And, importantly, it's not a decision that's compelled by some sort of hierarchical power structure. Responsible disclosure rules exist to protect everyone who uses the software. Like FOSS it's a system that benefits everyone which makes it worth participating in voluntarily.

Capitalism sometimes gives a cash prize as well, because of course they do.

If your server is running Linux you could run Termux on your phone and copy the files to the server via SCP. You could do it periodically with a script or you could run the command one time before you were going to travel across a suspect border.

Termux has the added advantage of being a generic app, without any specific purpose. There could be a situation where you don't want to have a 'Photo Downloader and Hider 9000' app on your phone.

If you just want something akin to iCloud/Google Photos, Immich is a good project to get started with.

It's modern journalism, only the headline really matters... only a small percentage of people are going to read the article anyway.

It says AI right in the title, that's going to bring all the attention and interaction on social media.

I'm think we're all tied of these puns

This is true, self hosting doesn't save you money. It is, in fact, more expensive and costs your time.

This is because modern technology isn't sold to you as a product, it's sold to you as a loss leader so that you use their products and let them spy on you so that they can sell that data to advertisers or anybody who can complete an ACH transfer.

Your options are

1. Use commercial services and sell your privacy in exchange for cheap consumer electronics which will be obsolete next year either by design or through engineered hardware failure.

2. Setup the services on your own devices and pay the cost in hardware and free time.

In one scenario you spend a bit more money, in the other scenario you don't have privacy. Everyone gets to choose which is more important.

At the top end, they're probably not subsidized... but if you're getting by on a $100 phone, it's likely a $100 piece of spyware that sold your data to pay for 2/3rds of your phone.

Is the bootloader unlockable or is this more locked down spyware pretending to be a tablet?

Boy if I had a nickle for every time that happened...

They're gonna IPO, get a shortcut into the Indexs so that Index Funds have to spend money (your retirement money) buying tons of stock at the near-IPO price, turning fake future profits into actual real US Dollars that the company can use to cash out executives and owners before the company explodes taking the economy with it. Elon already did this with SpaceX minus the economy exploding (so far).

I would not be invested in an index tracking asset which will include the tech or hardware companies involved in the AI supply chain.

I'm sure OPs RTX 3060 will have a tolerable effect on the school children and the rainforest.

Given the comments, I think this is probably the most relevant quote from the project creator (my bolds):

This project was built with AI, and that is part of the point.

I am not a reverse engineer or a security engineer, and I did not come up through whatever field normally teaches this. Almost all of the code here was written by AI coding agents (Claude) that I direct: I decide what to try, think up the angle of attack or the workaround, and the AI implements it quickly and comes back with a result I can check. That loop (an idea, an implementation within the hour, a measurement that says whether it held) is what made a project of this size possible for one person.

What I take from it: the bar for entry has dropped a great deal. The creativity is still yours (which problem to solve, which trick might get around it, when a result is not good enough), but the distance between having the idea and seeing whether it works is now very short.

Two rules kept it honest: every claim has to come from a test or a measurement (this README says what was measured and what was not), and nothing derived from the game goes into the repository. It is a study project; expect rough edges, and mistakes that a specialist would not have made.

Do vulnerabilities in voting software typically get disclosed to the public? Is there a CVE list somewhere that I can browse?

Anti-circumvention laws already exist and carry penalties including prison, it only takes one judge to create case law that this kind of reverse engineering is circumvention and then it'll disappear from the commercial Internet.

The safest way to go is to create the software from scratch, using reversed engineered proprietary software to make a different, but very similar, product is a legal minefield and largely unnecessary given that the same AI time that goes into reverse engineering the project could be directed towards making a clean room replacement which could completely avoid those legal issues while still replacing the product with a FOSS product.

Exactly, this is just the Internet Cafe business model. You pay for time to sit at a good PC, the business buys the hardware and pays the electricity.

If you want to avoid Microsoft but still want the ability to remotely stream to your low-end devices. You can setup Sunshine and Moonlight and use your own hardware to stream. There's essentially zero latency locally and comparable latency when streaming remotely.

It does require a tiny bit of DIY and some research/AI usage/YouTube video watching but it's a free way to make your own personal cloud streaming service if you're already gaming.

(It's also easier if you use Linux, yet another way to screw Microsoft out of your money and data.)

It depends on how complex you want to make it.

The simplest way would be to use a GUI tool like LM Studio, it has the benefit of selecting the model quants that are likely to work on your hardware and having sane defaults for the various parameters that you can work from.

Typically you'll want to increase the context window, the default is good for short one-off questions but if you're going to do RAG or coding tasks you will need a much larger context window. You will typically find this through trial and error, increasing until you hit memory limits or notice a large slowdown on tokens/second.

You can also have the model respond to an OpenAI-like API interface on a designated port/interface if you want to run non-chat tools through the same model.

Then, depending on the task that you're trying to solve, you may want to find an MCP server built for that task which can provide the ability for the AI to interact with tools in a way that's more controlled (giving an AI raw terminal access on a user's machine is typically a bad idea).

If you need something more flexible that you can incorporate into scripts and automation, llama.cpp has been popular for a while. Like Linux distros, everyone is going to have an opinion on the best and whatever you decide on will be the wrong choice. There's also Ollama and vLLM if you want to run a datacenter.

Model-wise it completely depends on the task. Do you need a multi-modal model that supports vision, is it going to be coding, is it trained to use tools/MCP servers, does it use reasoning, etc. The Qwen models are open weight and available in many different quants. The version 3.5 models run well on consumer hardware and can use vision, reasoning and tools.

I'm running it through Llama.cpp but it will likely be available via LM Studio as well.

The paper seems to argue unerringly in favor of the boss’s financial interests

To add to this, the SM64 project took a team of people 18 months.

The project in the OP was created on 12 October 2026, 9 days ago.