are back to old school google searches, like God intended.
I’ve had insight into a bunch of businesses in multiple industries putting stupid money into trying to find a use for this, caught up in the hype and worried about falling behind. While also being sold AI features by every vendor, who’re all doing the exact same thing.
Every single one is floundering and very unlikely to think this was a good use of resources a couple years from now.
LLMs and associated tech are here to stay, just like search algorithms and autocomplete and machine translation programs, and the clone tool. The hype will fade, though. They’re neat tools but, no, turns out we’re not on the verge of inventing Skynet, we just fooled people into thinking we were because a prominent hype-man/grifter was saying so as a sales tactic (Altman) and because the output is human language instead of numbers or whatever.
agreed. making the world worse.
Not gonna happen.
"I can't wait for the PC hype train to die"
"I can't wait for the internet hype train to die"
"I can't wait for the smartphone hype train to die"
"I can't wait for the EV hype train to die"
I suggest you don't wait too long.
"I can't wait for the Laser Disc hype train to die"
"I can't wait for the Betamax hype train to die"
"I can't wait for the HD-DVD hype train to die"
"I can't wait for the NFT hype train to die"
"I can't wait for the dogecoin hype train to die"
Treating LLMs as comparable to the internet is a great illustration of their point.
Anyone that relates GenAI's importance to that of the Internet must not have been around when the WWW actually launched.
"Internet" was not the killer app, email and instant messaging were. Email was free through your ISP and didn't require more than a $9.99/mo Earthlink connection. The alternative was Fax / Telex etc that had zero network effect outside of businesses and required a dedicated phone line and hardware.
LLM based GenAI on the other hand has been around for long enough that we know that it's main use case is limited to helping schoolkids cheat on their essays and polluting the Internet with factory-farmed social media "content".
Anyone relating LLMs to school kids cheating hasn't seen programmers be 3x as fast using it and journalists churning out articles 3x as fast by focusing on what they do best (gather and sort facts) and leaving the tedious make-easily-readable text writing to a machine.
One can certainly have opinions about how some people use it and how they check the quality of what comes out, but as long as it's not used to make up facts but merely to do the primitive busy work, like machines are supposed to do, I don't see how that's not just as revolutionizing as the fax/telex comparison you are giving.
> LLM based GenAI on the other hand has been around for long enough that we know that it's main use case is limited to [...]
Sounds exactly like what Bill Gates said in the early days of the internet. I don't have the exact quote, but I'm sure typing half a sentence full of grammar and spelling errors into ChatGPT would give me the quote including a link to its source. I should got get it fast before that tool disappears when the hype is over and we are back to old school google searches, like God intended.
> Not gonna happen.
I’ve had insight into a bunch of businesses in multiple industries putting stupid money into trying to find a use for this, caught up in the hype and worried about falling behind. While also being sold AI features by every vendor, who’re all doing the exact same thing.
Every single one is floundering and very unlikely to think this was a good use of resources a couple years from now.
LLMs and associated tech are here to stay, just like search algorithms and autocomplete and machine translation programs, and the clone tool. The hype will fade, though. They’re neat tools but, no, turns out we’re not on the verge of inventing Skynet, we just fooled people into thinking we were because a prominent hype-man/grifter was saying so as a sales tactic (Altman) and because the output is human language instead of numbers or whatever.
Working pretty heavily with these technologies since they hit mass appeal, I'm pretty confident, but I appreciate your concern