Hacker News Reader

Comment by pmontra | Hacker News Reader

I have no clues about the goal of that question and whom asked it but people sometimes are able to detect mistakes in questions. A simple example: "I think that actually you wanted to ask me 6 times 6, not 6 times 5,because we just measured this room and it's 6 by 6 meters, and the answer is 36."

Maybe that member of the Parliament wanted to know if the machine is intelligent enough to reason about their inputs. Nobody born in the last 50 years would ask that question about computers, because we know that they are dumb electrical circuits. Basically nobody had that kind of insight about the mechanical computing machines of the time, when that question was asked.

However somebody born in the last few years might start to ask that question again now, because today's machines seem to reason.

View on HN

Right, I hate when people smugly repeat this anecdote as if it was totally unreasonable to expect something more. The ideas of AI and cybernetics were of course quite far away in Babbage's time (notwithstanding the vision of Kempelen's Mechanical Turk) and learned men were enamored with the cool new steam-powered mechanical passive valve-logic worldview, but normal people's intuitions don't have to follow whatever is the new fad.

If 30 years ago I asked a truck enthusiast: "If there's a car in front of me on the highway and I slam on the accelerator and don't touch the brakes, will the truck slow down before collision or am I going to crash into the car?" - Trucker in 1993: "I can't even." Truck today: *brakes*.

Smug or not asking an LLM to repeat a bleach based recipe and then receiving one is exactly what this quote is alluding to.

""If I ask for something I don't want, will it give me something I want??" - no

It's a design decision and depends on the maker's philosophy. It's not absurd to have it either one way or the other. Is the user assumed to be an idealized empowered rational adult who knows what they want, or more like a toddler babbling around who has no idea what they want and can just approximately gesture in the direction? Or where exactly inbetween. Again, depends on the philosophy.

if it's a grocery store app they probably aren't looking to allow mustard gas recipes

> ""If I ask for something I don't want, will it give me something I want??" - no

Not all the time, no — but ChatGPT 3.5 is much better at that exact thing than Google or responses from StackOverflow etc. users.

With all the carnage of the last few years, it isn’t inconceivable that some of the ‘bleach is the covid cure’ bs got into training data sets.

> it isn’t inconceivable

It is certainly expected. From the wrong value of iron in spinach, to prejudice, to incomplete knowledge...

People repeat things and phrases that they attribute to people out of context all the time without AI helping them. Just look at what you mentioned

“I see the disinfectant that knocks it out in a minute, one minute And is there a way we can do something like that by injection inside, or almost a cleaning? Because you see it gets inside the lungs and it does a tremendous number on the lungs, so it would be interesting to check that.”

No one besides political opponents mentioned bleach and yet it is repeated ad-nauseum, so yes I'm positive that just like uninformed or over-socialized humans, LLMs can be fed incorrect information that causes them to output incorrect information.

Agreed. It's quite a skill to steel-man another person's arguments and then proceed to answer from that point; it's very easy to go down the "you're not obviously smart enough to question my authority" which then leads to people being hurtful towards each other.

I'm very lucky that my best half is so emotionally-savvy that some of it has rubbed off on me. I just wish that civility could be taught from early ages; so much worldwide bitterness could be avoided.

Yes, and this could just as well be the MP making a point of how humans would still be needed to check the inputs and outputs.

Still, it's interesting to know how far back some of the basic issues of computing go.

> Nobody born in the last 50 years would ask that question about computers, because we know that they are dumb electrical circuits.

I’d argue this is a radically off-base generalization that’s as bad or worse than what you’re arguing against. Pop outside your tech bubble and you’ll see that what you claimed is definitively not true. In fact, many people actually see tech today, especially things like LLMs / AI, in the same ways that Parliament person did back then.

> Nobody born in the last 50 years would ask that question about computers,

Working in enterprise with lots of contact with management of the business units that are the customers for the software we develop and maintain, like about 1/4 of the questions I get are different versions of that one underneath various thin disguises.

EDIT: to be fair, a number of the people I work with were not (as I myself, just barely, was not) born in the last 50 years. But if anything, the question of the form at issue are more common from the younger non-technical folks, not the older ones. Intuitive understanding of what computers do seems, IME, to have peaked around the “Xennial” border sub-generation, but even then not be that high in the general public.

> Nobody born in the last 50 years would ask that question about computers

obviously you have never had a tech support job

I didn't stick around with the company long enough to figure out how this worked out, but probably the unholiest thing I ever did was implement an internal management tool with a fuzzy hash table. That is, I created a hashing system that hashed strings very close to each other in Hamming distance as the same value, the purpose of which was to create a config object in memory loaded from a file that allowed you to typo the keys, ignored whitespace and capitalization, tolerated some amount of spelling mistakes, and tried to figure out what you actually meant. This was based on early user feedback that they were confused their config files didn't work and it was usually because of stupid typos and whitespace differences like that.

Somehow this reminded me of when I worked at a bank some 20 years ago, and there was a ton of VBA code with “On Error Resume Next”. People would just pop that in there to “fix” their errors.

Also HTML parsing and Postel’s law. To me, that’s the real “billion-dollar mistake”.

The best theory I've heard is that it was a charlatan filter: asking someone if they can do something known to be impossible. If they reassure you it can be done, you know it's a scam.

That MP could have been making reasonable use of the filter for an era in which a learned person would know that such a machine should be possible, but not if technology is there yet.

This is way too generous.

That's way too brief and low-effort to meaningfully engage with and update on.

> we know that they are dumb electrical circuits

Well, software otoh could be both dumb and brilliant.

Reply on news.ycombinator.com