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

I passionately hate hype, especially the AI hype

unixdigest.com

Post reply on HN
Print view View original

221–227 of 227 posts

Re: I passionately hate hype, especially the AI hype

#221

Post by Lerc »

the_af wrote: post #220
Lerc wrote: post #214

Earlier quoted context omitted.

I don't think there was any malfunction. The conflicting parameters implicitly contained permission to lie to the crew. The directive to take the crew to Saturpiter but also to not let them learn anything of the new mission directive meant deceiving. It's possible HAL's initial solution was to impose a communication blackout by simulating failures, then the crew reactions to the deception necciatsted their deaths to…

> I don't think there was any malfunction. The conflicting parameters implicitly contained permission to lie to the crew. The malfunction I mentioned happened earlier in the movie, when HAL incorrectly predicts a failure in a ship module (a communications antenna if I remember correctly), but when one of the astronauts checks it out and finds no hardware fault, HAL recants. After this incident the crew, knowing HAL c…

The Antenna 'malfunction' was what I was referring to. It seems quite a sensible first step to attempt to control the flow of information in order to carry out a deception. I think there was no 'psychosis'. Just a logical processing of the rules.

I don't think incorrect usage constitutes a malfunction. Hitting your thumb with a hammer is not a malfunctioning hammer. If the head came off and hit your thumb it would be. The distinction is that the first it is creating an unforseen result while performing the behaviour it was designed to do. The head coming off is the hammer not performing the function it is designed to do.

I think HAL was doing precisely what it was designed to do.

>This is equivalent to a poor prompt, it feels like splitting hairs.

I don't think so. I think it is implicit in the notion of aprompt that it is specifying a request in service of a single entity. Two different conflicting prompts in service of different people is a particular situation that should be distinct from poor prompting.

Re: I passionately hate hype, especially the AI hype

#222

Post by sidrag22 »

alganet wrote: post #169
sidrag22 wrote: post #154

Earlier quoted context omitted.

when dial up was initially the new thing you could also take that away and the world would keep chugging, same for phone lines, same for electricity. but you take away something the world has had time to become dependent on and suddenly the world has a harder time. The worlds dependency on LLM tools just hasnt had time to develop and that doesnt mean it wont. most people here are likely on the bleeding edge of utiliz…

Most people here, unless researchers actually browse HN (unlikely, that would be a dumb move), are equivalent to notepad bleeding edge users in regards to LLMs. If you don't train LLMs, you are just an user. I am sorry, that is the reality and it is cruel (for now). If "prompt engineering" becomes a thing, it means the tech is less impressive than it declares itself to be. It should be a natural language based interf…

you are going straight down from researchers building the LLMs to prompt engineers.. and there is a massive gap in the middle which is what i was speaking about. When i refer to tools that utilize LLMs, im not referring to tools built USING an llm to create it. im referring to tools which will serve a product which itself utilizes the LLM.

to use the same past example as simply as possible, computers utilize electricity. in this case LLMs would be the electricity. (this isnt me saying LLMs are as revolutionary as electricity, just trying to make a very clear example)

Re: I passionately hate hype, especially the AI hype

#223

Post by Nition »

skydhash wrote: post #101
Nition wrote: post #96

Earlier quoted context omitted.

One type of question that a 20%-failure-rate AI can still be very useful for is ones that are hard to answer but easy to verify. For example say you have a complex medical problem. It can be difficult to do a direct Internet search that covers the history and symptoms. If you ask AI though, it'll be able to give you some ideas for specific things to search. They might be wrong answers, but now you can easily search s…

> For example say you have a complex medical problem. Or you go to a doctor instead of imagining answers.

Medical was just one example, replace with anything you like.

As another example, you can give the AI a photo of something to have it name what that thing is. Then you can check the thing by its name on Google to see if it matches. Much easier than describing the thing (plant, tool, etc) to Google.

Re: I passionately hate hype, especially the AI hype

#224

Post by skydhash »

Nition wrote: post #223
skydhash wrote: post #101

Earlier quoted context omitted.

> For example say you have a complex medical problem. Or you go to a doctor instead of imagining answers.

