4 ms·

Interestingly as soon as a system works or we understand it's principles we stop using the term "artificial intelligence". It becomes image recognition, autonom

Interestingly as soon as a system works or we understand it's principles we stop using the term "artificial intelligence". It becomes image recognition, autonomous cars, or face detection. We mostly talk about artificial intelligence when we dont really know or understand what we are talking about :-)

AstralStorm 5y ago
Not really, the autonomous machine learning systems designed to solve games with limited observability are called AI properly, unlike say game scripts on fuzzy logic that are attempting to fool the user into thinking the machine is actually smart; or ones that work on fully observable games. (So no, AlphaZero is not exactly an AI.) And for an intelligence, they are still pretty bad at figuring things out.
Aicy 5y ago
Exactly, the same happens with the word "technology". Are scissors technology? They used to be but now they are comparatively too well understood and simple. Just like with AI, we label things technology that are on the forefront and not yet well understood and ironed out.
- klyrs 5y ago
  Yes, even simple machines are technology.
- webmaven 5y ago
  > Are scissors technology? To an anthropologist or archeologist scissors are most definitely a technology. As are basket weaving, flint-knapped tools, speech and writing, etc.
sidpatil 5y ago
This is known as the AI effect. https://en.wikipedia.org/wiki/AI_effect#AI_applications_become_mainstream https://en.wikipedia.org/wiki/AI_effect#AI_applications_beco...
aspaceman 5y ago
I look forward to rereading this comment repeatedly over the next decades. What pointless definition chasing.
cgearhart 5y ago
I think it’s more an issue of talking past each other. In order to build a program, you kind of need to specify the required capabilities, and then standard engineering practice is to decompose that into a set of solutions tailored to the problem. But intelligence is not just about a list of capabilities; they’re necessary conditions, but not sufficient. This is what leads to the AI effect conflict. We describe some capabilities that are associated with intelligence, build systems that exceed human performance ability on a narrow range of tasks for those capabilities, and then argue about whether or not that limited set of capabilities on a narrow domain of tasks is sufficient to be called “intelligence”. Recognizing faces, playing chess, and predicting the next few words in my email are all things I’d expect an intelligent agent to be able to do, but I’d also expect to be able to pick up that agent and plop it down in a factory and have it operate a machine; put it in the kitchen and have it wash my dishes; or bring it to work and have it help me with that. We already have machines that help me do all of those things, but none of them really exhibit any intelligence. And any one of those machines getting 10x better at their single narrow domain still won’t seem like “intelligence”.
6gvONxR4sf7o 5y ago
There's more nuance here than is usually given credit. I think it's more that once we understand the principles and the system works, we realize it never really needed AI. Here's how the story goes: Someone comes along and poses a problem. It seems like an automated solution will require building something like R2-D2. In other words solving it will need AI. Then someone else comes along and finds a way to solve it that looks nothing like R2-D2. Maybe they solve it with a massive dataset and a GLM. Turns out it never really needed AI. As a field, we're prone to thinking R2-D2 is necessary and sufficient for a task, when it keeps turning out it's a sufficient condition but not a necessary one for so many tasks.

AstralStorm 5y ago

Not really, the autonomous machine learning systems designed to solve games with limited observability are called AI properly, unlike say game scripts on fuzzy logic that are attempting to fool the user into thinking the machine is actually smart; or ones that work on fully observable games. (So no, AlphaZero is not exactly an AI.) And for an intelligence, they are still pretty bad at figuring things out.

Aicy 5y ago

Exactly, the same happens with the word "technology". Are scissors technology? They used to be but now they are comparatively too well understood and simple. Just like with AI, we label things technology that are on the forefront and not yet well understood and ironed out.

klyrs 5y ago
Yes, even simple machines are technology.
webmaven 5y ago
> Are scissors technology? To an anthropologist or archeologist scissors are most definitely a technology. As are basket weaving, flint-knapped tools, speech and writing, etc.

sidpatil 5y ago

This is known as the AI effect. https://en.wikipedia.org/wiki/AI_effect#AI_applications_become_mainstream https://en.wikipedia.org/wiki/AI_effect#AI_applications_beco...

aspaceman 5y ago

I look forward to rereading this comment repeatedly over the next decades. What pointless definition chasing.

cgearhart 5y ago

I think it’s more an issue of talking past each other. In order to build a program, you kind of need to specify the required capabilities, and then standard engineering practice is to decompose that into a set of solutions tailored to the problem. But intelligence is not just about a list of capabilities; they’re necessary conditions, but not sufficient. This is what leads to the AI effect conflict. We describe some capabilities that are associated with intelligence, build systems that exceed human performance ability on a narrow range of tasks for those capabilities, and then argue about whether or not that limited set of capabilities on a narrow domain of tasks is sufficient to be called “intelligence”. Recognizing faces, playing chess, and predicting the next few words in my email are all things I’d expect an intelligent agent to be able to do, but I’d also expect to be able to pick up that agent and plop it down in a factory and have it operate a machine; put it in the kitchen and have it wash my dishes; or bring it to work and have it help me with that. We already have machines that help me do all of those things, but none of them really exhibit any intelligence. And any one of those machines getting 10x better at their single narrow domain still won’t seem like “intelligence”.

6gvONxR4sf7o 5y ago

There's more nuance here than is usually given credit. I think it's more that once we understand the principles and the system works, we realize it never really needed AI. Here's how the story goes: Someone comes along and poses a problem. It seems like an automated solution will require building something like R2-D2. In other words solving it will need AI. Then someone else comes along and finds a way to solve it that looks nothing like R2-D2. Maybe they solve it with a massive dataset and a GLM. Turns out it never really needed AI. As a field, we're prone to thinking R2-D2 is necessary and sufficient for a task, when it keeps turning out it's a sufficient condition but not a necessary one for so many tasks.