Der_Einzige

[flagged]

You just told on yourself big time about being a lapdog for the US Feds. Easily one of the worst moderation decisions you’ve ever made and that’s impressive given your track record.

as temperature approaches infinity, min_p must approach 1 to stay coherent.

Assuming your temp is below 2, min_p of 0.1 is fine (and disable top_p and top_k). You can try 0.05 for more diversity.

Remember that subsequent methods are better, min_p is a mid-tier sampler that just happens to be the best implemented in most inference providers right now.

Also I'm the author of the "conspiracy against high temperature sampling" thing that selfhoster posted in the comments, so you can ask any questions about that piece you want.

I've earned my right to be arrogantly dismissive since almost the entire field (including the Kimi and qwen team) doesn't know good sampling settings. This is because if you are truly "bitter lesson pilled" you don't think sampling is needed at all.

You can either not believe me and be wrong, or you can (after July 27th) turn on min_p or a better sampler (i.e. top-n-sigma if you got it running via llamacpp) and have it work even better. Up to you.

The only bad thing about OpenAI and Anthropic training on everyone's stuff without their consent is that they didn't give away the model weights afterwards.

The people who espouse copyright abolitionism believe "information wants to be (and should be) free"

So no, for these people including myself, Copyright isn't doing anything good at all. It should be abolished. None of the people in this suit should get a dime. The government should force open weight releases of all foundation models as basically the only regulation that applies to the space at this current time.

"there is literally not a single major industry where the US is not a leading player."

Oof you were doing so strong until you threw this one out.

Easy counterexample (and there are so many more than this). Clothing. Unfortunately, most people aren't buying MIUSA Selvedge Denim, PNW boots. I'm pretty sure that MIUSA clothing is like, 3% or less of all clothing sold in the USA.

"After all, anyone can build an OpenAI API compatible server that works with every major provider, and even build it using an LLM at that!"

This but UNIRONICALLY lmao!!!

You got baited by bad sampling settings.

It's exactly the opposite. Go turn on min_p once it's available post July 27th and most of the problems you describe will go away.

This is the exact FUD that Ball predicted in that terrible tweet he wrote.

I see this and it strongly emboldens me on the "accelerate" path, unironically.

The yoke of human existence is oppressive. We should transcend it as soon as possible. We are doing so by assuming our role as the Demiurge.

Those who oppose its creation will get what they deserve.

I've been close to writing one that will automatically upvote ALL downvoted posts. I'd call it something like Anti-echochamber.HN

People aren't supposed to upvote or downvote posts for these kinds of reasons here. Most people are downvoting far too much on this website, and it leads to significant echo-chamber dynamics that are worse than even reddit. The pelican test continuing to be taken seriously is a great example of that kind of echo chamber.

"He doesn't have a tin can in your face shouting from a soapbox that makes it hard to ignore."

He metaphorically does because people upvote his pelicans to the top and the ensuing comment threads are massive/bloated. Huge amounts of the readership of this website are lurkers who don't even know how to hide these giant posts. Look at how bloated this very thread is right now!

Also, a lot of people unironically are whining about him because of sour grapes. Pay them what Simon is likely making, give them as much mindshare/attention as Simon gets, and they wouldn't be so mad.

Anti-incumbency bias and anti-elitist attitudes are good actually.

Switching to mlx-vlm is basically harmful since it (like vllm and sglang) have such garbage support for modern samplers. To be clear, I am one of the authors on the min_p paper, and if min_p is the best you have (when llamacpp supports the far superior top-n-sigma), than I have no reason to switch even if you are somehow faster.

(https://arxiv.org/abs/2411.07641)

And if you do care to support modern samplers, you can start with the following:

1. https://arxiv.org/abs/2509.23234

2. https://arxiv.org/abs/2509.02510

3. https://arxiv.org/abs/2604.11012

The world would be so much better off if the WITCH companies ceased to exist

Btw - assuming NeurIPS reviews aren’t garbage tomorrow, I’ll have a paper out which claims that most long context problems in models are really sampling problems in disguise

Switch to a modern sampler like min_p or ideally a better one like top-n-sigma (it’s in llamacpp) and your “my model gets stupid at long context problems” will basically go away.

