I understand, but the shift in user behaviour is significant and I think websites are not taking it into account. If the users move more and more to AI, and since Google introduced AI mode it’s only a question of time until it becomes the default, we will see more and more of what we thing are AI crawlers and less and less organic users.

AI seems to be the new middleman between you and the user, and if you block the middleman, you block the user. For people with hobby websites or established sites it may make sense because people either know of them, or getting more exposure is not a wish or requirement, but for everyone else, it will be painful.

I just realized an interesting thing - if I use Gemini, and tell it to do deep research, it actually goes to the websites it knows/finds, and looks up the content to provide up-to-date answers. So, some of those AI crawlers are actually not crawlers, but actual users who just use AI instead of coming directly to the site.

Soo… blocking AI completely could also potentially reduce exposure, especially as more and more people use AI to basically do searches instead of browsing themselves. That would also explain the amount of requests daily - could be simply different users using AI to research for some topic.

Point is, you should evaluate if the AI requests are just proxies of real users, and blocking AI blocks real users from knowing your site exists.

Anubis is the name of the tool. Also, Cloudflare just announced they have something against AI scrapers.

I share a Spotify family plan with friends, but I use Zotify to make backups, which I then host in Jellyfin.

Some things would, but not everything, so at best case, I would need to run 2 computers instead of one. I have a beefy spare M1 MacBook Pro, uses almost the same as the Raspberry but it’s horrible for selfhosting.

I thought the hidden cost is my power bill by having a PC run 24/7…

I wanted something that has OnlyOffice integration and basically an selfhosted Google Drive, so tried Nexcloud as the most popular solution… but… it was a pain to set up, its internal workings make handling reverse proxying a pain and it feels extremely slow.

I will try out Seafile, they seem to have just the 2 things I need and nothing more.

If you want cheap and simple - Time4VPS. I am with them for a couple of years and for the price there is nothing really to complain about. If you take 2y contract, the cheapest plans are like €2.5/month.

If you do decide to maybe go with them, it would be appreciated if you could use my affiliate link, otherwise… just type Time4VPS into google. They are based in Lithuania.

People will pay for solutions to their problems, and most people and companies don’t seem to want to hear - that we have the problem of AI being in everything.

The next BIG THING will have a single marketing label - no AI inside.

Actually, I need to update my github repos with “No AI inside” labels, stickers, etc. Might bring in more visibility.

Of course, you just need to make sure your power supply, however it is set up, needs to provide constant power over long term that the system needs at minimum. That means for ex. if you use solar, you need to have a big enough battery and array so that it can charge batteries throughout the day even when it’s cloudy so the system can be up overnight when there is no power generation happening. If you have for ex. a raspberry pi and a smartphone as a hotspot, you can probably have enough energy with an average UPS and one solar panel to make sure it will never go fully empty overnight, regardless how cloudy the day was.

Setup fail2ban

I cancelled my subscription because of technical reasons, but now I can add political reasons on top.

Asking people to do something will never work, telling people how something is better will trigger their curiosity to at least take a look.

It’s literally a fork. Open some files, they are exactly the same. Someone just exported easy notes and then uploaded it again with different icon and name so people won’t immediately see it’s a fork.

It’s open source standard, anyone can contribute…

Unless you can follow specific people on blocked instances, this is a fail. If my friend is on another instance which is blocked from my instance… whats the point of the fediverse? Might aswell go back to Facebook or X/Twitter. They are shitty but at least I can see my friends.

I don’t know or need to know the handle. I know my friends name and surname and that must be more than enough. Facebook doesn’t need “@facebook” and twitter doesn’t need “@twitter” to find people if they exist there. I know the feature is coming but it is the key to make it accessible to wide range of average Joes who don’t want to, in their own vision, be rocket scientists to find people on the fediverse. It needs to be as simple as on facebook or other networks.

It’s not dead but it has one big and massive issue that prevents mass adoption - discovery. If I can’t just write the name of my friends in search and find them no matter where they made their account - for an ordinary user, or one that comes from centralized services, this seems extremely alien and hostile.

And in the end, if you can’t find your friends, you want to interact with, what is the point of using the service?

Luckily, Mastodon is working on a discorvery protocol that should offer a way to find people across the board, which will hopefully make the Fediverse “appear” centralized to the average Joe while maintaining all the benefits of decentralization to the advanced users.

Its easier to just sign up and find everyone immediately, than to go learn what are instances and which one should you choose to make an account on, and then go and learn how to find other people that are not on that instance, or how to check do they have a mastodon account at all, then go and learn how to XY.

The “go and learn” is something that people, most of them, just don’t want to do. If you need to learn how to use something, this is the first indicator of a bad user experience. It should be obviously easy for a new person to get around.

I think there is a possibility to have a local client using Gmail, with a rule set to delete mails on Gmail as soon as they are synced locally. So in this case Gmail would be nothing more than a relay/proxy, while your mails would actually be offline.