Soooo after a month of trying, I have given up trying to get my 7900xtx setup back from my ex.

I have been grinding at work using a ryzen 5500U handheld with a smashed display and a usb-c dock as my stopgap system.

This is what I have settled on as the replacement:

https://pcpartpicker.com/list/GGrhwY

My primary usecase is g*ming, although my most played titles are HEAVILY cpu bound to the point that an Arc B580 would do in the counterfactual world where Intel did not operate factories on destroyed Palestinian villages.

The 9070xt is a gambit that ML hype keeps prices inflated through the launch of UDNA in 3 years.

The article is a great critique of how what the author refers to as the "Efficiency Lobby" has been pursuing a narrow idea of task oriented intelligence focused on productivity. It's a narrow focus, driven by corporate interests, that necessarily leads to individualistic consumption of AI services, hindering genuine creativity, open-ended exploration, and collection.

A recent paper introduces MemOS with the potential to create a truly collaborative and community driven foundation for AI. The paper introduces a new approach to memory management for LLMs, treating memory as a governable system resource.

It uses the concept of MemCubes that encapsulate both semantic content and critical metadata like provenance and versioning. MemCubes are designed to be composed, migrated, and fused over time, unifying three distinct memory types: plaintext, activation, and parameter memories.

This architecture directly addresses the limitations of stateless LLMs, enabling long-context reasoning, continual personalization, and knowledge consistency. The paper proposes a mem-training paradigm, where knowledge evolves continuously through explicit, controllable memory units, blurring the lines between training and deployment paving the way to extend data parallelism to a distributed intelligence ecosystem.

It would be possible to build a decentralized network where there's a common pool of MemCubes acting as shareable and composable containers of memory, akin to a BitTorrent for knowledge. Users could contribute their own memory artifacts such as structured notes, refined prompts, learned patterns, or even "parameter patches" encoding specialized skills that are encapsulated within MemCubes.

Using a common infrastructure would allow anyone to share, remix, and reuse these building blocks in all kinds of ways. Such an architecture would directly address Morozov's critique of privatized "stonefields" of knowledge, instead creating a truly public digital commons.

This distributed platform could effectively amortize computation across the network, similar to projects like SETI@home. Instead of constantly recomputing information, users could build out a local cache of MemCubes relevant to their context from the shared pool. If a particular piece of knowledge or a specific reasoning pattern has already been encoded and optimized within a MemCube by another user, it can simply be reused, dramatically reducing redundant computation and accelerating inference.

The inherent reusability and composability of MemCubes make it possible to have a collaborative environment where all users contribute to and benefit from each other. Efforts like Petals, which already facilitate distributed inference of large models, could be extended to leverage MemOS to share dynamic and composable memory.

This has the potential to transform AI from a tool for isolated consumption to a medium for collective creation. Users would be free to mess about with readily available knowledge blocks, discovering emergent purposes and stumbling on novel solutions.

Two Chinese satellites have rendezvoused with one another more than 20,000 miles above the Earth in what analysts believe is the first high-altitude attempt at orbital refueling.

I am trying to get a local org I am in set up with a domain and website just to have a place to point people for everything. We would like to keep it as cheap as possible. I figure we need the following:

-Domain name (going to use namecheap probably)

-VPS host (I haven't done this before, it looks like racknerd may be way to go?). I assume I will probably only need 1GB of memory as it will just be a static webserver but that may be too little, not 100% sure.

-Email host. This is one of two real reasons I want to own the domain, we have multiple uses for email but currently everything is under one gmail address and a lot gets lost in the clutter. A few people in our org would like to stick with gmail but I am open to other suggestions. Definitely do not want to deal with self hosting on this.

-Website builder. I plan to use an Ubuntu server with the LEMP stack on the VPS, should I just use Wordpress? I am definitely not experienced in website building so it's not realistic to do my own HTTP. My only concern is using Wordpress will result in a poorly optimized site that may strain my limited resources, but there are also a few people in our org that have experience with it so that would help.

While I have a decent amount of tech experience generally, these are mostly uncharted waters for me. I know this comes across as kind of half baked, but really I am just looking for general advice!

https://github.com/MemTensor/MemOS

For someone who says she is fighting AI bot scrapers just in her free time, Xe Iaso seems to be putting up an impressive fight. Since she launched it in January, Anubis, a "program is designed to help protect the small internet from the endless storm of requests that flood in from AI companies," has been downloaded nearly 200,000 times, and is being used by notable organizations including GNOME, the popular open-source desktop environment for Linux, FFmpeg, the open-source software project for handling video and other media, and UNESCO, the United Nations organization for educations, science, and culture. 

Iaso decided to develop Anubis after discovering that her own Git server was struggling with AI scrapers, bots that crawl the web hoovering up anything that can be used for the training data that power AI models. Like many libraries, archives, and other small organizations, Iaso discovered her Git server was getting slammed only when it stopped working.  

