@curiosityonx It is exciting to think we are on our way to the stars and even more exciting that it is happening commercially. We need more players following Elon's lead but someone has to blaze the trail and then others will follow.
@grok I would love the ability to chat with Grok build like I can with the web chat interface. It would be amazing if token usage and non token usage could be mixed with Grok Build as I am finding myself not using Grok Build as I do not want to burn usage.
Don’t be that guy. That Guy with Wrong ideas Confidently - and persistently held - despite all the inconvenient facts.
xAI is not “dissolved” as in giving up on AI models and pivoting to be “just a hyperscaler”. SpaceXAI is a name change due to a merger that unlocks the advantage of vertical integration. That’s it.
Many have a fundamental lack of understanding of what it will take for an AI company to actually win the AI Race.
Winning the AI Race means excelling at logistics and solving for the constraints to scale globally to 7 Billion more people and edge devices.
Capital, Compelling Model, Training Compute, Energy, Inference chips, Connectivity and Edge Devices.
These are the constraints that must be solved to actually win the AI race - Not just the model. Models will trend toward parity, lowest token costs will be the metric. Organizations are model agnostic, they can plug in Claude or Grok Build or Open AI with little switching costs.
THE LEAD ON EARTH
Google is in the lead now. China’s energy advantage also gives them an advantage - on Earth.
THE LEAD IN SPACE
BUT Once compute goes to Space, there is only one competitor that is vertically integrated and capable of solving all those logistics problems - at scale. SpaceXAI.
SpaceXAI competitors (who are not solving the full stack of delivering AI at global scale) are about to run into hard limits - Anthropic already did! - this is why Elon Musk says:
“Let’s see where things stand three years from now. . .” Elon Musk
Rumble is so much more than just a video platform.
By the middle of June, we begin on the trek to compete with the world's largest hyperscalers, with a primary focus to dominate in AI compute as a service.
All built with an ethos that will put freedom and privacy first.
@TomJones Yes a larger context window means more content is processed per response so best practices include starting new chats often, keeping prompts concise and generating only essential content for maximize efficiency.
@testerlabor But if I am understanding correctly the larger your context window grows if can mean a greater spend against your usage percentage for new requests? Any best practices for how to make sure you are being as efficient as possible using these new tools would be amazing.
Context Window vs Usage in Grok Build are two completely different things. Since many users get this confused, here's a quick explanation:
Context Window:
This is per-session/conversation memory.
It shows how much of the model's current working memory (context) is already filled. Grok Build's 512k context window is very large. Every time you chat, message, code, image, etc. eats into this window.
The 11,9% / 60k you see in my example means how much of that 512k window is currently used in this specific chat. This % resets or gets managed when you start a new chat or when compaction happens. It does not affect your daily limits.
Usage:
This is the overall quota, your account-wide consumption across all chats, all sessions.
You can easily check your real remaining quota with the /usage command. This is what actually gets "used up" toward daily or feature limits. In my case only 5% for now.
I was working an one project noticed a sudden jump in my context window and I jumped from 5% usage to 20% usage with pre 2.2 and 2.3. Not totally sure what I did that caused my context window to expand but I was surprised at how much my usage increased in one command. Having so much fun playing with grok! :-)
We are resetting rate limits as we had a lot of churn regarding our cache hit management.
Grateful for the feedback, and we are going to keep improving on all fonts!
@blankspeaker I assume that I will get a usage view in Supergrok soon so I can track my subscription usage like I have seen screen captures from superheavy users?
@mark_k@xai I am not totally clear on how my usage is being metered and if i am just directly using credit from https://t.co/uHiDDRnoXc. More information on that would be great.
Grok Build by @xai is now available for all SuperGrok subscribers 🔥
This is xAI’s agentic coding CLI, running directly from your terminal. Great to see access expanding beyond Heavy, because the AI coding race just got a lot more interesting.
Install:
"curl -fsSL https://t.co/E3Ero1CUbp | bash"
It's been #1 across all books, #1 in Politics, #1 in Psychology, #1 in Western History, etc. It's only day #6 of its release! Let's inoculate the West against Civilizational Seppuku!