Interviews with Viktor Lofgren (from Marginalia search engine) and Simon Descarpentries (from Meta-Press.es search engine).
FreeWebSearch Day is held each year on September 29. It is a day for freedom of information and democracy. Everyone can join by organising or attending events or other actions.
Simon’s Meta-Press.es lets you explore the news without middle man between news papers and your browser. The search engine in the form of a browser add-on helps you avoid the swamp of third-party trackers on most newspaper websites and news aggregators that give you little choice how to search and select.
-
Page of the event on NLnet website
-
The interview audio (45 minutes)
1. Interview With Simon Descarpentries from Meta-Press.es
"We should not rely on search engines to free us from the effort to know the world"
Meta-Press.es
Initially published here : https://nlnet.nl/news/2023/20231012-meta-press.html
The Meta-Press.es search engine lets you explore the news without middle men or trackers. Creator and lead developer Simon Descarpentries is a free software enthusiast for over twenty years and former Framasoft employee, he currently is CEO at Acoeuro.com and treasurer of Fund for Defense of Net Neutrality. We interviewed Simon for #FreeWebSearchDay. You can listen to the recording of the interview or read the edited transcript below.
Question: Can you tell us something about Meta-Press.es?
Answer: Meta-Press is a free software allowing everyone to search through online press. With one click you can search through 930 online newspapers. All these sources are indexed one-by-one by humans. Currently, if you query GAFAM news search engines, you’ll get results which include fake or forged newspapers. Some are just copying content from other sites to sell advertisement. These GAFAM news search tools can’t sort out the real sources from the false ones because everything in the process is automated. In Meta-Press we did it all by hand, so it is all verified by humans. Meta-press also includes about 700 newspapers which give free access to their content because that is the economic model they chose. Unlike Google News, Meta-Press does not lead you to a dead end: this exists but you can’t reach it.
1.1. No censorship
Meta-press is a Firefox browser add-on and is free software, accessible for everyone. It’s built with a software architecture which guarantees there are no bottle-necks, no single point of failure and no censorship. Because once you install the add-on in your web browser your requests are not send to a hypothetical Meta-Press server which could control things. Instead, it’s your computer that is instructed how to do the search. It’s your web browser that has gained the superpower to request results from nearly 1000 newspapers. Therefore it is virtually uncensorable because you have plenty news papers on the one side and plenty of computers on the other. And that is how the internet works : no central point that we all have to cross. The more people use Meta-Press the better it will work.
1.2. Problems with search today
Q: What problems do you see with search today?
A: My main concern with GAFAM news search engines is what I mentioned before: they serve fake newspapers. It is easy to feed GAFAM news search engines with fake content designed exactly to go through their ranking algorithm. And since nobody ever checks, you can fool them. It is possible already possible for you to fool them, so it is for a government or a company. It’s flawed, we need something else. It is bloated, dishonest and not working anymore. And it will only get worse because there is money at stake. And there are political stakes too.
1.2.1. Mass surveillance as an economic model
Another big problem with search online is that it’s currently dominated by a handful of megacorporations following no rules except their own and whose economic model is based on mass surveillance. They follow everyone, knowing who came back to the website and what they were interested in, what click they made… If you had a screen next to the newspaper, displaying all the information GAFAM extract from your online activity, you would just turn your computer off. They sell that information to companies, governments and political parties. That is not a conspiracy theory, we know it’s true from Snowden and the Cambridge Analytica scandal. It would be more comfy to live in a world where we could forget about those scandals but it’s a reality we have to face.
1.2.2. Decreasing accuracy of search results
A third problem is that the accuracy of Google search is decreasing. More and more content is created just to score high in their ranking algorithm. It shows up even if it was not exactly what you wanted to see. It’s called an injection attack. Google is an open security breach regarding injection attacks. That is a technical problem in search today.
1.2.3. We’re losing our ability to be inquisitive
And last but not least, regular humans are abandoning their ability to search in favor of those dishonest tools. You could compare it to GPS. When you use GPS for navigation you’ll slowly lose your ability to read a map and other navigation skills. If GPS was dishonest and provided you with the wrong information, you would be lost and abandon the tool. It is exactly the same thing with online search. The tool is dishonest and you have to abandon it. You need to work on your skills to search for things: cross check your information, compare your sources, publish your results so others can verify them. You must become a bit of a journalist. Like Viktor Lofgren of Marginalia Search said: the more you use Google the more you’ll become fenced into a small park Google allows you to reach. The best search tool is your brain.
1.3. Information has a price
Q: How does Meta-press address these problems?
