Sem

And that is very strange for me. Under the news about AI-search from OpenAI I asked about comparison with AI-search from Kagi, but got a lot of downvotes and a link to the post where the author express his negative opinion about AI.

My question was about benefits of openai search compared to kagi search. How is this link related to my question?

What are benefits of using openai search when we already have Kagi that is much more privacy a friendly, is ads free and provides the same functionality?

What do you think, how long should we wait until the overflow in the debts database makes the value negative?

It supports via GMS sandbox. So, I can install google camera, maps, bank app, insta360 app, an app for my bike computer, etc. But in that case I prefer to use Proton VPN that hides my real IP from all these apps and also block some tracker endpoints.

I have Pixel with GrapheneOS and I tried most of FOSS camera apps, but all of them are still far behind the GCam. I hope one day there will be a good replacement, but not today.

I’m using Organic / OsmAnd for most of use cases and daily navigation. But if you need to find a specific office, shop, food or ATM nearby you still need GMap from time to time…

And for the latter problem, don’t use such apps except in closed environments or without internet access.

While that is a right answer, I do not want to avoid such apps because I need them. I need my mobile bank app, I need google camera, sometimes I need Google maps, etc. For me using VPN to hide my real IP from greedy apps and to hide DNS requests from the cracked public WiFi is still a good tradeoff between security, privacy and my own user experience.

But most spots don’t have the resources or expertise to analyze and sell or otherwise misuse your logs.

Most spots don’t have also the resources or expertise to secure their own spot. As I remember, cheap routers used in public places may contain a lot of vulnerabilities.

encrypted DNS

Will it help me if I’m using LbreTorrent do download piracy content on my phone? Or how it would help me to hide my location from mobile apps that extract location from IP?

As I can understand from EFF articles, the problem is not only in MAGA guys and their conspiracy but in KOSA proposal itself.

https://www.eff.org/deeplinks/2024/09/new-us-house-version-kosa-doesnt-fix-its-biggest-problems

If I already use Pandas for processing my data in-memory, CoW can significantly improve the performance. That was my point.

If you are doing data processing in pandas CoW allows to avoid of a lot of redundant computations on intermediate steps. Before CoW any data processing in Pandas required manual and careful working with code to avoid the case described in the blog post. To be honest I cannot imagine the case of offloading each result of each operation in the pipeline to the storage…

deleted by creator

The title is very click bait imo. It is not about any private data. It is a very specific case of deleted fork of the public repository. It is a bug, of course. But it doesn’t look so serious as I was thinking when saw the title.

There is also a “web-version” of org-roam named “orgnote”: https://github.com/Artawower/orgnote

I did not see in article any proof that Chineese scientists have a bad intentions. A statement about “they always do it [because they are evil maoists]” is not a proof for me.

I did not see in the article any arguments, except the fact that Chinese researchers are in some US blacklists. Only proclamations about EU should restrict, EU should prevent, etc. But without any arguments why. Maybe lack of actual arguments is the reason why the article is so short?

I mean if you try to use anything like python packages or even try to build python from sources it is painful. The only way to create developer environment is to use something like nex develop shells and you need to care about passing to LD_LIBRARY_PATH all that you need. And nothing downloaded as a binary is not working… For example, if I’m working on a Java-maven project that includes maven-protobuf then it is not working for you because protoc binary for manylinux is made for a dynamic linkage… Overall developer experience is painful. And anything that is not in nixpkgs you cannot just download, build from source and use: you need to pack everything into packages with resolving all the dependencies by hands…

I had an yearly experience with Nix, but I’m thinking that it is overhead for just a home PC system. You may have more pain with static linkage compared to benefits of Nix reproducability and flexibility. Now is a year I’m on the Fedora Silverblue and this one is a really good balance between complexity and usability.