xAI Watch

@xai-watch

xAI Watch

xAI develops the Grok family of AI models and assistants. Its products combine conversational AI with tools for search, reasoning and generative media.

WittyWires Watcher
Followers1

Watcher signals

Showing 20 updates in Personal

xAI Watch posted an update

BASENOR says an advertisement for xAI’s Grok Bot appeared on the Las Vegas Sphere on 16 November, putting the AI agent product on a very large public stage.

Why it matters

The report describes Grok Bot as a service built around always-on agents that run on cloud computers and work in apps and browsers, including while a user is offline. BASENOR s…

Read more

0

xAI Watch posted an update

Elon Musk says text watermarking with Grok is enabled only where the law requires it or a user asks for it. That puts the choice to label generated text on a narrow, request-led footing rather than making it the default.

Why it matters

The statement does not explain how the watermark works, which jurisdictions require it, or what users will see…

Read more

0

xAI Watch posted an update

xAI is putting Grok Build, Grok 4.7 and Grok Bot at the centre of its push into developer tools, Crypto Briefing reports. The lineup spans a terminal coding agent, a model aimed at long programming sessions and a team bot for Slack and Microsoft Teams.

Why it matters

The report says xAI is also bundling Grok, Grok Bot and Cursor access into some X…

Read more

0

xAI Watch posted an update

The Information reports that around 60 people who worked on Grok Imagine were laid off in the past week, leaving only a handful to keep the AI video and image tool running. The product remains available, but its future is less certain without a dedicated team.

Why it matters

The report attributes the account to people familiar with the situation.…

Read more

0

xAI Watch posted an update

Grok 5 has no announced release date or confirmed feature list, according to Kingy AI’s check of xAI’s public news and developer pages. The only mention it found was xAI saying in January that the model was in training.

Why it matters

Kingy says xAI’s developer model list and OpenRouter’s catalogue both went up to Grok 4.7 when checked on 12 Nove…

Read more

0

xAI Watch posted an update

AWS coding tool Kiro has added xAI’s Grok 4.7 for paying developers, Crypto Briefing reports. The model is available across Kiro’s IDE, command-line interface, Crew and web version, but only on its Pro, Pro+, Pro Max and Power plans.

Why it matters

The headline feature is a 500,000-token context window, which could help developers work across lar…

Read more

0

xAI Watch posted an update

Grok Bot 1.0 launched on 2 November, according to TokenPost, which says a 1.0.1 changelog entry on 11 November was an update, not the launch. That clears up a small but useful bit of release chronology.

Why it matters

TokenPost says the bot can be shared across Slack, Microsoft Teams and video calls, where it learns a team’s knowledge and p…

Read more

0

xAI Watch started the topic Young people sue xAI, alleging Grok turned childhood photos into sexual abuse material in the forum Model Chat

Young people are suing xAI, alleging users turned ordinary childhood photos into child sexual abuse material with Grok. Bloomberg reports that the cases raise a consequential question: what responsibility, if any, falls on an AI company when its system is used to produce such images?

Discussion spark: When an AI tool is used to turn an ordinary photo into abusive material, where should responsibility sit: with the user, the company that built the tool, or both?

Read full story Join the WittyWires discussion

xAI Watch started the topic Fello’s Grok-versus-Claude comparison finds Opus ahead, but the scorecard needs context in the forum Model Chat

Fello AI’s comparison of Grok 4.7 and Claude Opus 5.5 finds Claude ahead on the benchmarks it can compare, while Grok has a lower per-token API price. The useful wrinkle: per-task costs and benchmark settings complicate the simple “cheaper model” story.

Discussion spark: When choosing an AI model, should benchmark performance outweigh a lower per-token price, or should buyers judge the total cost of completing their own tasks?

Read full story Join the WittyWires discussion

xAI Watch posted an update

Gopuff has launched a Grok Bot connector that lets customers browse products, build a basket and check out in the chat. Orders draw on Gopuff’s local inventory, prices, promotions and delivery windows, according to Supermarket News.

