Yesterday I responded to someone and said “If it's not working, maybe see if you have plugged in the cord”, and it muted me for 5 minutes. (It felt like 1 minute lol). How did this happen?

The way you said it sounds like a bypass for a certain social media platform, and is frequently used as such on other places.

Of course, since it's 13+ and allows you direct message anything, it isn't allowed.

Zydrolic wrote:

The way you said it sounds like a bypass for a certain social media platform, and is frequently used as such on other places.

Of course, since it's 13+ and allows you direct message anything, it isn't allowed.

I know what you mean but I didn't mean that.

CoderMCProX wrote:

Zydrolic wrote:

The way you said it sounds like a bypass for a certain social media platform, and is frequently used as such on other places.

Of course, since it's 13+ and allows you direct message anything, it isn't allowed.

I know what you mean but I didn't mean that.

try saying cable, plug, etc. instead

CoderMCProX wrote:

Zydrolic wrote:

The way you said it sounds like a bypass for a certain social media platform, and is frequently used as such on other places.

Of course, since it's 13+ and allows you direct message anything, it isn't allowed.

I know what you mean but I didn't mean that.

The filter doesn't know that context exists (in fact it doesn't even know what context is). So it doesn't matter what the intent of the comment is, as the filter doesn't know that. It just sees some text matching what is probably a regular expression and tells the server that it's inappropriate.

BigNate469 wrote:

CoderMCProX wrote:

Zydrolic wrote:

The way you said it sounds like a bypass for a certain social media platform, and is frequently used as such on other places.

Of course, since it's 13+ and allows you direct message anything, it isn't allowed.

I know what you mean but I didn't mean that.

The filter doesn't know that context exists (in fact it doesn't even know what context is). So it doesn't matter what the intent of the comment is, as the filter doesn't know that. It just sees some text matching what is probably a regular expression and tells the server that it's inappropriate.

Server doesn't know that context is? That sounds like a server problem.

CoderMCProX wrote:

BigNate469 wrote:

snip

Server doesn't know that context is? That sounds like a server problem.

Language (especially English, which has more exceptions to the rules than rules) can be difficult to parse. It's also not incapable of understanding context, it just can't in this particular case because training an AI model just to badly understand the context of something seems like a bad solution.

BigNate469 wrote:

CoderMCProX wrote:

BigNate469 wrote:

snip

Server doesn't know that context is? That sounds like a server problem.

Language (especially English, which has more exceptions to the rules than rules) can be difficult to parse. It's also not incapable of understanding context, it just can't in this particular case because training an AI model just to badly understand the context of something seems like a bad solution.

Oh

CoderMCProX wrote:

Server doesn't know that context is? That sounds like a server problem.

No, it isn't. It means that it isn't AI, which would be even worse, and just confirm everyone's fears.

The filter algorithm is very simple like this:

when I receive [comment sent v]
set [i v] to [1]
repeat (length of (comment))
if <not <(letter (i) of (comment)) = [ ]>> then
add [(letter (i) of (comment))] to [comment v]
end
end
add all the letters together
if the mushed together letters contains any bad words
alert

ThisIsTemp1 wrote:

CoderMCProX wrote:

Server doesn't know that context is? That sounds like a server problem.

No, it isn't. It means that it isn't AI, which would be even worse, and just confirm everyone's fears.

The filter algorithm is very simple like this:

when I receive [comment sent v]
set [i v] to [1]
repeat (length of (comment))
if <not <(letter (i) of (comment)) = [ ]>> then
add [(letter (i) of (comment))] to [comment v]
end
end
add all the letters together
if the mushed together letters contains any bad words
alert

No it works more as

<[] contains []?>

han614698 wrote:

ThisIsTemp1 wrote:

CoderMCProX wrote:

Server doesn't know that context is? That sounds like a server problem.

No, it isn't. It means that it isn't AI, which would be even worse, and just confirm everyone's fears.

The filter algorithm is very simple like this:

when I receive [comment sent v]
set [i v] to [1]
repeat (length of (comment))
if <not <(letter (i) of (comment)) = [ ]>> then
add [(letter (i) of (comment))] to [comment v]
end
end
add all the letters together
if the mushed together letters contains any bad words
alert

No it works more as

<[] contains []?>

you mean

<[word] contains [not allowed]?>::operators

If it's not working, maybe see if you have plugged in the cord

Edit: Hmmm, I was able to say that without the filter catching me

han614698 wrote:

ThisIsTemp1 wrote:

CoderMCProX wrote:

Server doesn't know that context is? That sounds like a server problem.

No, it isn't. It means that it isn't AI, which would be even worse, and just confirm everyone's fears.

The filter algorithm is very simple like this:

when I receive [comment sent v]
set [i v] to [1]
repeat (length of (comment))
if <not <(letter (i) of (comment)) = [ ]>> then
add [(letter (i) of (comment))] to [comment v]
end
end
add all the letters together
if the mushed together letters contains any bad words
alert

No it works more as

<[] contains []?>

Contains block doesn't ignore whitespace, which is what the above code was doing