5 ms·

Search engines were good at this, before the Internet optimized for it. Once the the Internet optimizes for ai search, it too will go the same way

by Den_VR 15d ago

Gareth321 15d ago
Search optimisation is deterministic. The algorithms are quite sophisticated but even an older LLM can easily work around them right now. There will be some kind of AI arms race but I think the net effect of that will destroy search as a platform. What arises from the ashes? I can only speculate. Possibly distributed search. Possibly something like torrent swarms. uTorrent's distributed search has been available for decades and it's surprisingly resilient. Possibly something like I2P or Tor. Possibly AI agents creating amorphous and transient search tasks/communities. Even in these early days it's possible to cobble together hobby search engines using various sources for cheap/free. I think Google sees the writing on the wall, which is why they're trying to turn Google search into an AI chatbot.
- megagpt4 15d ago
  Generative engine optimisation is a separate field. Websites that are SEO aren't necessarily GEO, they're having to be reoptimized for that. One strategy is to post fake conversations onto Reddit, where one bot how to do something, and another bot says the best way is with the product being optimised. LLMs have no connection to ground truth. It won't matter how many or how cleverly you network them to each other, they still can't tell the difference between real users saying a product is good and good spammers saying a product is good.
  - Gareth321 15d ago
    > LLMs have no connection to ground truth. It won't matter how many or how cleverly you network them to each other, they still can't tell the difference between real users saying a product is good and good spammers saying a product is good. I think this sounds much worse than it is. LLMs reason. They don't have an "opinion" about the truth. Meaning that the harness/prompt/tooling/data, determines the quality of the output. Using an LLM to navigate the truth is surprisingly effective right now with the right inputs and constraints.

Search engines were good at this, before the Internet optimized for it. Once the the Internet optimizes for ai search, it too will go the same way

by Den_VR 15d ago

Search engines were good at this, before the Internet optimized for it. Once the the Internet optimizes for ai search, it too will go the same way

Gareth321 15d ago

Search optimisation is deterministic. The algorithms are quite sophisticated but even an older LLM can easily work around them right now. There will be some kind of AI arms race but I think the net effect of that will destroy search as a platform. What arises from the ashes? I can only speculate. Possibly distributed search. Possibly something like torrent swarms. uTorrent's distributed search has been available for decades and it's surprisingly resilient. Possibly something like I2P or Tor. Possibly AI agents creating amorphous and transient search tasks/communities. Even in these early days it's possible to cobble together hobby search engines using various sources for cheap/free. I think Google sees the writing on the wall, which is why they're trying to turn Google search into an AI chatbot.

megagpt4 15d ago
Generative engine optimisation is a separate field. Websites that are SEO aren't necessarily GEO, they're having to be reoptimized for that. One strategy is to post fake conversations onto Reddit, where one bot how to do something, and another bot says the best way is with the product being optimised. LLMs have no connection to ground truth. It won't matter how many or how cleverly you network them to each other, they still can't tell the difference between real users saying a product is good and good spammers saying a product is good.
- Gareth321 15d ago
  > LLMs have no connection to ground truth. It won't matter how many or how cleverly you network them to each other, they still can't tell the difference between real users saying a product is good and good spammers saying a product is good. I think this sounds much worse than it is. LLMs reason. They don't have an "opinion" about the truth. Meaning that the harness/prompt/tooling/data, determines the quality of the output. Using an LLM to navigate the truth is surprisingly effective right now with the right inputs and constraints.