zlacker

Their giant teams are an organizational failure then.

Product reviews alone, whether it is enterprise software or sports clothing should be something that they can easily comb through by hand, as humans, and uprank sites that are putting out more than affiliate link assemblies.

>>basch+(OP)
Why wouldn’t they do this, if it’s scalable and profitable? Is your theory just “incompetence”? I can’t prove that’s wrong, but I feel the simpler answer is more convincing; that handing off page rank to an army of minimum wage call center workers (for every country & language in the world…) wouldn’t be more effective than the existing algorithms at filtering out spam.

>>basch+(OP)
Absolutely: the challenge is that any signal that you use to identify "good websites" from "bad websites" will be adversarially optimized by incredibly motivated people.

You are dealing with a moving target that has a huge financial incentive. It's a very difficult problem.

>>bbor+I1
Some form of paralysis. They dont have a single leader who has taken responsibility for it and corrected it. CEO/SVP has too many pots on the stove.

Im legitimately asking, who is responsible for Search at Google? Prabhakar Raghavan is SVP, Search, Assistant & Ads, and I click under him, he has 8 product groups reporting to him, and none of the people are responsible for Search. Yossi Matias is responsible for Search Engineering.

https://theorg.com/org/google

It may at first come off as a laughable answer, but Google Search has been in a directionless spiral since Marissa Mayer left. Her Yahoo tenure was not well received, but at Google she cared about the end quality of the product. Her title was Search Products and User Experience. Notice how we have gone from User Experience to Search Engineering, forgetting about the people who actually use the product.

>>Fede_V+J1
I don't agree. You start whitelisting good content manually. If babygearlab is the best result for baby gear, you start hardcoding it. If seriouseats is the best result for recipes, you hard code it. If someone better comes along, they get moved up the priority list.

You figure out a way to crowdsource certain decisions and establish who you can trust. Ask them questions with right and wrong answers. You start to tackle it one product category at a time. Instead of pagerank, which was a web of who linked to who" you start figuring which voters you have who consistently turn in good feedback.

This is some form of metamoderation that slashdot tried to implement.

If you are going to be a tastemaker, stop hiding behind "the algorithm" having some mind of its own that cant be controlled.