Logan Abbott’s Post

Ecosia and Qwant, Two European Search Engines, Join Forces on an Index To Shrink Reliance on Big Tech: Qwant, France's privacy-focused search engine, and Ecosia, a Berlin-based not-for-profit search engine that uses ad revenue to fund tree planting and other climate-focused initiatives, are joining forces on a joint venture to develop their own European search index. TechCrunch: The pair hopes this move will help drive innovation in their respective search engines -- including and especially around generative AI -- as well as reducing dependence on search indexes provided by tech giants Microsoft (Bing) and Google. Both currently rely on Bing's search APIs while Ecosia also uses Google's search results. Rising API costs are one clear motivator for the move to shrink this Big Tech dependency, with Microsoft massively hiking prices for Bing's search APIs last year. Neither Ecosia nor Qwant will stop using Bing or Google altogether. However, they aim to diversify the core tech supporting their services with their own index. It will lower their operational costs, and serve as a technical base to fuel their own product development as GenAI technologies take up a more central role in many consumer-facing digital services. Both search engines have already dabbled in integrating GenAI features. Expect more on this front, although they aren't planning to develop AI model development themselves. They say they will continue to rely on API access to major platforms' large language models (LLMs) to power these additions. The pair is also open to other European firms joining in with their push for more tech stack sovereignty -- at least as fellow customers for the search index, as they plan to license access via an API. Other forms of partnership could be considered too, they told TechCrunch. Read more of this story at Slashdot.

Qwant and Ecosia, two European privacy-focused search engines, are collaborating to create their own European search index. This joint venture aims to reduce dependence on Google and Microsoft's search indexes, innovate with generative AI, and cut costs amid rising API fees. Their initiative, the European Search Perspective (EUP), will enhance data sovereignty and privacy-first results, targeting France by early next year and Germany by 2025. This move aligns with Europe’s push for technological autonomy, especially in light of the growing role of AI and search data control. Credit: Natasha Lomas via TechCrunch #qwant #ecosia #europe #privacy #google #microsoft #searchindex

Google stands firm in its position that the changes will be a benefit to the web, and changes to the Search algorithm are just the start. Last week, Google CEO Sundar Pichai stood in front of a crowd at the company's annual developer conference and announced one of the most significant moves in the search engine's history. Going forward, Pichai said, Google Search would provide its own AI-generated answers to many of your questions, a feature called "AI Overviews" that's already rolled out to users in the United States. "The result is a product that does the work for you," Pichai said. "Google Search is generative AI at the scale of human curiosity." AI Overviews are just one of a slew of dramatic changes Google has made to its core product over the past two years. The company says its recent effort to revamp Search will usher in an exciting new era of technology and help solve many of the issues plaguing the web. But critics say the opposite may be true. As Google retools its algorithms and uses AI to transition from a search engine to a search and answer engine, some worry the result could be no less than an extinction-level event for the businesses that make much of your favourite content.

“Don’t Be Evil”, was Google’s launching Tech Bro tag. But leaving out the bit about making SEO waters a whirlpool of confusion from hell. Now we hear the truth - or do we? The recent leak of documents referring to “Google's Content Warehouse API” tends to only confirm what many have known all along. That Google can often put out a constant stream of contradictory messages. On topics such us, domain authority, CTR, sandbox, word count, font size weight, page titles, page disavows, dwell time, etc. Oldest game in the book. Keep ‘em guessing! Or in Google’s case – jumping through an endless fractal landscape of disappearing hoops. It’s why you’ll hear the most experienced and knowledgeable of SEOs on here repeatedly insist - you can’t trust anything what Google spokespersons officially or unofficially say! It’s liable to be the opposite the very next day. SEO true blood warriors state the only real way to develop an understanding of how the algorithm may behave and treat site content, is to: “Constantly analyse and test”, and in real time. Each individual site and its pages. Doing the hard yards. No secret sauce. No red button. If you want to clear those confused muddy waters for precious sparkling droplets of insight. Here’s one the leaked document revealed: Google's ranking system is a series of microservices rather than a single algorithm. Yep – a hydra of many hundreds of heads, each seeking to feed on a different aspect of the content. And each subject to “Twiddlers” – modifying the ranking functions to adjust search results before being displayed. Crucially, the leak does emphasise and reconfirm the importance of quality content, user engagement, and strategic link building. The very three elements SEOs and content providers have long repeatedly insisted are the backbone of sustained search results and ranking. Unless AI sweeps it all away in the next 5 years in a torrent of prompted, auto-predicted but, ultimately, fake content. Yep - real evil being done here. No document leaks needed to see this monster of a lie! #Googleleak #Googlesearch #seocontent

