State of the Game

I'm betting i can target as good as such a system by just showing ads related to the content people are actually trying to consume ...than trying to target the individuals or groups of people by directly categorizing them. and my method would require no data collection.
it is just what has been done since decades already, just that in the 20th century you got a catalogue of target groups from mailing services and in those catalogues were a whole bunch of target groups and you could select a couple of them and the data base would then just include those addresses, which were matching all criteria. Advertisers want to minimize the addresses they have to pay for, but which are not matching their target group - so this is quite important on the advertiser-side, that the target group is met as closely as possible.

Your method is failing already when we click on video links in this forum - that is often stuff we want to see in this moment to react to it, but are otherwise not interested in - if you would now get ads about those, they would miss their target group but will eventually have to be paid for nevertheless - if it is pay per view and not pay per click for example.

have you seen those "people who watched Obsidian Ant videos are watching this as well" comment on some channels or videos on youtube - that is labelling you and putting you into an interest group - and suggestions are made based on the underlying categorized group - namely those people, who watched videos from Obsidian Ant and their behavior as a group.
 
Last edited:
suggestions for similar videos on streaming sites do a much better job at targetting me as a consumer than the ads even on those same networks.

I'm pretty sure the ad networks just straight lie to the advertisers to sell more impressions to get more money out of them...ruining the perceived efficacy of the entire system as far as the consumers are seeing. And the advertisers are too checked out to notice.

the advertisers that are paying attention cut their ad budgets significantly a couple years ago and noticed absolutely no change in sales.
 
suggestions for similar videos on streaming sites do a much better job at targetting me as a consumer than the ads even on those same networks.

I'm pretty sure the ad networks just straight lie to the advertisers to sell more impressions to get more money out of them...ruining the perceived efficacy of the entire system as far as the consumers are seeing. And the advertisers are too checked out to notice.

the advertisers that are paying attention cut their ad budgets significantly a couple years ago and noticed absolutely no change in sales.
not sure they lie to their customers - results of deep learning or other AI algorithms are not per se logical - they are successful in regards to the fitness function or on their way to get better at least - but how they do it in the end can be very surprising, because a "winning" algorithm can sometimes even appear to be illogical. Think of AlphaGo for example, who won against the world's top player 5 times in a row - and some of it's moves are so surprising, even looking like a lost opportunity - and still it won.

We are working with complex adaptation, not with deep learning, and it is there the same - those algorithms survive, which are better than others, but those are not necessarily logical and one might have a hard time to figure out why the algorithm is doing something. We had to accept, that irrational stuff is part of the game and might be used by the AI, if it provides an advantage - and that some of the found algorithms cannot be engineered, because they would have to deal with infinite complexity and unpredictable chaos - whereas the AI is purely experience based and adapts to actual events, it doesn't have to take the complexity of parameters into account and chaos is just something temporarily arbitrary, which is hard to adapt to, so it will favor more logical paths meanwhile, because those provide a higher adaptation chance. But in the end it will be successful.

i think, if ads are not getting any success at all, the target group is simply the wrong one or the product or service is not up to the task to get more or better customers. But an AI can eventually help to analyze the data, how customers have used the ads and reacted to it - eventually the ad itself is just not selling the product well enough or at all.

Well, i will give an example actually - think of an investment, one needs money to invest - people, who have a house and paid off most of the mortgage could increase their mortgage and invest the money instead to have a house mainly paid off already. This would be a good idea, if the investment is not too risky and outweights the additional costs of the now higher mortgage - do you think people will do that, even it is a good idea?- Probably not, they worked for their house all their life, they won't risk it. And so an ad for an investment which suggest this financing method is likely to fail - and gain no customers, even the target group might be correct - people with a nearly paid off house.

Edit.: well, i checked the match of alphaGo against Lee Sedol - Lee actually won game 4, AlphaGo resigned - but still, a win of 4-1 for AlphaGo.
 
Last edited:
Sorry, there is only one burger gif

FLAME-GRILLED.gif
 
@Rubbernuke & the rest of the crew...

Not banned yet...

 
@Rubbernuke & the rest of the crew...

Not banned yet...

Brilliant! And a good piece of PR to get the community even more in love with Sally than they already seem to be... even the ardent Doom™sters seem loathe to attack Sally's postings compared to any other FDev representative
Good to see a bit of bonhomie on the threads instead of the usual moaning snowflakes... 👍👍
 
Brilliant! And a good piece of PR to get the community even more in love with Sally than they already seem to be... even the ardent Doom™sters seem loathe to attack Sally's postings compared to any other FDev representative
Good to see a bit of bonhomie on the threads instead of the usual moaning snowflakes... 👍👍
We did liaise with Sally before going ahead. she was obviously up for it :ROFLMAO:
 
@Rubbernuke & the rest of the crew...

Not banned yet...

Absolutely brilliant! And I knew in my heart that Sally wouldn't mind!
 
Back
Top Bottom