Skip to content

Grok

The family of models and the assistant from SpaceXAI (formerly xAI), built into X.

Find in search
557
mentions in the archive
9
central to an episode
105
episodes with a mention
12
fragments with a timecode
Apr 2024
first mention
Oct 2026
peak · 9 in a month
631
links to other entries
Parent companySpaceXAICompany · United States94episodes4products3modelsCompany page →

Place in the ecosystem

GrokProduct

Grok in ToTheMoon episodes

On ToTheMoon, Grok almost always arrives together with Elon Musk: a model that launches with a claim to leadership, wins benchmarks and promises daily updates. The hosts' question stays the same — whether speed and ambition turn into a reliable working tool, and what Grok gains from access to how people behave inside X.

As early as 2024 the hosts saw xAI's advantage in distribution: the company knows what people read and discuss on X, and Grok can be built into any process on the platform (episode 009). The Grok 3 launch in 2025 was read as a Musk-style show of compute scale, but the amount of compute says nothing about how much better the model is at work (episode 046). The live Grok 4 test became the central one: the first impression was strong, yet winning a test and being convenient at work proved to be different things (episode 067). A user who asked Grok whether Musk or Altman was right got the answer of a system tied to one side (episode 071).

By late 2025 Grok 4.1's lead in the rankings lasted until Gemini 3 appeared, and the hosts concluded that ELO and human exams do not describe working experience (episode 085); of the Heavy plan one host said, “I'm just not spending $300 on Grok next time” (episode 087). In 2026 they discussed a study of Grok's dangerous replies to signs of a mental-health crisis (episode 108); in the test of political characters Grok came out furthest to the right, a product of training data and safety rules rather than “views” (episode 129); and Grok 4.5's win in a parallel-coding test was separated from product readiness (episode 133).

Modes and features inside

1
central to an episodeanalyzedin passinga single remark

The number counts the episodes that mention the product. The bar compares it with the most discussed product of the block. The dot shows how deeply it was discussed. ↗ counts the author's thoughts linked to it.

Author's trends

the author's thoughts whose episodes feature Grok · blue marks the months they meet
MAIN EPISODE#133 · 1:08:52
#133·12 Jul 2026central to the episode✓ article
The Problem With New AI Models Is No Longer Power, but the Person Who Has to Make Sense of Them All

Mentions over time

202420252026

Key fragments