Search engines have always been substring-based. Google, e.g., searches for articles that literally contain the words you typed. Yes, it also tries synonyms and applies other heuristics, but at its core, it's a substring lookup algorithm. Some things are impossible to find using today's search engines, and we've all just gotten used to it. The next obvious trillion-dollar idea dawned on me: A search in LLM's "meanings" Embedding Space. Re-index the entire internet as meaning vectors in Embedding Space! --- Imagine being able to search the internet (or your local chats) for: 1. An intricate romantic feeling. 2. A joke you once told. 3...∞. Tone, complex emotion, and every other textual characteristic. These are all directions in Embedding Space. An algorithm can find a known vector that points in a similar direction. Everything is searchable by meaning, gist, complex emotion, tone, etc. --- Today, looking to read a story posted by someone who conveys very similar, intricate romantic emotions often leads to a deluge of clichéd male/female narratives. The algorithm I propose will pre-process the entire internet (and local data on your phone) by forwarding every piece of text through an LLM and storing the resulting embedding vector. Given a query, It will find vectors that point in a similar direction (that is, are close to the query vector in non-zero entries. This is done fast using algorithms like 'Faiss', 'Annoy', etc). It will then yield matches of a unprecedented quality. --- The local version of the algorithm (a meaning-based search for chats and files) is easily implementable, and could be made a start-up today. However, the internet version requires immense compute power for pre-processing. Being an advancement over traditional search engines, it's reasonable that it will be implemented by giant companies: Microsoft and Google.

Certainly! Here’s a simplified explanation of Bing’s updated AI search and the issues with Google's search algorithms: ### Bing’s Updated AI Search 1. **Introduction of AI-Powered Search**: Bing has introduced a new AI search experience designed to enhance user engagement and ensure website owners continue receiving traffic. 2. **New Layout**: The updated layout on Bing includes: - **Three Panels**: - **Table of Contents (Left)**: Helps users navigate easily through main topics and related subtopics. - **AI Answers (Center)**: Provides answers from AI with links to original websites. - **Organic Search Results (Right and Below)**: Displays traditional search results and additional organic links. 3. **Encouraging Clicks**: Bing has structured its search results to maintain website traffic by ensuring users still click through to sites, unlike other search engines that focus on zero-click searches. 4. **User-Friendly Navigation**: The layout anticipates related questions and provides an easy way to explore more information through the table of contents. 5. **Feedback and Rollout**: Bing is slowly introducing this new search experience, gathering user feedback to ensure it maintains traffic to websites. ### Google’s Search Algorithm Issues 1. **Acknowledge Imperfections**: Google admits its ranking algorithms are not perfect and acknowledges feedback about poor search results. 2. **Feedback from Users**: Users have raised concerns about the quality of search results, particularly when low-quality content ranks higher than it should. 3. **Scaling Challenges**: Google's process of indexing and ranking trillions of web pages is automated and scaled, making it challenging to address all issues instantly. 4. **Commitment to Improvement**: Google is committed to refining its algorithms by reviewing feedback and working on rewarding high-quality content. 5. **Ongoing Efforts**: Despite updates aimed at improving search results, some site owners feel their content has been unfairly affected, and Google continues to address these concerns through ongoing adjustments. By focusing on maintaining a healthy balance between AI-generated answers and organic search results, Bing aims to enhance user experience and support site owners. Meanwhile, Google is working on improving its algorithms to ensure high-quality content is appropriately ranked.