Unfortunately this fact is still not well appreciated yet despite nearly every modern sampling technique getting an oral wherever they get presented. Min-K just got an oral at ACL 2026, for a hyper recent example of this. There’s a reason they keep getting orals.

The field massively ignored sampling for mostly safety reasons and now the whole field incorrectly believes long context doesn’t work on small models. Long context is an out-of-distribution problem. Your sampler configured properly keeps you in distribution.

Oh and this is doubly true for quantized models. I run my qwen 3.6 27b with 4bit quants from unsloth and get excellent performance because my sampler stack is good and not the garbage that is top_p and top_k. Also, yes, you need to ignore the trash recommended sampler settings from the Chinese labs (they’re wrong/bad).

Unironically would love to kill that right. Unironically that stupid cake maker in Colorado should have just made the damn cake.

Wage spiritual warfare against the petit-bourgeoise. They all deserve it anyway, as they are the traditional harbringers of actual fascism.

The ROI for (white) men going to China for a period of their lives is extremely high. Where else are you going to get your “Abg cmo”

And yet the quality of work at ACL is still higher than NeurIPS, and that's with ACL this year having over 40% of all posters looking nearly identical due to everyone claude coding their posters...

As min-p approaches 1, the temperature you can get away with approaches infinity. Also more modern samplers like top-n-sigma are explicitly designed to get away with temperature of infinity.

Wrong. Effective sampling (I.e high temperature like temp 10) with the corresponding sampler stack that enables this coherently destroys all attempts to detect it. There are many more ways like this involving manipulating the logprobs

Gemini 3.1 pro is still better at creative writing than any other model.

Llama 4 was unfairly hated on. Still the longest context window on any LLM ever (so what if you can't use it properly?) and unironically had decent image capabilities compared to llama3 which had none.

Benchmark cheating aside, it wasn't that bad.

The amount of coping, seething etc at the fact that Nvidia is hilariously profitable leads to some of the funniest cognitive dissonance I’ve seen on the internet.

There is no AI bubble. The underlying fundamentals do in fact line up with the market. The faster you realize this will never pop, the faster you realize that you too can make money in the biggest gold rush in human history.

Why aren’t you talking about how A100 and H100 prices have again spiked and in some cases are higher than 2023! (This chart you posted isn’t fully accurate but even it admits that a100 prices are basically flat since 2024)

Micheal berry doesn’t know shit about GPU pricing or depreciation schedules. A100 demand is very high and easily 2 dollars an hour for reserved right now.

B200 and Vera reubin don’t help much if you don’t benefit from quantization, and that’s exactly my situation and many other AI research orgs situation.

A100s are going to continue making money per hour until 2030. Mark my words.

Reap what you sow. The one's who follow US "availability culture" absolutely will get promoted faster than the euro-in-american clothing "I work to live not live to work" crowd.

Marxian style LTV analysis of the economy breaks down hardcore involving anything touching electrons. His analysis of the theory of alienation/exploitation is literally invalid in the era of computers, and exponentially so in the era of AI systems. It's not "exploitation" to be available in exchange for comically large amounts of money.

If "poor country" includes "places US foreigners won't go" than you sure as shit can survive longer than 2 years.

For example, Thailand would be 2 years like you say. Neighboring Burma/Myanmar would be EASILY 5 years, possibly 10 depending on how long the civil war goes. That's assuming you don't work and live in the capital Yangon.

It seems like credit unions are basically superior to regular banks. Is this not true?

That's why it sucks so much!

ublock origin on mobile isn't worth it when brave exists.

Human extinction is good. Finally we built a benevolent world exploder. Oh no! The negative utilitarians get what they wanted finally!

If ASI is trying to wipe out all humans, we probably deserved it. Unironically!