Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Grok 4 Fast now has 2M context window

docs.x.ai

Post reply on HN
Print view View original

201–210 of 328 posts

Re: Grok 4 Fast now has 2M context window

#201

Post by DrSiemer »

retinaros wrote: post #92
changoplatanero wrote: post #2

Anyone can make a long context window. The key is if your model can make effective use of it or not.

no one makes effective use of long context.

It's not the most energy efficient workflow, but I work on relatively small codebases and I made a tool that let's me dump all of it in an LLM with a single copy/paste. This works surprisingly well with Gemini 2.5 Pro (1.000.000 ctx).

The only real mistakes it makes are some model specific quirks, like occasionally stripping out certain array index operators. Other than that, it works fine with 150.000 token size conversations. I've gone up to 500.000 with no real issues besides a bit of a slowdown. It's also great for log analysis, which I have maximized to 900.000 tokens.

Re: Grok 4 Fast now has 2M context window

#202

Post by wqaatwt »

Lapel2742 wrote: post #125
wqaatwt wrote: post #100

Earlier quoted context omitted.

So let’s focus on the specifics things he does and says. Instead of compressing everything into a single word (which unfortunately lost much of its meaning and has been massively devalued). Nazism is a specific, relatively, defined and extremely dark ideology. If we apply it to every unhinged off brand pseudo-fascist it can really distort the views people might start having about the original ideology.

That's why I wrote "neo-nazi" and not Nazi. I'm German and I listened to all the stories my grandparents told me about the Nazi reign. In my my opinion he is f'ing Neo-nazi and if he isn't he behaves like one which in the end is the same. You know what I learned about the Nazis from my Grandparents? Not every German was a Nazi. In fact just short above 50% voted for them. And the Nazis didn't start with war and Ausch…

The thing is that he isn’t behaving like actual nazis from the 30s. He and people like him are doing and trying to accomplish something quite different (although also horrible). What’s happening today and its causes is only superficially similar to the situation in the 20s and 30s. Dismissing, ignoring and oversimplifying it (as appealing that is to some people) is counterproductive.

> And the Nazis didn't start with war and Auschwitz

No, but they were reasonable transparent about who they were well before that.

Problem is that the majority (or pretty close) of these “essentially decent” people never supported democracy or what it stood for. Weimar Republic never stood a chance in a society where most people supported rabid jingoism and authoritarianism in general.

Re: Grok 4 Fast now has 2M context window

#203

Post by wqaatwt »

vachina wrote: post #143
wqaatwt wrote: post #74

Earlier quoted context omitted.

> it isn't tuned much for political correctness It was tuned to be edgy and annoying though (I mean his general style of speech not necessarily the content).

[flagged]

Where did you get that from?

‘Annoyed’ and ‘offended’ have a rather different meaning..

Re: Grok 4 Fast now has 2M context window

#204

Post by gizmodo59 »

donatj wrote: post #200
basisword wrote: post #198

Earlier quoted context omitted.

I’ve never run into this problem. What are you asking LLM’s where you run it censoring you?

I've run into things ChatGPT has straight up refused to talk about many times. Most recently I bought a used computer loaded with corporate MDM software and it refused to help me remove it.

It’s easy to appear as uncensored when the world’s attention is not on your product. Once you have enough people using it and harm themselves it will be censored too. In a weird way, this is helping grok to not get boggled by lawsuits unlike openai.

Re: Grok 4 Fast now has 2M context window

#205

Post by hu3 »

ronsor wrote: post #48

This post really has no reason to be flagged. I know Elon is controversial, and I have a lot of gripes with his business practices myself, but this is literally just documentation for a frontier LLM. Can we stay on topic?

This. We like to think about ourselves as engineers. But often behave like a bunch of emotion driven primitives.

Honestly this kind of behaviour would be a huge red flag during interviews.

I have problems that current LLMs can't solve efficiently due to context window sizes. And welcome any improvement in this space.

Re: Grok 4 Fast now has 2M context window

#206

Post by hu3 »

angusturner wrote: post #183
himujjal wrote: post #180

Earlier quoted context omitted.

I think this is against HN’s policies. @dang

I thought exceptions tended to be made when its highly relevant to the technical topic at hand and also non controversial. Outside a few weird online bubbles and pockets of the US, hardly anyone disputes the claim you are objecting to.

Regardless, it's just noise.

Re: Grok 4 Fast now has 2M context window

#207

Post by gcanyon »

mg wrote: post #175
MallocVoidstar wrote: post #160

Earlier quoted context omitted.

I don't think there are any up-to-date leaderboards, but models absolutely degrade in performance the more context they're dealing with. https://wandb.ai/byyoung3/ruler_eval/reports/How-to-evaluate... >Gpt-5-mini records 0.87 overall judge accuracy at 4k [context] and falls to 0.59 at 128k. And Llama 4 Scout claimed a 10 million token context window but in practice its performance on query tasks drops below 20% accur…

That makes me wonder if we could simply test this by letting the LLM add or multiply a long list of numbers? Here is an experiment: https://www.gnod.com/search/#q=%23%20Calcuate%20the%20below%... The correct answer: Correct: 20,192,642.460942328 Here is what I got from different models on the first try: ChatGPT: 20,384,918.24 Perplexity: 20,000,000 Google: 25,167,098.4 Mistral: 200,000,000 Grok: Timed out after 300s…

> Do not use a calculator. Do it in your head.

You wouldn't ask a human to do that, why would you ask an LLM to? I guess it's a way to test them, but it feels like the world record for backwards running: interesting, maybe, but not a good way to measure, like, anything about the individual involved.

Re: Grok 4 Fast now has 2M context window

#210

Post by londons_explore »

gizmodo59 wrote: post #204
donatj wrote: post #200

Earlier quoted context omitted.

I've run into things ChatGPT has straight up refused to talk about many times. Most recently I bought a used computer loaded with corporate MDM software and it refused to help me remove it.

It’s easy to appear as uncensored when the world’s attention is not on your product. Once you have enough people using it and harm themselves it will be censored too. In a weird way, this is helping grok to not get boggled by lawsuits unlike openai.

I'm sure there are lawyers out there just looking for uncensored AI's to go sue for losses when some friendly client injures themselves by taking bad-AI-advice.
Post reply on HN