I will make next big search engine!
- Thread starter Buzzika
- Start date
Monrox
Power Member
- Apr 9, 2010
- 617
- 588
Right now, what could be a big hit are sites that are excluded from showing. Not delisted illegal material or spammy nonsense but stuff that's not to the SE liking. They are relevant and interesting but nobody can reach them. E.g. when Bing is showing 'some results have been removed' at the bottom. What results? Which results? Where are they? Also they must be good enough as they have to be excluded from page 1.
For example trying to find how to enable right click on some flash banner to stop it from doing well whatever irritating things it's doing returns no hits other than setting the options during development. Now, there is NO way that someone has not yet come up with a solution. But the info never gets shown.
Such a safe haven will attract MANY users and it will get promoted by word of mouth just like G was in the beginning. But it is a mundane task to (write the software to) sift through all that is actual spam so I don't mind sharing the idea.
For example trying to find how to enable right click on some flash banner to stop it from doing well whatever irritating things it's doing returns no hits other than setting the options during development. Now, there is NO way that someone has not yet come up with a solution. But the info never gets shown.
Such a safe haven will attract MANY users and it will get promoted by word of mouth just like G was in the beginning. But it is a mundane task to (write the software to) sift through all that is actual spam so I don't mind sharing the idea.
Buzzika
Supreme Member
- Jul 8, 2009
- 1,446
- 1,776
Welll maybe I may not succeed in the next best AI stuff... Even if I go half way through I will be really great at what I do...If you succeed in making an AI that can recognize content, a puny perfect internet search engine would be the smallest of your income. The search won't be even needed, the AI will generate the best content on the fly by sifting through all that is available and picking the best stuff. But even that is nothing, actually the whole world will never ever have to do any amount of work again as what we do now is mainly guiding dumb machines and animals to do the actual work. All that is, if the AI can be kept submissive of course
Google isn't achieving anything on its own. They simply take into consideration what real people find good and link to it. Count those links (that's what the huge data centers are for) and put the winners on the respective top spots. Don't forget to patent it as well.
So you either create an AI that can read pages and see the best match (a fundamentally different approach than today's search engines) OR come up with a better idea to use the activities of internet users for free.
The current strong AI's problem is making a computer want to do something. This can be achieved either through punishment (pain) or reward (pleasure). If it could be made to feel either one, it will be motivated to complete a task. We can simulate an amoeba with all its organelles but it justs sits there, it's not moving towards a crystal of digital sugar... it's lacking an urge aka motivating behavior. Don't think LISP or any other current language can help here.
However one day some genius will figure it out as it always happens. But usually such persons are not very social, don't have an agenda and also are not motivated by some other desire like making money, needing to succeed or wanting to be famous.
What I am currently thinking is an AI capable to some respectable extent of natural language processing. The finest project i have seen in public is the DeepQA project of IBM which uses the IBM watson computer...