they are playing with being fast, not good lol
Grok: "I can answer any math question in less thant 500ms" User: "what is 2319321x232" Grok: "3201521321" User: "this is wrong" Grok: "but it was less than 500ms""
Grok 4.5
961–970 of 1001 posts
Re: Grok 4.5
#962Earlier quoted context omitted.
The joke is that McDonald's spends hundreds of thousands of dollars to identify new locations - traffic studies, visibility, demographics, nearby traffic generators, site characteristics, drive-thru feasibility, etc. They have one of the most rigorous processes in the industry. Burger King's process is to open a location across the street.
I think the story was Starbucks -> Seattle's Best Coffee, not McD's --> BK. but it does work I guess.
I actually heard it as Walgreen's -> CVS
Re: Grok 4.5
#963Earlier quoted context omitted.
Instead of a VPN, might I suggest running litellm proxy on a server in the US and connecting to that
What would the advantage be?
You have control over how the connections work and what they do, rather than just relying on the VPN exit server routing to get you to the right place.
With things like a fly machine or whatever, you might even be able to run it for free
Re: Grok 4.5
#964It seems to be extremely economical - 4x better reasoning efficiency compared to Opus while being priced at $2/$6. For comparison, GPT 5.4 is $2.5/$15, GPT 5.5/5.6 are $5/$30, Opus 4.8 is $5/$25, Fable is $10/$50. And by benchmarks (unless they gamed them), seems to be at around Opus 4.7 level, which is what Elon mentioned in https://x.com/elonmusk/status/2074911038286295049 . I guess the Cursor data was very useful.
until I exceed my Claude Max/$100 sub (have not hit a wall so far), the pricing of these models isn't relevant to me
Re: Grok 4.5
#965It takes a lot of digging to find their cache pricing - it's $0.50, which is unusually expensive. The vast majority of input tokens are normally cached, so this is actually the price that matters. I wonder why it's so high.
Huh so that's why it's hard to find. They probably haven't properly optimized their caching, or they are just trying to make more money from there.
Re: Grok 4.5
#966Earlier quoted context omitted.
Just to add context (no pun intended), OpenAI also charge differently based on context usage with GPT 5.5 being $5/30 below 200K, and $10/45 above. Anthropic have a fixed price regardless of context usage. These per-token pricing schemes aren't directly comparable though since these models all use different numbers of tokens, even for input (Anthropic's recent tokenizer change generates 30% more tokens for exact same…
I didn't know that, thank you. Does anyone know why they would charge more for higher context usage (other than they can)?
Think of the entirety of the context (the full thread of conversation, all tool call output, etc.) as one message that's been submitted to the LLM all at once just so it can generate the next token (which is just the next word or even syllable.) Once that token is generated it is appended to the context, and the entire context is once again used to generate the next token. Keep doing that until the entire response is generated. The larger the context, the more stuff the model has to pay attention to as it generates each next token.
To use an analogy: imagine your friend is the author of an unfinished book. They die with 19 chapters written, and on their death bed ask you to write the 20th chapter. Assuming you're up to the task, you can only do this well if you take the time to absorb the entirety of what's been written so far.
This is how LLMs work. Context caching is an optimization on top of this, but it has its limits.
Re: Grok 4.5
#967Earlier quoted context omitted.
Its more depressing to see that we have such a low morality, that it apperenly doesn't matter what you do as long as you make great tech? My philosophy is, that the better i have it, the more responsibliltiy I can/have to carry. People in tech are above avg successful and we have a thinking job. I would imagine that our world would be better if more people would care if they see a Tesla and feel a little bit bad. Yes…
I came to this thread trying to understand the technical capabilities of this release and how it fits into the current competition. I want to see a discussion about this, but I can't find it, because every reply is like yours, about philosophy and morality. It's frustrating, because I can separate the physics and the philosophy when I examine something. I can be interested in understanding how a nuclear bomb works, a…
So we just let it go?
No consequences anymore?
No reminder of what Grok stands for?
Lets forget eh?
Re: Grok 4.5
#968I echo the top comment - the politics here have gotten out of control. Like it or not, Elon and his companies have changed the world for the better. You are allowed to separate the human from the human's impact sometimes.
I kind of doubt that Twitter/X is any sort of net benefit. Boring company was a wash, Tesla is a net positive, although it also has some issues. Paypal was a huge advancement, although it's still deeply flawed. Grok may end up important, but right now it's kind of just hovering around 'acceptable' second-tier.
But SpaceX has the potential of driving some of the grandest and most revolutionary accomplishments of the 21st century. That's going to be what determines if high-schoolers recognize his name two hundred years from now.
Re: Grok 4.5
#969It seems to be extremely economical - 4x better reasoning efficiency compared to Opus while being priced at $2/$6. For comparison, GPT 5.4 is $2.5/$15, GPT 5.5/5.6 are $5/$30, Opus 4.8 is $5/$25, Fable is $10/$50. And by benchmarks (unless they gamed them), seems to be at around Opus 4.7 level, which is what Elon mentioned in https://x.com/elonmusk/status/2074911038286295049 . I guess the Cursor data was very useful.
How does it compare to Chinese APIs? It doesn't seem like xAI is meaningfully more competent or any single bit more honest than Chinese labs anyway, so you might as well send tasks straight to China unless theirs is substantially cheaper.
[dead]
Re: Grok 4.5
#970Earlier quoted context omitted.
I think US based folks perhaps aren't aware of what Musk has been doing in Europe politically. Specifically promoting - appearing at conferences, applying political pressure as well as public statements etc - actual fascists. Not right wing controversial figures etc, but literal authoritarian, anti-migrant, anti-EU, anti-democracy fascists. Not to be glib but if this was Mussolini-AI, or Pinochet-AI or what have you,…
They do know, they just don't care. As long as it doesn't affect them, it's "not political". Frankly, I am afraid. I see many takes here that resemble Russian way of thinking and being "apolitical" that I thought I'd never see from Americans.
Slashdot went the same way more than a decade ago. It could reflect more the microcosm of the aging, wealthy, silicon valley ycombinator community than a sea change in attitudes among tech people or US intelligencia in general.