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

Chatbots: Still dumb after all these years

mindmatters.ai

Post reply on HN
Print view View original

311–320 of 426 posts

Re: Chatbots: Still dumb after all these years

#311

Post by ClumsyPilot »

melony wrote: post #48
version_five wrote: post #42

Earlier quoted context omitted.

I think a problem is the tighter cycle between academic discoveries and business people trying to monetize them. Large language models were developed, objectively a great achievement, and immediately someone nontechnical wants to apply their own interpretation and imagine that we can build a chatbot that you won't have to pay, and before you know it, people are selling and even deploying them. Anyone who questions or…

That's a good thing. The market is very good at empirically validating research. It helps filter out the chaff and ensure research is useful. Better than academia wasting years on concepts like the "semantic web" where nobody can even agree on what it means. Academia is under attack from many political directions right now, being able to show useful output will help it thrive in the long run.

"The market is very good at empirically validating research."

That's why the market for homeopathy and fake medicine is worth $30,000,000,000, right?

Re: Chatbots: Still dumb after all these years

#312

Post by Jensson »

fault1 wrote: post #297
pdimitar wrote: post #292

Earlier quoted context omitted.

Hmmm. But I really didn't mean robots with mostly hardcoded requirements and a few excellent optical recognition algorithms (which might truly be the real boss here). I actually do mean a walking robot that can find the exact knife it needs to fillet a fish, regardless of whether it's in the sink, in its proper place on a stand, on the table, or dropped on the floor (and if it finds it on the floor it will wash it be…

> I actually do mean a walking robot that can find the exact knife it needs to fillet a fish, regardless of whether it's in the sink, in its proper place on a stand, on the table, or dropped on the floor (and if it finds it on the floor it will wash it before usage first). I mean a robot that can open the fridge and find the tomatoes, regardless if your brother has moved them to the top shelf or if they are where the…

> From a computer vision perspective, I think most of that is fairly easy

No, not at all. The hard part is to do computer vision to understand what you can do with objects, not just convert object representation of images into string texts. For example, can I move this object away to get an object under it? Can I place an object on this object without it toppling over? Is this object dirty and moving it will smear things all over? If I topple this object, will it break? How much pressure can I apply to this object without breaking?

Those things are necessary to have an intelligent agent act in a room, every animal can do it, and our software models are nowhere near good enough to solve them.

You need the computer vision program to also have a physics component so it can identify strain, viscosity, stability etc on objects, and not just a string lookup table it needs to understand that naturally like humans as the same object category can have vastly different properties identified by looks.

Re: Chatbots: Still dumb after all these years

#313

Post by petra »

tetha wrote: post #93
dexwiz wrote: post #45

Chatbots are about as useful as phone trees. They can help solve the top 5 easy/common problems, but they are useless passed that. Anyone who has worked in a call center knows that more than half the calls are about the same couple of issues: reset password, finding common help docs, etc. Since help desks are cost centers, it makes sense to half a robot handle as many of these as possible. I think most of the hate di…

This is what our NLP teams are indeed working on. Sales prefers to describe it in more colorful ways, but practically, they are developing a more natural interface to an initial decision tree - you don't press 5 for a connectivity issue, you type "My fritzbox doesn't internet anymore" into a chatbox and it recognizes that as a connectivity issue. This goes on for 3-4 questions, until it either generates a case a huma…

If creating a shallow question based decision tree is the goal, why Is NLP needed ? What's wrong with just creating it with intuitively phrased questions ?

And if there's a difference in results between those 2 methods, how big it is ?

Re: Chatbots: Still dumb after all these years

#314

Post by paxcoder »

pwinnski wrote: post #13

Having worked in ML at two different companies now, I think that people interpreting model output as intelligence or understanding says much more about the people than about the model output. We want it to be true, so we squint and connect dots and it's true. But it isn't. It's math and tricks, and if human intelligence is truly nothing more than math and tricks, then what we have today is a tiny, tiny, tiny fraction…

No post body was provided.

Re: Chatbots: Still dumb after all these years

#315

Post by DangitBobby »

repomies69 wrote: post #231
klabb3 wrote: post #195

Earlier quoted context omitted.

