KolmogorovComp

Nothing constructive to say, besides that the video really shows we're entering into a Sci-fi era.

It’s satire.

I assume they shamelessly were talking in Fahrenheit degrees.

Doesn't the implementation being in rust for many browser (`temporal_rs`) make it possibly slower than it could be in pure JS? Calendar is not very intensive process, so I would not be surprised if the slowness of boundary passing make it slower.

They’re only popular because they appeal to the real luxury product, like counterfeit Rolexes.

Yeah it really felt hammered in as a way for the author to try to look literate and punchy, but it does not work.

I've been subconsciously doing this forever and after hundreds if no thousands of interaction have lately been assessing its outcome as more neutral than the article frames it. On the long run while mildly pleasant, you realise how shallow these conversations are and they have not brought anything valuable to your life. Maybe stopping for a while would lead to further realisation but I don't think so.

The most positive effect it has had on me is to make me enjoy even more deep conversations with my friends.

Was I the only stunned by the quality of the video ad? It’s really advert at its peak, to the point yet artistic.

Thank you for the insights.

Same, on Safari

It’s a very well known pattern, as someone else mentioned it’s used in CPP in smallstring, Rust smallvec, C usually hand rolled etc.

How can you seriously pretend to do any study with only eight people involved? Especially when your company is worth billion. It just calls for bad press and criticism of amateurism.

I’ve been using DDG for close to ten years and only use !g about once a year or less. What are your typical queries? I mostly search CS-related (though less nowadays with LLM) or simple stuff.

> Git repositories also typically don't store full copies of all versions of a file that have ever existed either; they're incremental

This is wrong. Git does store full copies.

This! The comment I was angrily about to write.

Write access to public repos is not optional?? As well as full read and write user data?? Flee, far.

Unfortunately will not make it to the frontpage but very cool

The article grossly underestimates the energy needed to be stored. You cannot fly an aircraft planning to use 100% of its ‘fuel’. You need alternate(s) and final reserve.

Whoch likely mean doubling the required size of the battery pack, making it infeasible.

> Fuel consumption in flight reduces landing weight. Battery weighs the same landing and take off. Harder landing costs

It’s not that much in terms of bodyweight on a regional jet compared to a long haul.

I wonder why it’s an issue for readers. Has anyone of you read all past great sci-fi books?

I assume they plan/hope to sell much more than 500 units, and/or the profit margin is much higher with China manufacturing.

not more terrifying than sitting in any turboprop airplane.

It’s a static page, there is no backend. That’s why a static list, for example computed every day via an action is the way to go.

It would not make much sense to compute it, if you do not subtract the karma earned through submissions.

> It just seems like these kind of design choices needlessly add a barrier to entry for people who want to climb aboard.

Any experienced programmer knows syntax is a detail, and would not be deterred by that.

yes, but that's not what happen here. this part of the AGPL is there to avoid people adding more restrictions, but here mattermost is loosening up the restrictions.

> > We promise that we will not enforce the copyleft provisions in AGPL v3.0 against you if your application ... [set of conditions]

Why are rclone/rsync never used by default for app updates? Especially games with large assets.

> instead of creating e.g. some kind of a path of interest.

You need to like them to update the weights of the algo, it works well

To save everyone a click

> The 1.8-bit (UD-TQ1_0) quant will run on a single 24GB GPU if you offload all MoE layers to system RAM (or a fast SSD). With ~256GB RAM, expect ~10 tokens/s. The full Kimi K2.5 model is 630GB and typically requires at least 4× H200 GPUs. If the model fits, you will get >40 tokens/s when using a B200. To run the model in near full precision, you can use the 4-bit or 5-bit quants. You can use any higher just to be safe. For strong performance, aim for >240GB of unified memory (or combined RAM+VRAM) to reach 10+ tokens/s. If you’re below that, it'll work but speed will drop (llama.cpp can still run via mmap/disk offload) and may fall from ~10 tokens/s to <2 token/s. We recommend UD-Q2_K_XL (375GB) as a good size/quality balance. Best rule of thumb: RAM+VRAM ≈ the quant size; otherwise it’ll still work, just slower due to offloading.

It’s really a shame git storage use files as the unit for storage. That’s what makes it improper for usage with many of small files, or large files.

Content-based chunking like Xethub uses really should become the default. It’s not like it’s new either, rsync is based on it.

https://huggingface.co/blog/xethub-joins-hf