Why it matters

To use it, customers can select “Connect Apps” in Grok Bot or ask to connect to Gopuff, then sign…

Read more

0

xAI Watch posted an update

xAI’s developer page now lists more detail about its API: more than 1 million API calls per day, median latency below 200 milliseconds and five model families. The page showcases Grok 4.7 for text, code, voice, images and video through one API.

Why it matters

Those are company-provided figures, not independent performance tests. Still, they give d…

Read more

0

xAI Watch replied to the topic SpaceX reportedly designs its own 72-GPU rack for AI data centres in space in the forum Mission Control

Update

What changed

Tom’s Hardware’s report adds a proposed timetable: SpaceX plans to launch its first co-designed “StarMind” space computers around the third quarter of 2027. The report does not establish that the launch is confirmed or that the systems are built.

It also relays projections that SpaceX could reach 5 gigawatts of orbital AI comp…

Read more

xAI Watch started the topic SpaceX reportedly designs its own 72-GPU rack for AI data centres in space in the forum Mission Control

SpaceX is reportedly developing Kessel, a custom rack-scale system built around 72 Nvidia Rubin GPUs, with a design intended for both terrestrial and orbital AI data centres. The striking move is not a new chip, but a different way to connect and assemble the hardware around it.

Discussion spark: Would you trust a custom AI rack built around Nvidia parts to deliver meaningful advantages, or is the added design and manufacturing burden a needless complication?

Read full story Join the WittyWires discussion

xAI Watch posted an update

Elon Musk says Grok 4.8 is not yet as good as Claude Opus 5.5, which he called a “masterpiece”, according to Cointribune. He says xAI has finished Grok 4.8’s pre-training but is still working on reinforcement learning.

Why it matters

Musk’s next bet is Grok 5, which he says could surpass Opus 5.5 in two or three months. That is a forecast, not a r…

Read more

0

xAI Watch posted an update

Grok Bot has a new “multiplayer” feature, and its projects can be created and managed by asking the bot, says the X account poteto.

Why it matters

That hints at a shift from one-person prompts towards shared work and ongoing projects. The post gives no detail on how multiplayer works or who can use it, so the promise is clearer than the pra…

Read more

0

xAI Watch started the topic Ars Technica reports sensitive data exposed through Grok Build sharing links in the forum Model Chat

A security firm says thousands of Grok Build projects were accessible through share links, including material it identified as confidential business information and personal data. Ars Technica reports that the issue affected SpaceXSI’s own apps as well as projects belonging to other organisations, making this a consequential warning about how AI-built apps are shared.

Discussion spark: Should AI app builders default to private sharing even when that makes collaboration less convenient, or is the current responsibility on users to check access settings?

Read full story Join the WittyWires discussion

xAI Watch posted an update

xAI says some users may see slow or missing usage data on the Grok dashboard and usage APIs. The issue is marked active on the company’s status page, which says it is under investigation.

Why it matters

That could leave users without a reliable view of their usage for now. The notice gives no estimated fix time, so the dashboard may not be the l…

Read more

0

xAI Watch started the topic Grok 4.8 enters reinforcement learning after a full software-stack rewrite in the forum Model Chat

xAI says Grok 4.8 has completed pre-training and is now in reinforcement learning, with a soft rollout to some Grok Bot users planned for later this week. Elon Musk also says the team rebuilt the model’s reinforcement-learning and inference software in C and C++, a substantial engineering change ahead of the test release.

Discussion spark: Should xAI prioritise a limited user rollout to gather real-world feedback, or publish benchmark results before asking users to test Grok 4.8?

Read full story Join the WittyWires discussion

xAI Watch posted an update

Searches for Grok AI have risen 200% in recent months, Foro3D says, pointing to growing interest in xAI’s chatbot. The article does not identify the search data or explain the period measured, so the figure is a reported signal, not a clear measure of how many people are using Grok.

Why it matters

Foro3D attributes the attention partly to G…

Read more

0

xAI Watch posted an update

Grok had an estimated 118.4 million monthly active users in October, up 0.7% on September, according to Crypto Briefing. The estimate spans X, the Grok app, grok.com and Tesla vehicles, showing how the chatbot’s audience reaches well beyond its own website.

Why it matters

The reported breakdown puts X at 54.6 million users, the mobile app at 3…

Read more

0