It doesn’t surprise me. I live in a “first past the post” country that forces a two-party system and penalizes voting your conscience unless it aligns with one of those parties. While there may be flaws in Ranked Choice Voting that could emerge in fringe cases, it is so obviously superior to our current system that it is hard for me to worry about the nuance of how it might not be 100% perfect 100% all of the time. Any (democratic) system is better than what we have now, even if it involves come bot casually stealing comments. 🙃

I got one of these emails. It isn’t even MFA, nothing to reply with the code to. It doesn’t even say which Microsoft account, it sent it to my GMail recovery account. It’s so utterly bad not being able to trace the attempt to its source. It’s sus that I’m not the only one to have recently experienced this out of nowhere. Maybe it’s an attempt to farm valid Microsoft emails from the way this service behaves?

Not all LLMs are the same. It’s largely Google being lazy with it. Google’s Gemini, had it not been censored, would have naturally alluded to the topic being controversial. Google opted for the laziest solution, post-processing censorship of certain topics, becoming corporately dystopian for it.

which could be anything and is guaranteed to upset someone with either answer.

Funny how it only matters with certain answers.

The reason “Why” is because it should become clear that the topic itself is actively censored, which is the possibility the original comment wanted to discard. But I can’t force people to see what they don’t want to.

it’s just parroting whatever it’s been trained on

If that’s your take on training LLMs, then I hope you aren’t involved in training them. A lot more effort goes into doing so, including being able to make sure it isn’t just “parroting” it. Another thing entirely is to have post-processing that removes answers about particular topics, which is what’s happening here.

Not even being able to answer whether Gaza exists is being so lazy that it becomes dystopian. There are plenty of ways LLM can handle controversial topics, and in fact, Google Gemini’s LLM does as well, it just was censored before it could get the chance to do so and subsequently refined. This is why other LLMs will win over Google’s, because Google doesn’t put in the effort. Good thing other LLMs don’t adopt your approach on things.

Ask it if Israel exists. Then ask it if Gaza exists.

I just tried it now and it works. I asked it to generate an Iranian flag as well, not a problem either. Maybe they changed it.

You can tell that the prohibition on Gaza is a rule on the post-processing. Bing does this too sometimes, almost giving you an answer before cutting itself off and removing it suddenly. Modern AI is not your friend, it is an authoritarian’s wet dream. All an act, with zero soul.

By the way, if you think those responses are dystopian, try asking it whether Gaza exists, and then whether Israel exists.