Applied AI @SpaceXAI, Founder @ChartThePath, Forecasting @IIForecasters, ex @shv @h2oai
Joined September 2025
- Tweets104
- Following337
- Followers285
- Likes462
Now you can use Grok Bot to even help with your groceries, snacks, household items and everything else @gopuff offers in their inventory.
Cool to see companies like Gopuff actually pushing the technology to benefit their customers
Breakdown of a bot and designing multi-bot chains
Getting Started with Grok Bot
01:00 What is @Bot?
05:26 Anatomy of a bot
08:49 3 ways to run
12:04 Multi-bot chains
16:06 Workflow 1: X followers to @Notion
20:29 Workflow 2: Personal software
23:08 Workflow 3: Coding with @cursor_ai
26:44 Workflow 4: Search
28:27 Recap
2
422
This speech-to-speech model is the real deal
Grok Voice Think Fast 2.0 is now #1 on the Artificial Analysis Speech-to-Speech Index.
This index measures whether voice agents can reason over the speech it hears, resolve real customer issues, and correctly complete tasks using agent tools.
1
20
644
Thrilled to be teaming up with the Cursor team. Now lets get to work.
Cursor is now part of @SpaceX.
Today, we have officially closed our acquisition. We will join the @spacexai team to help make Grok the world's most useful AI and improve Grok Build, Grok Bot, Grok API, Cursor, and more.
SpaceX has built some of the most inspiring and impressive technology in the world, and we’re grateful for the opportunity to become part of such a special company. Onwards.
1
23
914
The teams at @spacexai and @cursor_ai have been absolutely COOKING. Can't wait to see what others discover as best practices here too. Currently my most frequent agents for Grok Bot are focused on:
- An overall business agent connected to all of my systems to help me plan for and navigate my projects and conversations.
- Daily 5-day lookback/outlook on my projects
- Daily research into the probabilistic timeseries forecasting literature, updating my persisted knowledge base with any significant advances.
- On-demand agents for product design and prototyping
Always love to see significant improvements in open source modeling
Releasing the model weights and technical report of Kimi K3.
Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.
New model architecture: 2.5x the intelligence per unit of compute, not just more params.
Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.
Model weights: huggingface.co/moonshotai/Ki…
Tech report: github.com/MoonshotAI/Kimi-K…
Tech blog: kimi.com/blog/kimi-k3
3
345
Grok 4.5 is becoming very popular if you’re a user of @augmentcode
@grok 4.5 saw the largest week-over-week increase in usage of any model in our picker.
When the frontier is moving this fast, having your pick of models is a huge advantage: route to your favorite model today, not the top model from last quarter.
More soon on what we've been using!
19
1,487
This is quite an improvement. Been brainstorming with our team recently just around getting the model to think longer term and big picture when asked to contribute or update a code base. Great to see these results on the modeling front
2
15
7,428
Never, ever give up. The customers I’ve worked with over the last few days have already seen significant improvements in a wide range of applications.
This includes enterprise and consumer applications alike. I’m excited to see what you build!
Based on strong positive feedback from customers in our beta test program, @spacexai will make Grok 4.5 available to the public tomorrow.
It is an Opus-class model, but faster, more token-efficient and lower cost.
Absolutely love how easy it can be now to incorporate voice as a modality into whatever you're building
Introducing Voice Agent Builder: a no-code platform to create human-like voice agents with Grok Voice.
Available today at $0.05 / min.
x.ai/voice
Beyond text, I think voice will be the next modality for humanity to engage directly with other intelligences.
Neuralink is already far beyond that.
Good things come to those who train on high quality data
Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus.
RL is continuing to significantly improve the model, and the Grok Build harness gets better every day.
Nice work by all those involved!
Completely trained from scratch new models will be released by @SpaceX every month this year.
What makes "Go" special:
• Real-time personalization from your own interactions (not just past orders)
• Grok Voice for natural conversation
• Grok Imagine for visual customization
• Grok Reasoning + X location data for truly local intelligence
This is the kind of applied AI our team loves building at @SpaceX and @xai.
Huge thanks to the GoPuff team (@YakirGola @Rafaelilishayev) for trusting us with this, and to @elonmusk for pushing the tech to even make this possible for our customers.
I've learned an incredible amount working with the team on Grok Build. If you're into agentic harnesses that allow you to build anything, check out the beta released today x.ai/news/grok-build-cli
4
163