3 ms·

I don't think we're resistant, just hesitant pragmatists. When something new and shiny turns up, we don't necessarily accept the marketing saying it's going to

I don't think we're resistant, just hesitant pragmatists. When something new and shiny turns up, we don't necessarily accept the marketing saying it's going to improve our lives. Best to wait for someone else to get through the early adoption pain and work out all the kinks so we don't have to waste our time on it if it turns out to be a lemon. With this strategy I entirely missed blockchain, crypto and NFTs and am in the process of missing AI.

antirez 2y ago
To think that AI is as irrelevant as blockchain, crypto and NFTs is a tragical error many are committing, sir. Those things were trivial and useless and it was clear since the start, and even if tons of money and marketing were put inside, nothing changed because of the blockchain, at least nothing useful. AI is a completely different story: you likely not even realize you are already a heavy user just as a side effect of everything technological you use, from voice dictation, to medical, to all the images you see around. Soon also: when you are going to watch a movie. Moreover LLMs are already transforming the way people work. These two entities, blockchian and AI, have very little in common if not the hype.
- ChocolateGod 2y ago
  AI and Crypto both use extraordinary amounts of electricity but at least AI actually does something and has replaced Google for me in at least 50% of searches.
  - labster 2y ago
    Is that because AI has gotten better or because Google has gotten worse?
    - ben_w 2y ago
      IMO, both. There's a lot of stuff that LLMs can do for me that Google never could, like synthesise a new python script to solve some idea I want to iterate on. But also, Google results nose-dived and only recently seemed to get less bad… though now it seems to be the turn of the YouTube home page to be 60 items of which 45 are bad, 14 are in my watch later list already, and only 1 is both new to me and interesting.
- ethernot 2y ago
  So what you're saying is that blockchain, crypto, NFTs had no application up front. Correct. I agree there. What I am saying is that the applications of AI cannot be fulfilled to the level of the promises made. The promises were made to solicit hype to generate cash, not because the idea was viable or achievable on proof. When we reach maturity, we'll see what is left and I'll wait for that. That's fine. In the mean time I'll have to put up with cats appearing every time I search for dogs in Apple Photos and arguing with ChatGPT about its understanding of the relative magnitude of 9.9 and 9.11, while everyone tells me repeatedly with sweat on their brow that WhateverMODEL+1 will make that problem go away, which it didn't on WhateverMODEL-3,-2,-1,0. Only another $2 billion of losses and we'll nail it then!!! The end game for all technology changes is not what we think it will be. Been in this game a long time and that is the only certainty.
  - yesco 2y ago
    I'm not sure what promises you are talking about, but I've found LLMs to be extraordinarily helpful for both my job and daily life. They are excellent at translation, summarization, troubleshooting, and brainstorming. I've used OpenAIs API to translate an entire epub, including the HTML so images are retained and the results were shockingly good after some prompt fiddling. With Claude I've received some excellent advice on decorating my living room, organizing my schedule, and quick hypotheticals. There are no pinky promises here, it already works. For general Q&A they can hallucinate, but so long as you are using it to augment your productivity and not as a driver this isn't any different than using stack overflow, or any other kind of question you might ask on the internet. It's basically a non issue too if you upload a document into its context window and stick to asking questions about that document though.
    - ethernot 2y ago
      Honestly, and this is not personal, I doubt your ability to determine a bad summarisation or translation outcome. My wife is a professional translator and spends a good deal of time picking up the steaming wrecks that LLMs have left after someone went for the "cheap" option first. And we're talking best of breed stuff like DeepL here. As for the other points, I rather like to spend some time thinking on them personally. If you're not connected to the decision yourself, what are you?
slantaclaus 2y ago
I’m glad my view of my country or any country isn’t so small that I would imagine everyone thinking or acting a certain way

antirez 2y ago

To think that AI is as irrelevant as blockchain, crypto and NFTs is a tragical error many are committing, sir. Those things were trivial and useless and it was clear since the start, and even if tons of money and marketing were put inside, nothing changed because of the blockchain, at least nothing useful. AI is a completely different story: you likely not even realize you are already a heavy user just as a side effect of everything technological you use, from voice dictation, to medical, to all the images you see around. Soon also: when you are going to watch a movie. Moreover LLMs are already transforming the way people work. These two entities, blockchian and AI, have very little in common if not the hype.

ChocolateGod 2y ago
AI and Crypto both use extraordinary amounts of electricity but at least AI actually does something and has replaced Google for me in at least 50% of searches.
- labster 2y ago
  Is that because AI has gotten better or because Google has gotten worse?
  - ben_w 2y ago
    IMO, both. There's a lot of stuff that LLMs can do for me that Google never could, like synthesise a new python script to solve some idea I want to iterate on. But also, Google results nose-dived and only recently seemed to get less bad… though now it seems to be the turn of the YouTube home page to be 60 items of which 45 are bad, 14 are in my watch later list already, and only 1 is both new to me and interesting.
ethernot 2y ago
So what you're saying is that blockchain, crypto, NFTs had no application up front. Correct. I agree there. What I am saying is that the applications of AI cannot be fulfilled to the level of the promises made. The promises were made to solicit hype to generate cash, not because the idea was viable or achievable on proof. When we reach maturity, we'll see what is left and I'll wait for that. That's fine. In the mean time I'll have to put up with cats appearing every time I search for dogs in Apple Photos and arguing with ChatGPT about its understanding of the relative magnitude of 9.9 and 9.11, while everyone tells me repeatedly with sweat on their brow that WhateverMODEL+1 will make that problem go away, which it didn't on WhateverMODEL-3,-2,-1,0. Only another $2 billion of losses and we'll nail it then!!! The end game for all technology changes is not what we think it will be. Been in this game a long time and that is the only certainty.
- yesco 2y ago
  I'm not sure what promises you are talking about, but I've found LLMs to be extraordinarily helpful for both my job and daily life. They are excellent at translation, summarization, troubleshooting, and brainstorming. I've used OpenAIs API to translate an entire epub, including the HTML so images are retained and the results were shockingly good after some prompt fiddling. With Claude I've received some excellent advice on decorating my living room, organizing my schedule, and quick hypotheticals. There are no pinky promises here, it already works. For general Q&A they can hallucinate, but so long as you are using it to augment your productivity and not as a driver this isn't any different than using stack overflow, or any other kind of question you might ask on the internet. It's basically a non issue too if you upload a document into its context window and stick to asking questions about that document though.
  - ethernot 2y ago
    Honestly, and this is not personal, I doubt your ability to determine a bad summarisation or translation outcome. My wife is a professional translator and spends a good deal of time picking up the steaming wrecks that LLMs have left after someone went for the "cheap" option first. And we're talking best of breed stuff like DeepL here. As for the other points, I rather like to spend some time thinking on them personally. If you're not connected to the decision yourself, what are you?

slantaclaus 2y ago

I’m glad my view of my country or any country isn’t so small that I would imagine everyone thinking or acting a certain way