I wouldn’t call it a shit show, they were first in E2E encrypted collaborative messaging platforms and get all the hits. Moderation turned out to be incompatible with E2E encryption. People kicked out of channels can still decrypt because keys revocation hasn’t been made.

Matrix server is a resource hog and a mess, and the protocol has never took off to spark alternative implementations.

Encryption doesn’t scale well for big channels.

They got many things right, but got crippled by many things they couldn’t possibly predict.

I recommend reading this publication from Denys (Radis) Prokopenko — Commander of Azov Army Corps.

https://www.pravda.com.ua/eng/columns/2025/07/14/7521623/

The question is not when World War III will start. The question is whether we are capable of recognising its course and structure before it reaches its decisive phase.

Anthropic or not, the recent streak of supply chain and also targeted attacks got me packing into VM.

Now I run all devtools and install packages only inside the VM and either connect Zed / VSCode via ssh or run terminal ide inside the VM.

No host mounts whatsoever, all communication Host to Guest strictly through git.

I have a few examples of people replaced by AI.

Photoshop slop cheap ads on the streets and TV are mostly AI slop now.

Content creators, publishers, media, music industries are all heavily affected.

The company I know paid $2K for wireframe UX design instead of standard $10K UX + UI. UI part was done by AI ($8K in human labor was replaced by AI)

I know a few companies stuck with ancient software that have been paying salaries to generations of programmers, this software is being replaced by AI, quite fast too.

I know positions which are in essence “receive a fax and type it into the system” right now being replaced by AI.

This is neither pro- nor anti- AI stance. Saying that a chatbot will ultimately replace humans is lying — they need to justify trillions in valuations.

Saying that AI is not severely changing our society is ignorance and head in the sand tactics.

The majority of Internet traffic are bots. I’m terrified of the upcoming elections with all the deepfakes and bot campaigns.

People should not reinvent the wheel. FOSS Android has already solved many problems which Linux always will have, by design.

Non-preemtive scheduler which doesn't take app usage into account will drain battery fast. Non-managed app platform is always a security risk.

The absence of a single app framework will always lead to a zoo. Any average Gnome installation bring all KDE and vise versa because different apps use different GUI and Desktop convenience libraries.

Google is trying to kill Android by moving more and more core functionality from OSS into proprietary services.

We need to fight back through forks and legislations. Android has started as a Linux phone 15 years ago? And has already solved a ton of inherent problems.

No, this is have enough canned food scenario.

If the money would wipe out from the economy, that is less money for everyone. First round of bankruptcies will also wipe out contractors and debt issuers of the bancrupted companies, and so on. Mass layoffs. People cannot afford restaurants, or car washes, more businesses out, more mass layoffs.

In 2001 tech bubble the money haven't been lost. Virtual valuation dropped sharply. But the economy recovered in a year.

This is like 2008, money have been poured into concrete and silicon chips. They were spend, gone for good. It will take a few good years to recover if the bubble would burst.

I'm missing artisanal, carefully crafted software from before LLM times. Microsoft Teams, Eclipse, Elastic, Lotus Notes even, and every website, well-thought and carefully optimized, taking its 30 second to load carefully picked handcrafted ads with broken CSS and color palette causing epileptic seizures. But sure.

Software has been horrible for the last 20 years. Photoshop slop was replaced by AI slop. LLMs have replaced software bootcamps.
Here's Tom with the weather.

Mainstream phone as in "I don't need to debug it via terminal"

The issue I'm pointing at: safe, long battery life, snappy maps while driving is what took AOSP more than 5 years.

It would be very unfortunate to discard all that work and start from scratch.

I made a full comment in this thread. The bottomline is

1. Sandboxing of resources both hardware (gyroscope, network devices) and data (photos, music) takes a lot of trial and error.
2. There is a need for an ecosystem: i.e. apps sould be able to create calendar events, or access shared mediaplayer — also with permissions
3. Developers need to adapt to the software ecosystem
4. Hardware companies e.g. smart watch, projectors, TV need to adapt

It all takes years.

Linux phones are around for enthusiasts since Nokia N900 (which was/is a masterpiece) — yet nothing is remotely close to a mainstream phone.

It took 6-10 years for Android to take shape.

On Linux, every app has full access to your browsing history, clipboard (passwords), photos with geo-tags, music, list of other installed apps, contacts. Unrestricted battery and network access -- it's a tracking paradise. And all it takes is one supply chain attack on npm install with typical 4000K dependency packages

I missed that part

The time from opening the browser to having a fully loaded site is minutes.

I think it depends on the region. To me, full browser restart with reconnect is maybe 10 seconds tops, usually less. I use Tor Browser as a default one on my phone, and it opens random links quite okay.

For me, the main issue is exit node blocking, then I need to restart the browser 1-2 times.

This question is unironically very deep. As it's privacy we're talking, you decide what to trust on your own.

My understanding is that Tor provides anonymity for my threat model (ad-tech corporations).

But trust need to be placed somewhere. Do we trust Mozilla? All their emploees? Do we trust OSS? Does anybody actually review open-source code? What about supply chain attacks?

I am, a nobody, was personally invited to a Contagious Interview (a person, pretending to be a client for consulting was trying to place a rootkit on my machine via GitHub repo).

What about AI-assistet coding that actively tries to eliminate security gates?