I tend to use DuckDuckGo most of the time, but its quality has collapsed considerably. Google is terrible now and only worth it for reverse image search. I haven't tried Brave much and SearX-NG was a pain to set up for being a de-facto Google and Bing wrapper. There's Kagi but I don't really have $120/yr for a search engine, nor do I want to deal with how they handle logins and session management.

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

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?