A: Meta-press addresses the fake news problem because it only searches through sources that are validated by humans. You’re guaranteed to search in real newspapers with articles written by real humans. Humans who were trained and paid for it, which is what we call journalists. This type of information has a price and we should pay that price. If you do not pay for the information you get, than it’s you who is the product sold in the transaction. And you’re not getting the information that you need. So you should pay for the services you use (or run them on your own computer). If you don’t change the way the world is turning, it will continue to turn the wrong way.
1.4. A solution to decentralized indexing
There are multiple projects working on free software search engines to make general purpose search available such as YaCy, Searx and Marginalia. They are addressing a difficult problem: how to make a distributed index of the world that is reliable and honest. With Meta-Press we addressed this by limiting our scope to online newspapers. Newspapers provide honest indexes, because their reputation is at stake. With Meta-Press we just stitched those indexes together. This is a small window to the rest of the world with only information created by journalists. But hopefully information that covers the entire world because journalists are looking everywhere.
General purpose indexes are a complicated problem and I am better at solving simple problems. When I started the project it was small compared to Google. But I decided I would pull one hair out of the head of Google. I invite you to do the same. Get your own hair off the head of Google and we will win quite fast. It is my collibri approach to this thing. You address a small part of the problem but you address it well.
1.5. The future of search should be collaborative
Q: What do you think will happen with search in the future?
A: ChatGPT could be the end of Google Search in a year if it continues to rise like this. Not because the results are better. ChatGPT is a stochastic parrot. It has no idea what it says. You can ask it what the French presidents of the Republic are and you get the right list. Ask it who the female presidents are and you’ll also get a list, while no woman ever been French PR yet. Despite that ChatGPT works because it is simple for people to use. That is always what wins (the simplest, not the best).
ChatGPT works with so-called AI algorithms and fortunately for us they are not the same kind of AI as in the Matrix or Terminators movies. It’s just a statistic matrix and this technology, which is 40 years old, can be used for the good of humanity. For instance you have Pl@ntNet. You send it the picture of a plant and it will tell you what plant it looks like. This is access to knowledge. This is the same technology but used in a good way. This is search made well. iNaturalist.org helps you find out which insect you are seeing. BirdNET from Cornell University will tell you what bird you are hearing.
These are examples of online search tools I am the most excited about. They are collaborative efforts. The more you use it, the more accurate it will get. Take Pl@ntNet for instance, if you upload a picture of a rare plant (something that is missing in their database) they will display a pop-up inviting you to upload more photo’s of it once it will be flowering or when the seeds will be fully grown to improve the database. That way humanity works together to get better knowledge of what surrounds us. That looks like the way to go for me. Going to Mars is not an interesting thing to do as long as we don’t have maps of the ocean floors.
1.6. We mustn’t rely on search engines to know the world
Q: Is that how you would like to see search evolve in the future? Becoming more collaborative?
A: Yes. There are two sides of the problem. Is it search that has to be improved? Or is it the way we publish things? Searching for something in a library is easy because it is an organized world. You have shelves, you have books in alphabetical order, that works great. If we publish things better we won’t have a problem searching for them. But hoping that search engines will free us from the efforts to know the world and to sort it, is a bad way to go. Something that won’t work and will catch people in the glue like small birds.
Keep in mind the limits of the dream sold to us nowadays: artificial intelligence is just another algorithm to sort things automatically and we have absolutely no idea of what the content is. Tools won’t solve society’s problems. If humans work together it will be better.
Collaborative efforts to discover the world, to map it, like OpenStreetMap is the way to go. People working together and giving each other knowledge that will help to know the world and to search through it. For instance, you can help Meta-Press mapping the newspapers of the world. It is a collaborative effort. The project is open to all your contributions. Help us to map the world! And thanks to NLnet it will soon be possible for people who aren’t computer science engineers to do this, I promise.
1.7. Ways to get involved with Meta-Press
Q: How can we contribute to Meta-Press?
A: You can add your own sources to Meta-Press. On the Meta-Press.es website you find much documentation on how Meta-Press sees the world and how it can read your favorite newspapers. It’s currently a long process available for someone who knows how to make a CSS selector. But with funding from NLnet I am working on an interface in which you will just have to copy paste the address of the source and click to point where is the search engine, where is the title of the results, where is the link, the date and it will be enough.
Once you have this newspaper definition it’s recorded somewhere in Meta-Press and you have a button in the settings to manage your local sources. You will have the JSON object (text) describing how to fetch the results from this source. You can send it by mail to Meta-Press or you can create a pull request on FramaGit.org. You can also reach us via IRC or Mastodon.
Adding sources is the contribution you can do. But every kind of help to the project would be welcome. You can help with translations (which are managed with Weblate.org, a great free software project and enterprise). You can also help the project by just speaking about it. Introduce it to people or to your local university. There are a lot of configurations possible in Meta-Press to search for specific topics or in one language or country (among 75 of them currently). Just try it, use it and make it known.