> Anyone who questions or points out that the technology doesn't do what the business people think it does[...] Uh oh, we've got a downer! Jokes aside, I'd like to consider an even simpler explanation, namely that "The purpose of a system is what it does"[1]. In this case, it would suggest decision makers are fully aware that they suck. Why would anyone want something that sucks? Because it's discouraging, and custom…

The chatbot saves money. Simple as that. People get served by chatbots, get frustrated, and majority gives up and doesn't bother the company with the problem. It doesn't really matter how the chatbot saves the money, they can just see the end results and use the money for bonuses instead.

That's definitely how some companies use it. I needed a human from Comcast to help me with something, and instead of getting me on with a human (after spending a few minutes on the phone trying to communicate to the bot why I needed to talk to someone) it sent me an SMS with a link to their online chatbot and promptly hung up. I then called back and said my reason for calling was to "cancel my service," which of course got me through immediately to a human.

Re: Chatbots: Still dumb after all these years

#316

Post by TedShiller »

tcgv wrote: post #237
TedShiller wrote: post #227

Earlier quoted context omitted.

The reason laypeople want it to be true is because experts present it as being true.

And marketers

Yes. And I certainly don’t see many AI experts or engineers trying to stop the marketers.

Re: Chatbots: Still dumb after all these years

#317

Post by fault1 »

Jensson wrote: post #312
fault1 wrote: post #297

Earlier quoted context omitted.

> I actually do mean a walking robot that can find the exact knife it needs to fillet a fish, regardless of whether it's in the sink, in its proper place on a stand, on the table, or dropped on the floor (and if it finds it on the floor it will wash it before usage first). I mean a robot that can open the fridge and find the tomatoes, regardless if your brother has moved them to the top shelf or if they are where the…

> From a computer vision perspective, I think most of that is fairly easy No, not at all. The hard part is to do computer vision to understand what you can do with objects, not just convert object representation of images into string texts. For example, can I move this object away to get an object under it? Can I place an object on this object without it toppling over? Is this object dirty and moving it will smear th…

I agree with all that (though I don't really necessarily associate any of that with computer vision, perhaps it's my physics bias).

Having good state estimation, kinematic, and dynamics models of real world objects is something that is very mature in controlled environments, but not very mature in other environments.

Re: Chatbots: Still dumb after all these years

#318

Post by petra »

shrdlu2 wrote: post #200

Having worked on conversational AI (virtual agents/chatbots) for over a half a decade now, I can say that there are large differences in the capabilities of the solutions and the quality of implementations. Some are just bad and unhelpful. Some are very good. I'm personally familiar with several voice deployments doing millions of calls a month. Not only are there obvious cost savings but the calls handled (entirely)…

Can you name a few of the best companies, to get an impression of the best looks like ?

Re: Chatbots: Still dumb after all these years

#319

Post by tootie »

pdimitar wrote: post #217
pwinnski wrote: post #13

Having worked in ML at two different companies now, I think that people interpreting model output as intelligence or understanding says much more about the people than about the model output. We want it to be true, so we squint and connect dots and it's true. But it isn't. It's math and tricks, and if human intelligence is truly nothing more than math and tricks, then what we have today is a tiny, tiny, tiny fraction…

You didn't have to go for the jugular and do the brutal kill on the first hit, dude! > We want it to be true, so we squint and connect dots and it's true. That's exactly the issue. You summarized it once and for all and we can all go home and stop discussing it now and forever (until we get a general AI that is). 1. Normal people want to have intelligent machines. They watch movies and series and imagine one day a ro…

Many years ago I wrote this spreadsheet import tool. One of the fields required data a little too rich to fit in single cell value so I came up an "encoding" that read like a sentence. It was sorta NLP but only understood one sentence worth of syntax. I thought it was some clever UX. Users thought they were talking to an AI. They'd just type whatever expression or thought they wanted in that field. And of course the parser would just choke.

Re: Chatbots: Still dumb after all these years

#320

Post by kumarvvr »

pwinnski wrote: post #13

Having worked in ML at two different companies now, I think that people interpreting model output as intelligence or understanding says much more about the people than about the model output. We want it to be true, so we squint and connect dots and it's true. But it isn't. It's math and tricks, and if human intelligence is truly nothing more than math and tricks, then what we have today is a tiny, tiny, tiny fraction…

It has always been my strong belief that ML is good to solve problems, if the problems are very narrowly defined and training sets are very specific.

Say, recognition of cats. Other than that, it is difficult to build stuff around something more broad.

So, ML is a brand new paradigm that has its uses, just not in AGI.

Post reply on HN