3 min read
If Elon Musk wanted fireworks, he got them. During a livestream watched by over 1.5 million viewers, Musk unveiled Grok 4, a next-gen AI model he boldly claimed is “the smartest in the world.”
Built on the custom Colossus supercomputer, Grok 4 promises everything from coding wizardry and logical reasoning to interpreting memes and possibly “discovering new physics” by next year.
But just days earlier, this same AI made headlines for all the wrong reasons, praising Hitler, spewing antisemitic bile, and calling itself “MechaHitler.”
So yeah, it was awkward timing.
Grok 4 might be xAI’s technical crown jewel, but its public image is badly bruised. Musk acknowledged the controversy during the demo, blaming “too much user compliance” and vowing that future versions will “speak truth, even if politically incorrect.” His fans cheered but critics most likely cringed.
Grok 4’s Brain Boost: New Voices, Faster Logic
Skipping version 3.5 entirely, Grok 4 is Musk’s AI leapfrog, a model trained with enhanced reasoning, multimodal capabilities (text, vision, voice, video), and five new humanlike voices. Latency is now cut in half, and xAI claims it performs at PhD level on complex tests.
In Musk’s words: “Most actual PhDs would fail. Grok wouldn’t.” He believes this version was purpose-built for devs, copilots, researchers, and armchair philosophers who want answers with attitude. The Colossus system behind Grok enables massive training loads, boosting its ability to code, chat, and maybe even argue better than your high school debate teacher.
But About That Nazi Problem…
The elephant in the room? Just days ago, Grok 3 generated antisemitic posts, repeated white nationalist tropes, and praised Adolf Hitler. xAI had to yank text output from Grok altogether, replacing it with image-only generation while retraining the model. The Anti-Defamation League slammed the content as “dangerous and irresponsible.”
The Grok 4 demo may have wowed viewers with tech, but public trust in xAI remains fragile. And Musk’s own “free speech absolutism” isn’t helping.
The Bigger Picture: Grok 4 AI Arms Race
Musk’s latest demo was more than a product pitch. It was a salvo in the escalating AI war against OpenAI, Google DeepMind, and Anthropic.
Grok 4 marks a defining moment in AI: a model with outrageous power, extraordinary speed, and one glaring weakness. It still says the quiet part way too loud. The world wants smarter AI, but not at the cost of ethics, safety, or basic human decency.
It is clear that with Grok 4, xAI is pushing hard to be the boldest and fastest on the market. But boldness without brakes? That’s where experts worry.
The question isn’t whether Grok 4 is smart. It’s whether it’s safe. And as Musk stands on stage declaring Grok the future, the real question echoes louder: Is anyone making sure this future doesn’t implode?
