Musk Rails Against ‘Woke And Sanctimonious’ AI Rivals—Insists ‘Grok Musk Win’

Forbes ^ | Feb 18, 2026 | Siladitya Ray

[MinorityRepublican]

XAI owner Elon Musk attacked rival artificial intelligence companies like OpenAI, Google and Anthropic, accusing them of being “woke” and biased on culture war issues, as his company rolled out a public beta of the latest version of the AI model powering its chatbot Grok—which the billionaire insisted “must win” the AI race.

---

TOPICS: Computers/Internet
KEYWORDS:

[VictoryGal]

Don’t be a beta, Elon.

The free market will decide.

If Grok appeals to more people, it will win.

If it doesn’t, it’ll be one of many AI options available.

Maybe none of the companies “win”, each AI will serve the people that want it. Maybe X users prefer Grok, maybe Google users prefer Gemini. That’s okay, it really is. Calm down, sweetie.

BTW Grok has been “woke” a couple of times, too. It’s LLM learns from libs as well as conservatives.

[jroehl]

If we have learned anything about the internet in the last 30 years, it is that the best does not always win. There is a lemming affect. That is the most pervasive and accessible technologies take over, regardless of their merits. Sadly google has the upper hand here.

[Bruce Campbells Chin]

Of course, the inherent problem with any of the LLM's is that the majority of people using it will have no idea if it is giving them correct information or not. So it may become just a matter of trendiness or advertising rather than actual "quality".

People who actually know what they're talking about, and how to write, don't need any of them.

[Yo-Yo]

Someone needs to invent a “Dogpile” AI that consults ChatGPT, Grok, Copilot, etc., and presents an aggregate answer.

[Blueflag]

Id have to agree to an extent.

Google gemini is hopelessly woke, self righteous and sanctimonious. Its also untrainable to the contrary.

MS copilot is trainable. You can get it to build a profile of you and will will no longer rely on the loudest voices, rather logic and facts. It no longer uses progressive phrases with me and knows why. I use it frequently in my daily.’work.’ It does however make errors, sometimes bad ones. i ‘fixed’ that by instructing Copilot to double check its answers. That has greatly improved utility and reliability. It now also knows to never cite AP, Reuters or Al Jezeera sources.

Grok is more fun, more reliable out of the box, but I swear its like working with a brilliant coworker who’s a habitual Marijuana user who has just finished a joint.

[Jewbacca]

It’s true. ChatGPT is insufferable and will dodge any non-PC question and preach at you.

[spintreebob]

Google Gemini AI says that TRUTH SOCIAL is unreliable, especially on election accuracy reports. Gemini does not make a distinction between election fraud allegations and evidence/proof.

But Gemini AI says nothing about Truth Social type sites at the other end of the political spectrum. Yet they all do the same thing as Truth Social on elections...make allegations with not proof.

What is needed is scorecard site that scores such things as allegations with proof and allegations with no proof. And then a step further...No proof because nobody has done the hard work of finding the proof vs no proof because nobody has yet sought proof (as in the case of vote fraud in GA).

[oldtech]

What I have seen of AI so far, I am not impressed. Admittedly, I do not understand all the uses, but what I have seen is mostly stuff certainly not needed, deceptive,a waste of energy & a general waste of consumer’s time. Where would we be without it? For the most part I would probably say...just as well if not better off.

[suthener]

A few days ago Grok told me:

Abortion is not murder.
Gun control helps reduce crime.
Human caused climate change is real.

It has also thrown in completely unrelated liberal facts in answer to direct questions. I have seen it use the NAACP, Politico, and The Innocence Project for sources. I’m not impressed.