Interested in this post from Avinash Kaushik about search efficacy and search changes driven by AI. Not evidence-based and thoroughly anecdotal, but I suspect that it depends on the domain of the search. For instance, if searching for sales trends show demographic or geographic skewing, a followup query will be needed. Search in the Amazon store is quite poor when color, size or other specific attributes are included (perhaps intentionally?); you don't think to include these initially but are guided by the inappropriateness of the results. As an aside, I know it is a bit beside your point, but I found the breezy promo prose of the @Perplexity result off-putting. As though it's search was similarly casual. OTH it was good at highly focused queries, such as " Does the TP-Link TL-SG2008 switch support a DHCP server?". Contrasted with Gemini, which said no, got it wrong, and created a temporary search dead end. Perhaps I am an atypical user, but I am happy to spend another few seconds framing my question. If I am researching plug-in hybrid vehicles, don't show me hybrids without plugs. I have gone round and round in the Amazon store or Google searches into the store with frustrating results when constraints clearly articulated in the search are seemingly ignored. In other words, give me expedience when retrieving a simple fact from Wikipedia, but understanding the #intent of a question is sometimes more important than guessing at The Answer. I suspect those extended interactions with search can also sometimes yield marketing insight. #knowledgegraph #AI #cognition

Avinash Kaushik's article on Google's conservative move in "AI Search" from 'results' to 'answers' suggested that it might be time to try other "AI Search" options along with a Google search. I tried a common, top-of-mind question: "What is the best way to get from Bristol to Clarkson's Farm" with Google and Perplexity.ai. I'm not great with change, and I will miss clicking through the Reddit top result from Google, but Perplexity provided such a good answer, with website source references, I have to follow this advice. "How would you optimize your website for showing up in perplexity answers" seems like the next best question to ask. Some familiar looking #SEO guidelines in this answer.

Google faces escalating costs with the rapid growth of spam and the impending adoption of generative AI in SERPs. Implementing SGE incurs significantly higher costs compared to current methods, potentially impacting ad revenue as searchers’ needs are met more quickly reducing the number of placement opportunities. #seoconsultant One approach to reducing costs I’ve observed is a more stringent policy for what gets included in Google’s index. Anecdotally, clients with complex business models using template pages are seeing more frequent deindexing, even for pages that have been stable for years. #seoprofessional In some cases, JavaScript is the culprit. However, Google has already figured out how to deal with sites with high amounts of injection. Google should identify these pages, add them to the render queue and crawl them after the data has been injected. #seo So why is this not happening? Could this be another resource that Google is struggling to manage efficiently with the growing amount of spam? #SEOspecialist Being deindexed is a nightmare for website owners and SEO professionals. It’s like being taken out of the game entirely. #digitalmarketer However, experience has taught me that aligning with Google’s initiatives can significantly benefit our clients’ sites. #marketer The solution is straightforward: schema markup. Aside from aiding search engines in understanding content more efficiently, it can also provide a great cost-saving for Google. #marketing Schema markup aids Google’s crawlers and machine learning algorithms in understanding web content more efficiently and cost-effectively. #digitalmarketing It can play a role in reducing Google’s operational costs. The premise is that aiding Google in minimizing the resources needed for crawling, indexing and understanding your site will lead to improved visibility.

Google is moving aggressively into "AI Search," it still feels conservative. At last week's event, Philipp Schindler demoed an odd query: "how to pick the right men's running shoe for beginners with high arches" He showed the results in the new Google, at 12:58 here https://t.ly/k0Kt- . I cannot replicate those results on live Google. That is ok. But, one defining feature of the next gen search is that humans are unlikely to seek with another variation of "more keywords in my query" search. If I know there is a "super intelligent thing" at the other end, it is more likely I would switch... From: "help me search better". To: "help me get to the answer faster". The query would then be: "which is the best men's running shoe for beginners with high arches" I typed that into Google and Bing, you can see the results below. Google is still in "old Search mode." Bing has adopted "new Search mode." After Google I/O 2024 announcements, my expectation was that Google would be now in "new Search mode." Answers vs. results. Part of this is understandable, Google has a lot more to lose than Bing. It is likely to move carefully and slowly. For me, it is about tempering the expectations from Google, waiting to see them come to life a bit more slowly. [Disclosure: I am a Google shareholder.] The very best result for my "help me get to the answer faster" query mode was provided by Perplexity. You can see the result here: https://t.ly/dY6hO BOOM! Action: As an Agency, at Croud we have been aggressively seeking, learning, experimenting, researching, all sorts of ways that Organic and Paid Search experiences are changing (for humans and search engines) - our effort to ensure our clients are riding the wave vs. being smashed by it. Recommendation: I highly recommend you use other options to search along with Google. Let your mind open to how humans are seeking information - it is not how it was.