Well, you’d also be forgiven for thinking ‘how on earth can a social website chatbot be a white supremacist?’ And yet xAI managed to prove that is a legitimate concern.

xAI has a shocking track record of poor decisions when it comes to training and prompting their AIs. If anyone can make a partisan coding assistant, they can. Indeed, given their leadership and past performance, we might expect them to explicitly try.

What’s their incentive to do this? What do they gain by making a partisan model instead of one that just works well?

You really can’t think of ANY advantage to becoming a perfected propaganda machine? Not one?

Enlighten me. How would a partisan coding model help.

Competence in every field is correlated for LLMs. Better coding probably means more competent rhetoric and more competent Swahili-Latin translation. But only "probably", the causation is being argued about.

Perhaps you’ve never heard of Tay?

Microsoft did pioneering work in the Nazi chatbot space.

Fwiw Tay was unintentional and was shut down immediately upon realization… very good case study for safety folks!

Even on regular Grok, I've seen it disagree with fundamental consensus viewpoints of people on the right. You're reading a lot of comments from people who have never used Grok in any way.

I urge anyone who disagrees to use grok and get it to say something obviously untrue and right wing. I have used it many times and it is clearly balanced.

I mean the subject is if a coding model has a bias but the regular one does not.

FYI you can try Grok for free on their website and see for yourself.

I am sure this of course a good faith argument and no need to once again teach the point of everything being, in a sense, political.

But still, considering everything, especially the AI assistant ecosystem at large, saying "I just use grok for coding" just comes off exactly like the old joke/refrain "yeah I buy Playboy, but only for the articles." Like yeah buddy, suuure.

It was a good faith question. I use perplexity for my searches/research today. I have NO intention of moving that to X (although grok maybe one of the models I can use underneath, I haven’t explicitly enabled it).

I don’t use social media in general, maybe YouTube but it’s been a real challenge to get rid of all the political content - both left and right wing.

I have used both Grok and Perplexity, and I've recently decided to just use Perplexity, even when I tell it to use Grok under the covers, I like the way Perplexity organizes things.