Nothing will surprise readers most readers here. But still this is a deep, wide-ranging and important analysis Beware of the Google AI salesman and its cronies
2 Likes
This is a good article.
I can definitely see the conflict of interest between publishers who need web traffic to fuel their businesses, and Google who is decreasing the traffic it sends to those publishers.
I wonder if some kind of everything-app ends up killing Google as both publishers and consumers get tired of scams.
The Trump administration is doing everything possible to promote scams. So, the abuse demonstrated in this article is not going anywhere.
I saw this video of Ente where you were interviewed:
You weren’t very clear about what Mojeek will plan to do with AI.
The reason I will use Mojeek (more) is the hope to have an independent (non US) search index, of which the results will never be influence by AI. Not now and not in the future.
If Mojeek wants to add AI, then please do that as an option. I want to have good-old classic search results forever.
Welcome @marc-io
Mojeek is primarily a keyword-based search engine. Pages returned in the results are ranked objectively based on the keywords you enter and your location.
Starting last year, Mojeek changed the algorithm and allowed the semantic meaning of words to affect the ranking of search results. This change is based on AI. And, only English language results are affected.
If you don’t like this, it looks like the ssoe URL parameter is still functioning. So, you can manually turn off semantic ranking by adding &ssoe=0 to your search results URL.
For example, if you search for the keyword “aubergine” then you can add the URL parameter like so:
# Before
https://www.mojeek.com/search?q=aubergine
# After
https://www.mojeek.com/search?q=aubergine&ssoe=0
Navigating to the second example URL will show you search results that are not affected by semantic ranking.
You can also make this change permanent by manually setting a web browser cookie for Mojeek.
Finally, Mojeek added optional Summaries. These are large language model responses to your keywords which appear alongside your search results and don’t impact the ranking. Summaries can be toggled on the Mojeek Settings page.
References
| Mojeek Discourse
1 Like
Some other AI-based features in Mojeek:
- Mojeek - Limited Relevance Widget - this is likely using a combination of heuristics and AI.
- TESTING: Related Queries - Related queries are generated using an LLM. I still have this feature enabled on one of my machines but I must admit I stopped looking at it long ago.
1 Like
I don’t see this as a bad thing per se, but I would if it’s using external resource heavy LLM’s. If “AI” is just a algorithm based on a local database, then I’m ok with that.
1 Like
Next week, @Colin will probably chime in on Related Queries and any other AI stuff they’re working on.
Thank you for your questions and responses. There’s a lot of confusion about what AI is, so first some context before addressing points raised.
Much of the confusion is fueled by media, blitzscalers and Big Tech, which often present Generative AI as either a panacea or a threat. People here are aware of this hype, but to keep things grounded, here’s a brief overview.
Today, “AI” usually refers to Generative AI, especially for text prediction. This means generating text by predicting the next token — a task performed by Large Language Models (LLMs) using transformer-based decoder architectures, like OpenAI’s GPT. These are the basis for today’s chatbots.
Earlier transformer models, like Google’s BERT, used encoder-only designs and are not used for text generation, but for tasks like classification.
Generative LLMs are known to confidently output incorrect information — so-called hallucinations. This is an inevitable byproduct of the decoder design. Reducing this remains an active area of research.
Finally, LLMs are just one part of the broader AI landscape, which was more accurately called machine learning before 2017 — a term coined in 1959, reflecting decades of diverse models and techniques.
At Mojeek it is NOT our mission to generate information.
Our mission to provide links to sources of information. We do use machine learning where it helps to do this, as explained below.
With reference to posts above:
We did not change the core algorithm of Mojeek, but we did include a semantic scoring in the overall scoring used for ranking results. This improved the relevance of search results, for what is still a keyword based search engine. For the semantic scoring we use a local encoder-only model similar to BERT. That is not AI in modern parlance.
These use our own internal scoring algorithm used for ranking. So no AI in modern parlance is used.
We do believe in the value of “good-old classic search results” and unlike everyone else it seems, we are not aiming to be an answer engine. Where resource light, preferably local, machine learning models and techniques can help us be a better search engine we use them, and would be foolish not to do so.
We use generative AI, in two experimental cases, as follows
The first is the Summary feature which:
- Is a feature specifically designed to assist users in more efficiently locating the search results to click through too.
- Is designed to distill information from only the search results.
- Is off by default
- Has a button which can be removed from the display and a tab which can be removed (in the preferences menu)
The second is the Related Queries feature which are testing but have not rolled out.
As for the rest of Mojeek which is of course a vast array of IP built up in-house for 20 years, we do use some machine learning techniques and models. And of course we do develop our own algorithms. These models/agorithms are all run locally and consume low amounts of power.
“No Tracking. Just Search” is mission statement not a cliche.
3 Likes
This is what I liked to hear. Thank you!
1 Like