They shouldn't be trying such a desperate solution without trying chlorox and ivermectin first.

It was an LLM that chose the target.

Yes, the data was outdated. The LLM still chose the target.

Yes, it was a human who ultimately made the decision to launch the attack. They trusted the LLM, which chose the target.

I'm not saying it would not have happened if not for the LLM, but if the LLM wasn't involved there would have been more humans involved, and less trust in the data, and maybe someone would have caught the error in time.

Regardless, my point is that they are already using AI to make decisions about who to murder. That is a fact even the Pentagon admits, and Peter Thiel brags about.
=Smidge=

Just don't leave the country after you participate.

For that matter, you may suffer remote retaliation even if you stay home. I would leave this to the military or other government branch.

Also there is the question of legality, government operation or not. It sounds to me like vigilante justice. And countries who like their anti-American cybercriminals will certainly treat it as illegal. (Hence the advice not to leave the country.)

We certainly need a way to stop cybercrime, international or otherwise. But this doesn't strike me as a good way to go about it.

Finally a prediction: in the not too distant future, one or more offshoots from this will turn into unmanageable domestic criminal organizations.

Arms races always end well for everyone involved...

> Nobody with any sanity whatsoever would honestly suggest that at this point.

They're already using AI to identify military targets.... that's why they blew up a school full of children with cruise missiles, remember?

=Smidge=

Third option: The same person/people who has influence over Trump also has influence over everyone else in the party. That being big donors and foreign powers, mostly.

But in terms of DC politics it's mostly fear of breaking ranks since that'll get you ostracized.
=Smidge=

It's cute you don't think he's going to do it anyway shortly after the deal closes...

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.