Medical was just one example, replace with anything you like. As another example, you can give the AI a photo of something to have it name what that thing is. Then you can check the thing by its name on Google to see if it matches. Much easier than describing the thing (plant, tool, etc) to Google.

Having the wrong information can be more detrimental than having no information at all. In the former case, confident actions will be take. In the latter case, the person will be tentative wich can reduce the area of effect of bad decisions.

Imagine the lambda person confronted with this:

  sudo rm -rf /
What is the better situation, having no understanding of what it does or believing that another action will take place?

Re: I passionately hate hype, especially the AI hype

#225

Post by the_af »

Lerc wrote: post #221
the_af wrote: post #220

Earlier quoted context omitted.

> I don't think there was any malfunction. The conflicting parameters implicitly contained permission to lie to the crew. The malfunction I mentioned happened earlier in the movie, when HAL incorrectly predicts a failure in a ship module (a communications antenna if I remember correctly), but when one of the astronauts checks it out and finds no hardware fault, HAL recants. After this incident the crew, knowing HAL c…

The Antenna 'malfunction' was what I was referring to. It seems quite a sensible first step to attempt to control the flow of information in order to carry out a deception. I think there was no 'psychosis'. Just a logical processing of the rules. I don't think incorrect usage constitutes a malfunction. Hitting your thumb with a hammer is not a malfunctioning hammer. If the head came off and hit your thumb it would be…

Hm, I think you're right about the reason HAL initially reports the failure. It's made explicit in the novel, and only hinted at in the movie -- though then spelled out in the followup "2010", I reviewed my notes ;)

I still think this is firmly in the realm of "bad prompting", and if so, it's not outrageous to think it could happen in the real world, were LLMs connected to hardware peripherals handling mission critical stuff.

Re: I passionately hate hype, especially the AI hype

#226

Post by nl »

YeGoblynQueenne wrote: post #82
nl wrote: post #74

Earlier quoted context omitted.

Sure, databases didn't get as much hype but that's partly because they are old. Look at something more recent: "cloud", "social networking", "blockchain", "mobile". Plenty of hype! Some delivered, some didn't.

Which of those things claimed it would be "like the renaissance" or that we'd cure all diseases? In the clip I link above Hassabis says he hopes that in 10-15 years' time we'll be looking back on the way we do medicine today like the way they did it in the middle ages. In 10-15 years time. Modern medicine - you know, antibiotics, vaccines, transplants, radiotherapy, anti-retrovirals, the whole shebang, like medieval…

A quick search:

"Blockchain will cure cancer": https://archive.trufflesuite.com/blog/blockchain-will-cure-c...

"The Cloud will cure cancer": https://techcrunch.com/2012/03/29/cloud-will-cure-cancer/?gu...

"Why The Cloud could hold the cure to all diseases" https://www.cloudcomputing-news.net/news/why-cloud-could-hol...

"Bitcoin is bigger than the renaissance": https://hackernoon.com/bitcoin-is-bigger-than-the-renaissanc...

I could go on, but you get the idea.

Re: I passionately hate hype, especially the AI hype

#227

Post by Nition »

skydhash wrote: post #224
Nition wrote: post #223

Earlier quoted context omitted.

Medical was just one example, replace with anything you like. As another example, you can give the AI a photo of something to have it name what that thing is. Then you can check the thing by its name on Google to see if it matches. Much easier than describing the thing (plant, tool, etc) to Google.

Having the wrong information can be more detrimental than having no information at all. In the former case, confident actions will be take. In the latter case, the person will be tentative wich can reduce the area of effect of bad decisions. Imagine the lambda person confronted with this: sudo rm -rf / What is the better situation, having no understanding of what it does or believing that another action will take pla…

The process I'm suggesting is:

1. You have a complex or vague question that you can't search easily via Google etc

2. You ask the AI and it converts that to concrete searchable suggestions (in this case "sudo rm -rf /")

3. You search "sudo rm -rf /" to check the answer.

Step 3 is designed to (hopefully) catch this kind of problem.

Post reply on HN