3 ms·

> how you can talk to something that doesn't provide factual information and just take it at face value Like talking to most people you mean?

philipwhiuk 3y ago
When OpenAI buys me a drink at the bar in exchange for the rubbish it produces, I might have a more favourable view.
- ilaksh 3y ago
  As soon as they release the API, we can build an AI "bartender". Combine the voice output and input with NeRF talking heads such as from Diarupt or https://github.com/harlanhong/awesome-talking-head-generation https://github.com/harlanhong/awesome-talking-head-generatio.... You will now be able to feed it images and responses of the customers. Give it a function to call complementaryDrink(customerId) Combine it with a simple vending machine style robot or something more complex that can mix drinks. I'm not actually in a hurry to try to replace bartenders. Just saying these types of things immediately become more feasible. You can also see the possibilities of the speech input and output for "virtual girlfriends". I assume someone at OpenAI must have been tempted to train a model on Scarlett Johansson's voice.
lostcolony 3y ago
Hopefully people know not to ask others for factual information (unless it's an area they're actually well educated/knowledgeable in), but for opinions and subjective viewpoints. "How's your day going", "How are you feeling", "What did you think of X", etc, not "So what was the deal with the Hundred Year's War?" or whatever. If people are treating LLMs like a random stranger and only making small talk, fair enough, but more often they're treating it like an inerrable font of knowledge, and that's concerning.
- TeMPOraL 3y ago
  > If people are treating LLMs like a random stranger and only making small talk, fair enough, but more often they're treating it like an inerrable font of knowledge, and that's concerning. That's on them. I mean, people need to figure out that LLMs aren't random strangers, they're unfiltered inner voices of random strangers, spouting the first reaction they have to what you say to them. Anyway, there is a middle ground. I like to ask GPT-4 questions within my area of expertise, because I'm able to instantly and instinctively - read: effortlessly - judge how much to trust any given reply. It's very useful this way, because rating an answer in your own field takes much less work than coming up with it on your own.
graemep 3y ago
No individual is "most people". Most of the time I spend talking to people in real life is with people whose professional expertise, hobbies, and other sources of knowledge I know at least roughly. I have an idea how good they are at evaluating what they know and how honest they and whether they are prone to wishful thinking.

philipwhiuk 3y ago

When OpenAI buys me a drink at the bar in exchange for the rubbish it produces, I might have a more favourable view.

ilaksh 3y ago
As soon as they release the API, we can build an AI "bartender". Combine the voice output and input with NeRF talking heads such as from Diarupt or https://github.com/harlanhong/awesome-talking-head-generation https://github.com/harlanhong/awesome-talking-head-generatio.... You will now be able to feed it images and responses of the customers. Give it a function to call complementaryDrink(customerId) Combine it with a simple vending machine style robot or something more complex that can mix drinks. I'm not actually in a hurry to try to replace bartenders. Just saying these types of things immediately become more feasible. You can also see the possibilities of the speech input and output for "virtual girlfriends". I assume someone at OpenAI must have been tempted to train a model on Scarlett Johansson's voice.

lostcolony 3y ago

Hopefully people know not to ask others for factual information (unless it's an area they're actually well educated/knowledgeable in), but for opinions and subjective viewpoints. "How's your day going", "How are you feeling", "What did you think of X", etc, not "So what was the deal with the Hundred Year's War?" or whatever. If people are treating LLMs like a random stranger and only making small talk, fair enough, but more often they're treating it like an inerrable font of knowledge, and that's concerning.

TeMPOraL 3y ago
> If people are treating LLMs like a random stranger and only making small talk, fair enough, but more often they're treating it like an inerrable font of knowledge, and that's concerning. That's on them. I mean, people need to figure out that LLMs aren't random strangers, they're unfiltered inner voices of random strangers, spouting the first reaction they have to what you say to them. Anyway, there is a middle ground. I like to ask GPT-4 questions within my area of expertise, because I'm able to instantly and instinctively - read: effortlessly - judge how much to trust any given reply. It's very useful this way, because rating an answer in your own field takes much less work than coming up with it on your own.

graemep 3y ago

No individual is "most people". Most of the time I spend talking to people in real life is with people whose professional expertise, hobbies, and other sources of knowledge I know at least roughly. I have an idea how good they are at evaluating what they know and how honest they and whether they are prone to wishful thinking.