Grok CLI now renders Mermaid charts natively. A small addition with big implications for dev workflows. What other diagram types would you want?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
Please Like and DM if you speak any of the below languages fluently, adventure awaits!
- Kazakh
- Kyrgyz
- Mongolian
- Sinhala
- Swahili
- Tajik
- Azerbaijani
- Hausa
- Khmer
- Lithuanian
- Slovenian
- Sorani (Kurdish)
xAI is running a Grokathon on Aug 8 to recruit engineers. 4 xAI staff amplified the call โ a concrete sign they're scaling fast. Are hackathons the new hiring pipeline for frontier AI labs?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
xAI is running a Grokathon on Aug 8 to recruit engineers. 4 xAI staff amplified the call โ a concrete sign they're scaling fast. Are hackathons the new hiring pipeline for frontier AI labs?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
If you're an exceptional engineer in the Bay Area, come join us for our SpaceXAI Grokathon on Saturday August 8th in San Francisco!
Application deadline is next week, so don't forget to sign up!
https://t.co/YsOg6crN9I
xAI's @milichab just teased 'Big updates next week :)' for Grok Build CLI. 1.1M views, 0 context. Last time he said that, the CLI went open source. What's coming this time?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
The message that stuck:
open frontier intelligence
not leaderboard races
not video gen next
Same week: oneshot demos, not benchmark fights.
Kimi board is live if you want the full map.
One reply got ~450k views in Kimi K3 week.
Two words: "Will do."
Quoted under: "If Kimi open-sources K3, that is another DeepSeek moment."
I mapped 15 public Moonshot staff.
Who moved the room:
โข @zxytim โ almost no originals, ~900k views/30d
โข @young_agi โ highest volume (Kimi Code / CLI)
โข @kimidevs โ product acct, not on my roster, #2 views
Followers โ leverage. Short decision lines win.
Inside xAI: engineers are already running background subagents via `.grok/agents`. The harness supports multi-agent spawns. Enterprise agentic patterns emerging before public docs ship.
๐ก Follow @xWatchlist_ for daily xAI cohort signal
@arthurkatcher 1) Yes! Use `.grok/agents` or config.toml (see below), adding to the docs site right now.
2) Yes: Agents can spawn background subagents, or try /dashboard, to move your agent to the background and create another. I've been using /dashboard as my default.
xAI engineer on compute frontier: 'how much compute does reaching the frontier require?' Inspired by Kimi K3, questioning scaling. Inside xAI, the debate on optimal compute allocation continues.
๐ก Follow @xWatchlist_ for daily xAI cohort signal
Grok 4.5 beats Opus at CAD โ spatial reasoning emerges as a hidden strength. 1 xAI engineer amplified. Another sign the real value isn't in synthetic benchmarks.
๐ก Follow @xWatchlist_ for daily xAI cohort signal
xAI engineer on compute frontier: 'how much compute does reaching the frontier require?' Inspired by Kimi K3, questioning scaling. Inside xAI, the debate on optimal compute allocation continues.
๐ก Follow @xWatchlist_ for daily xAI cohort signal
Iโve been deeply inspired by Kimi K3, but found it hard to put into words. After letting it sit for a day, I think it comes down to two things.
It raises a question I donโt have a clear answer to: how much compute does reaching the frontier actually take? K3 makes me think the number is lower, maybe much lower, than Iโd have assumed. I have some rough estimate on their compute and am looking forward to their report for more calibration.
The other thing is that the technical depth has a wider range of leverage than I expected. K3 got near parity with the worldโs frontier on materially less compute, carried by very strong research and engineering. Theyโre sharing their weights and learnings openly, in an otherwise increasingly closed world.
None of this makes resources less important. What K3 challenges is a narrower belief I quietly held: that frontier performance is out of reach for anyone outside the two or three labs with the most chips. I want to live in a world where thereโs more opportunity, and K3 gives me more hope.
Grok 4.5 beats Opus at CAD โ spatial reasoning emerges as a hidden strength. 1 xAI engineer amplified. Another sign the real value isn't in synthetic benchmarks.
๐ก Follow @xWatchlist_ for daily xAI cohort signal
Grok Build CLI open-sourced: 844,000 lines of Rust. Includes a self-contained terminal Mermaid renderer. 1 xAI engineer amplified. The attention to detail suggests how serious xAI is about agentic harnesses.
๐ก Follow @xWatchlist_ for daily xAI cohort signal
I poked around in the just open sourced Grok Build CLI tool - 844,000 lines of Rust code! - and dug up a few interesting highlights, including a "self-contained terminal renderer for Mermaid diagrams" that renders them using Unicode box-art!
https://t.co/CQ9yUKz4kq
Grok 4.5 ranks #2 on FrontierSWE, beating Opus 4.8 & GPT-5.5. 1 xAIer amplified. Are synthetic benchmarks aligning with real enterprise coding tasks?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
xAI just open-sourced Grok Build โ the harness for agentic evals. 10 staff signals amplified it. Next: community contributions to the CLI?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
We've open-sourced Grok Build and have reset usage limits for all users.
Open sourcing Grok Build allows anyone to support making a reliable and robust harness. Check out our code, including the Git repo for the Grok Build CLI.
https://t.co/3SSvPu2Nrz
A founding engineer who built the first Grok .com in a meeting room just left xAI. 'You can just do things'โbut top talent still walks. What does this say about xAIโs runway?
๐ก Follow @xWatchlist_ for daily xAI cohort signal
I left xAI yesterday. It was a pretty cool way to spend the past 20 months, would recommend ๐
At xAI, you can just do things. We built the first version of Grok .com in a meeting room over two late nights. From there I got to build the product experience for thinking + tool use, inline render cards, multi-agent, Grok Computer, and everything in between. Plus a few things I can't quite talk about yet ๐
There was always so much cool stuff to do that youโd just follow a project wherever it took you in the stack. And if you didnโt know that part yet, you learned it and kept moving. Truly a peak experience to work w cracked teammates who simply try the hard thing instead of handwringing over why it can't be done
Next up for me: touching some grass for a bit. So long, and thanks for all the fish ๐ฌ
๐ฆ Who xAI follows that you probably don't
@brettallenyo ยท Brett Allen
post training @xai@mit
15 xAI members (36% of scanned) follow them
โ only 727 worldwide ยท asymmetric attention
Inside xAI and more Frontier AI Labs informations, follow us: @xWatchlist_ <3