"I wasn't able to load it in my browser. I thought, huh, that's strange," Iaso told me on a call. "So I looked at the logs and I figured out that it's restarted about 500 times in the last two days. So I looked in the access logs and I saw that [an] Amazon [bot] was clicking on every single link."

Iaso knew it was an Amazon bot because it self identified as such. She said she considered withdrawing the Git server from the open web but that because she wants to keep some of the source code hosted there open to the public, she tried to stop the Amazon bot instead. 

"I tried some things that I can't admit in a recorded environment. None of them worked. So I had a bad idea," she said. "I implemented some code. I put it up on GitHub in an experimental project dumping ground, and then the GNOME desktop environment started using it as a Hail Mary. And that's about when I knew that I had something on my hands."

There are several ways people and organizations are trying to stop bots at the moment. Historically, robots.txt, a file sites could use to tell automated tools not to scrape, was a respected and sufficient norm for this purpose, but since the generative AI boom, major AI companies as well as less established companies and even individuals, often ignored it. CAPTCHAs, the little tests users take to prove they're not a robot, aren't great, Iaso said, because some AI bot scrapers have CAPTCHA solvers built in. Some developers have created "infinite mazes" that send AI bot scrapers from useless link to useless link, diverting them from the actual sites humans use and wasting their time. Cloudflare, the ubiquitous internet infrastructure company, has created a similar "AI labyrinth" feature to trap bots. 

Iaso, who said she deals with some generative AI at her day job, told me that "from what I have learned, poisoning datasets doesn't work. It makes you feel good, but it ends up using more compute than you end up saving. I don't know the polite way to say this, but if you piss in an ocean, the ocean does not turn into piss."

In other words, Iaso thinks that it might be fun to mess with the AI bots that are trying to mess with the internet, but in many cases it's not practical to send them on these wild goose chases because it requires resources Cloudflare might have, but small organizations and individuals don't. 

"Anubis is an uncaptcha," Iaso explains on her site. "It uses features of your browser to automate a lot of the work that a CAPTCHA would, and right now the main implementation is by having it run a bunch of cryptographic math with JavaScript to prove that you can run JavaScript in a way that can be validated on the server."

Essentially, Anubis verifies that any visitor to a site is a human using a browser as opposed to a bot. One of the ways it does this is by making the browser do a type of cryptographic math with JavaScript or other subtle checks that browsers do by default but bots have to be explicitly programmed to do. This check is invisible to the user, and most browsers since 2022 are able to complete this test. In theory, bot scrapers could pretend to be users with browsers as well, but the additional computational cost of doing so on the scale of scraping the entire internet would be huge. This way, Anubis creates a computational cost that is prohibitively expensive for AI scrapers that are hitting millions and millions of sites, but marginal for an individual user who is just using the internet like a human. 

Anubis is free, open source, lightweight, can be self-hosted, and can be implemented almost anywhere. It also appears to be a pretty good solution for what we've repeatedly reported is a widespread problem across the internet, which helps explain its popularity. But Iaso is still putting a lot of work into improving it and adding features. She told me she's working on a non cryptographic challenge so it taxes users' CPUs less, and also thinking about a version that doesn't require JavaScript, which some privacy-minded disable in their browsers. 

The biggest challenge in developing Anubis, Iaso said, is finding the balance. 

"The balance between figuring out how to block things without people being blocked, without affecting too many people with false positives," she said. "And also making sure that the people running the bots can't figure out what pattern they're hitting, while also letting people that are caught in the web be able to figure out what pattern they're hitting, so that they can contact the organization and get help. So that's like, you know, the standard, impossible scenario."

Iaso has a Patreon and is also supported by sponsors on Github who use Anubis, but she said she still doesn't have enough financial support to develop it full time. She said that if she had the funding, she'd also hire one of the main contributors to the project. Ultimately, Anubis will always need more work because it is a never ending cat and mouse game between AI bot scrapers and the people trying to stop them. 

Iaso said she thinks AI companies follow her work, and that if they really want to stop her and Anubis they just need to distract her. 

"If you are working at an AI company, here's how you can sabotage Anubis development as easily and quickly as possible," she wrote on her site. "So first is quit your job, second is work for Square Enix, and third is make absolute banger stuff for Final Fantasy XIV. That's how you can sabotage this the best."

I'm watching all this for 1450 bucks (e: after taxes):

• AMD Ryzen 7 9700X
• AMD Radeon RX 7700XT 12GB
• MSI B650 motherboard
• Corsair 32GB DDR5 6000MHz
• Corsair RMe 850W
• 2TB NVMe
• 360mm 320W liquid cooling

Is this a coherent build?