Freenet is also based on middle-out compression.

No. Trying to push 30TB into Freenet will be interpreted as an attack on the network by other peers and they'll disconnect from your peer.

Freenet is better viewed as a communication medium than a storage medium. Data is prioritized according to demand, somewhat like an LRU cache. You can't upload 30TB and expect the network to preserve it for you.

On liability, running a Freenet peer is more like running network infrastructure that automatically routes and caches other people's traffic than intentionally publishing it. US law explicitly recognizes this distinction in the DMCA's provisions for transitory network communications and system caching. We discuss this in more detail in the FAQ.

You obviously haven't tried Freenet. You can join our room on River and you'll see no spam, no fraud, no hate speech, none of that.

Decentralized doesn't mean unmoderated. Systems built on Freenet can be moderated however their creators and communities want. River is moderated.

The difference is that nobody running one service gets to decide what everyone else is allowed to run.

1) 30B, dense. Optimized to fit Q4 quantized in a 24GB card with a speculative decoding model as well.

2) Because reporters don't know what weights are and assume you don't know either.

3) Way better than Gemma 4 on text tasks, slightly better on multimodal. Numbers below are all: Benchmark: Muse Glimmer score Gemma 4 31B score difference

Artificial Analysis Intelligence Index: Muse Glimmer: 35 30 +5
MCP Atlas (Public): 75.5 54.2 +21.3
DeepSearch QA: 74.6 61.7 +12.9
SWE-Bench Pro: 51.2 36.9 +14.3
SWE-Bench Verified: 76.0 66.6 +9.4
OSWorld-Verified: 65.9 58.5 +7.4
GAIA2: 43.3 36.4 +6.9
WildClawBench: 47.6 37.6 +10.0
TerminalBench 2.1: 51.7 43.4 +8.3
Tau3-Banking: 23.5 15.1 +8.4
MMMU Pro: 74.0 73.0 +1.0
Charxiv Reasoning: 78.8 77.7 +1.1
OmniDocBench v1.5: 75.8 72.5 +3.3
ScreenSpot Pro: 75.4 75.9 -0.4

4) 128k tokens

5) No, sadly.

What you want is a DGX Spark.

It costs $4000.

And yes, there are fundamental advantages to cloud services, such as large-scale batching, little idle downtime, high speed, and hardware optimized to the specific models / serving needs. That said, one can weigh that off against sovereign control over your server...

Define "tolerable".

The model in question - Muse Glimmer - with DFlash/speculative decoding - will probably get you ~60 to 124 tok/s on a 3090 (a quite dated GPU). For a 5090, it's said to clock in at 233,4 tok/s. On CPU you're looking at maybe 3-5 tok/s.

If you call that "tolerable", I guess you're more patient than me? And as mentioned, you're not just wasting time, but also wasting a lot of power too - CPU is a very power-inefficient way to run ML models.

If you insist on CPU, this isn't the right kind of model anyway. You want to take advantage of the fact that you probably have lots of (comparably cheap) RAM, and compensate for the fact that you have (comparably) terrible memory bandwidth, and for that, you want a MoE with a high total parameters but a low active parameters. Not a dense model like this.

Then I was basically aiming at Mac Minis

That's very much a special case which you didn't mention in your post that I responded to, but still the answer is "meh". You couldn't run it at all on a 16GB Mac Mini, and I think you'd struggle to run it at all on a 24GB (because you have to share the ram with the OS, the inference server, etc). For the base Mini you might get 10-12 tok/s, and for the M2 Pro / M4 Pro, maybe 25-35 tok/s. Still pretty far from a GPU, though.

This model is designed for >= 24GB GPUs.

MoEs don't save RAM (for a given quality), they increase it. You have to store all of the parameters in memory, not just the active params. But inference only uses a subset of the total params for each token, so it reduces the memory bandwidth requirements and improves token generation rate. But this comes at the cost of a higher total param count for a given quality.

Meta has always been releasing open models (Llama was famously the first powerful open model). The change has actually been in the opposite direction, with Muse Spark *not* being open.

Zuck descrbed his motivation way back when, about how they got burned with Facebook on app stores, in that Apple and Google could basically bully them however they wanted, on whatever extractive terms they wanted, and there was nothing Meta could do about it. He's now paranoid about "others controlling the platform", and wanted to make sure that doesn't happen with AI, that they have their own AI base to work with.

I mean, if you want your model to be terribly slow and terribly power inefficient, sure, go ahead and run it on CPU.

This really isn't a good model for CPU. For CPU, you want a MoE with a large number of total params but a tiny number of active params. Something like DeepSeek V4 Flash 0731 if you have at least 128GB of RAM - you might get 2-3 tok/s or so on that. The goal is to minimize the memory bandwidth requirements per token, at the cost of a greater total RAM footprint.

For GPU, you're highly VRAM limited but not bandwidth limited, so your best option is generally a dense model (non-MoE) with speculative decoding to make up for the performance limitations.

You can use it fine on 24GB with a good size context with KV cache compression. They've also designed it to use speculative decoding to boost performance instead of a MoE, so that they can pack as much capability as possible into a dense model. Seems like a winner.

Probably LRO vs. Danuri

Except instead of pursuing policies that would help open models at the expense of Big AI, anti-AI people pursue just the opposite - policies that can cripple open source AI but have no realistic ability to meaningfully hurt big AI. It's so short sighted.

Actually, no, what they usually do is much more insidious: fingerprinting rather than watermarking. The fingerprint isn't actually included in the audio, it's included in a database. If they want to tell if the track was generated, they just try to match the fingerprint in the database. Can't filter it out of the track like you can with a watermark.