Musk’s chatbot stumbles again

No living human has had a week as tumultuous as Grok, the Elon Musk-sponsored AI that lives inside X for our, and its, amusement. If people were still making the Downfall Hitler meme videos, Grok’s progress would be an apt topic. Last week, Grok started spewing out anti-Semitic posts after a flurry of troll prompts.

Soon after, X shut down its newly-created “MechaHitler,” saying “We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts. Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X. xAI is training only truth-seeking and thanks to the millions of users on X, we are able to quickly identify and update the model where training could be improved.” That same day, X CEO Linda Yaccarino resigned, having sensed that the ship had taken on an insupportable amount of water.

Quite conveniently, Grok was now in the news, so Musk could immediately announce the new Grok 4, which he proclaimed, at an unveiling live on X, was the “smartest AI in the world.” He said, “It really is remarkable to see the advancement of artificial intelligence and how quickly it is evolving. AI is advancing vastly faster than any human.” He also said, “Grok 4 is smarter than almost all graduate students in all disciplines, simultaneously,” which, given the recent political track record of grad students, will hopefully take care of the anti-Semitism once and for all.

But how smart is Grok, really? It’s hard to accurately gauge. Cockburn exclusively use AI engines to amuse myself with generated images from prompts like “Godzilla eating pizza with angry clowns at a child’s birthday party.” Full access to Grok 4 costs $300. But reviews are already surfacing from people who have started Grokking. They’re not exactly glowing. “After using @grok 4 for about 6 hours so far, it is nowhere near as good as @elonmusk and co advertised,” said one X user. “Basic coding keeps breaking down.” “It’s all just hype,” says another.

“Grok 4 sucks. It can’t do simple tasks,” laments another user. “Grok 3 was better at understanding my needs and actually adjusting. Grok 4 just repeats the same thing over and over.”

Needs must be met. And while there are a small number of people wandering around X who say that Grok 4 is fast, and that it works well, the grumpy Daniels and Peters in tech far outnumber them. What if Grok 4 isn’t the dawn of SkyNet or the coming of HAL9000? What if it’s just buggy software that the biggest computer nerds on the planet can’t get to work properly? When MechaHitler finally comes, I don’t know exactly what it’s going to look like, but I doubt it’s going to be sloppy, or slow.