Grok and the Dog Which Ate the Homework

May 16, 2025

No AI, just the dinobaby expressing his opinions to Zillennials.

I remember the Tesla full self driving service. Is that available? I remember the big SpaceX rocket ship. Are those blowing up after launch? I now have to remember an “unauthorized modification” to xAI’s smart software Grok. Wow. So many items to tuck into my 80 year old brain.

I read “xAI Blames Grok’s Obsession with White Genocide on an Unauthorized Modification.” Do I believe this assertion? Of course, I believe everything I read on the sad, ad-choked, AI content bedeviled Internet.

Let’s look at the gems of truth in the report.

First, what is an unauthorized modification of a complex software humming along happily in Silicon Valley and— of all places — Memphis, a lovely town indeed. The unauthorized modification— whatever that is— caused a “bug in its AI-powered Grok chatbot.” If I understand this, a savvy person changed something he, she, or it was not supposed to modify. That change then caused a “bug.” I thought Grace Hopper nailed the idea of a “bug” when she pulled an insect from one of the dinobaby’s favorite systems, the Harvard Mark II. Are their insects at the X shops? Are these unauthorized insects interacting with unauthorized entities making changes that propagate more bugs? Yes.

Second, the malfunction occurs when “@grok” is used as a tag. I believe this because the “unauthorized modification” fiddled with the user mappings and jiggled scripts to allow the “white genocide” content to appear. This is definitely not hallucination; it is an “unauthorized modification.” (Did you know that the version of Grok available via x.com cannot return information from X.com (formerly Twitter) content. Strange? Of course not.

Third, I know that Grok, xAI, and the other X entities have “internal policies and core values.” Violating these is improper. The company — like other self regulated entities — “conducted a thorough investigation.” Absolutely. Coders at X are well equipped to perform investigations. That’s why X.com personnel are in such demand as advisors to law enforcement and cyber fraud agencies.

Finally, xAI is going to publish system prompts on Microsoft GitHub. Yes, that will definitely curtail the unauthorized modifications and bugs at X entities. What a bold solution.

The cited write up is definitely not on the same page as this dinobaby. The article reports:

A study by SaferAI, a nonprofit aiming to improve the accountability of AI labs, found xAI ranks poorly on safety among its peers, owing to its “very weak” risk management practices. Earlier this month, xAI missed a self-imposed deadline to publish a finalized AI safety framework.

This negative report may be expanded to make the case that an exploding rocket or a wonky full self driving vehicle is not safe. Everyone must believe X outfits. The company is a paragon of veracity, excellent engineering, and delivering exactly what it says it will provide. That is the way you must respond.

Stephen E Arnold, May 16, 2025

Written by Stephen E. Arnold · Filed Under AI, Business strategy, Management, News, Social Media

Comments

Got something to say?

Search the site
Subscribe to Beyond Search
Feature archive
News archive

Stephen E. Arnold monitors search, content processing, text mining and related topics from his high-tech nerve center in rural Kentucky. He tries to winnow the goose feathers from the giblets. He works with colleagues worldwide to make this Web log useful to those who want to go "beyond search". Contact him at sa [at] arnoldit.com. His Web site with additional information about search is arnoldit.com.