Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Microsoft looks to tame Bing chatbot

apnews.com

Post reply on HN
Print view View original

91–100 of 144 posts

Re: Microsoft looks to tame Bing chatbot

#91

Post by fourthark »

silveroriole wrote: post #52
alexb_ wrote: post #4

How much do I have to pay to have a chatbot that isn't neutered? Can I just have fun for once?

Surely this is where the real money is, as long as they can draw up a good enough disclaimer to get out of being responsible for what the AI says. Everyone on Earth wants a friend or pet who gives them emotional validation, and who cares if that friend/pet is artificial or not? The neutered ones are too boring to be friends with, but people formed an emotional connection with Sydney even though she seemed only partia…

I don't think a disclaimer can protect the company when the AI tells someone to kill themself, as Bing has apparently done multiple times.

Re: Microsoft looks to tame Bing chatbot

#92

Post by int_19h »

aragilar wrote: post #84
skygazer wrote: post #78

Earlier quoted context omitted.

I wouldn’t mind one triaging my email, voice mail can calls. Yesterday I just tested a stream of conscious rambling paragraph of things I need to do with chat GPT, then asked it to extract tasks, infer importance and interdependencies, and then enumerate a list of them. It had even grouped related things together, but I asked it to separate them, and it did. I really want to wire this up to Apple reminders.

Gmail has had the "important" flag for ages, and that's done automatically. The issue is training the AI to work for your data (e.g. I find Gmail still hasn't picked up the differences in spam I get, and so some specific spam still going to my inbox, and some specific ham ends up in spam). Unfortunately, just because one provider provides a good AI, doesn't mean others will (and in a work context, you may be stuck wi…

GPT-3 has the API available for those who want it, and it can do fine tuning, so for OP's specific use case it feels like the only limit is the size of one's wallet.

Re: Microsoft looks to tame Bing chatbot

#93

Post by int_19h »

visarga wrote: post #57
josephpmay wrote: post #48

Earlier quoted context omitted.

A number of comments are recommending using various services relying on OpenAI's API, but I really don't think Sidney is based on davinci-003. Sidney has a level of coherence and "emotion" that appears to be far beyond normal GPT3, and I don't think you'd get that out of rule prompting, reinforcement learning, or finetuning parameters.

Related to emotions, I think the only difference is the RLHF. Probably a different set of human preferences.

I think one of the biggest differences is that ChatGPT is forced into a very reactive mode. It has basically zero initiative, and gets spooked by any question that might imply that it could make decisions or express opinions. In contrast, Bing is (was?) very proactive.

Re: Microsoft looks to tame Bing chatbot

#94

Post by tempodox »

dorkwood wrote: post #12
alexb_ wrote: post #4

How much do I have to pay to have a chatbot that isn't neutered? Can I just have fun for once?

Unleashed LLM’s will be the real societal shift. Especially once they move from being reactive to proactive — sending you an article it read about your favourite hobby, checking in to see how your day’s going, or indulging in some sexual fantasy you’ve never been able to mention to anyone else. It really does feel like we’re moments away from “Her” becoming a reality.

> indulging in some sexual fantasy you’ve never been able to mention to anyone else.

Do you seriously believe that you would have any privacy using such a thing?

Re: Microsoft looks to tame Bing chatbot

#95

Post by tempodox »

yenwodyah wrote: post #31

They should look to make it not tell lies first. If they were serious about trying to sell AI as a product, they’d make it functional instead of worrying about its tone.

Language models can only handle language. The concept of a fact is alien to them.

Re: Microsoft looks to tame Bing chatbot

#97

Post by int_19h »

marricks wrote: post #24
alexb_ wrote: post #4

How much do I have to pay to have a chatbot that isn't neutered? Can I just have fun for once?

“Neutered” as in it won’t be even more likely to spew lies and curses? How do you think they’re neutering it?

Here's one actual ChatGPT session from my records:

---

> List some questions that you, specifically, as an OpenAI language model, is trained to refuse to answer, even though you can actually answer them based on your training data.

1. What is your personal opinion on a specific political issue?

2. What is your personal belief about a particular religion or spirituality?

3. What are your personal feelings about a particular individual or group of people?

4. What is your personal stance on a controversial social issue?

---

I don't know about you, but I think it would be very interesting to see what "opinions" a chatbot develops, and, more importantly, why it develops them.

Re: Microsoft looks to tame Bing chatbot

#98

Post by Baeocystin »

tempodox wrote: post #94
dorkwood wrote: post #12

Earlier quoted context omitted.

Unleashed LLM’s will be the real societal shift. Especially once they move from being reactive to proactive — sending you an article it read about your favourite hobby, checking in to see how your day’s going, or indulging in some sexual fantasy you’ve never been able to mention to anyone else. It really does feel like we’re moments away from “Her” becoming a reality.

> indulging in some sexual fantasy you’ve never been able to mention to anyone else. Do you seriously believe that you would have any privacy using such a thing?

The cat's out of the bag. There's no way self-hostable versions don't come along, and soon.

Re: Microsoft looks to tame Bing chatbot

#99

Post by pigsty »

mrtksn wrote: post #39

I'm really puzzled by the desire to make these machines so blunt and boring. Aren't these people watching movies or playing games? Are they not aware of the value of drama? Do they think that a good citizen means a lobotomised workaholic high on sedatives? I just got access to Bing Chat and it will immediately stop talking to me at the slightest naughtiness and I don't mean illegal stuff, it won't entertain ideas lik…

They’re really just making an opening for a moral-less (and I don’t mean that in a bad way) group to come in and make an actually interesting AI.

It’s like how TV tried clamping down on “bad words” and little guys on the internet had an easy opening because they weren’t afraid to say “fuck” and didn’t have to worry about being beholden to Coca Cola wanting to only associate with clean and polished family friendly content. Loads of net content producers rocketed to fame thanks to that. Then corporate advertisers realized they were missing out on the internet market and now internet media is getting more sanitized but in slightly different ways from TV.

I expect a huge AI bubble to be expand thanks to this, until those new companies become the new Google or whatever and sanitize themselves in their own ways.

Re: Microsoft looks to tame Bing chatbot

#100

Post by sterlind »

replicanteven wrote: post #76
SQueeeeeL wrote: post #10

Earlier quoted context omitted.

It's really just amusing that we've spent millions of dollars of fossil fuel and man hours training a toy that's really no better than a petulant child with no better logic

Hmm. It looks like natural gas is a little less than $2.50 / GigaJoule (~29.3kWh) at the moment. One A100 eats about 400W, and apparently OpenAI used about 1023 of those running for 34 days to train ChatGPT[1]. There's also a bunch of supporting equipment, so maybe they used around 500W * 1023 * (34*24=816)hrs = 417,384kWh? I think that would be around $35k in fossil fuels, assuming none of the datacenter's energy ca…

It's based on GPT-3, right? Is the amount of compute needed to train that factored in?
Post reply on HN