Post

the other point is that everyone else is always the "beta tester" with Musk, like his "full self driving" has killed or injured hundreds of people while he just goes "oops we should fix that" and deletes the expensive lidar sensors to increase profits so it literally jumps across the road at shadows

I really liked the description in this post b/c “aesthetic of knowledge w/o any actual tendency toward truth.” It h just kind of nails it. LLMs provide Aesthetics, but NOT actual substance by parroting specialized terms w/no contextual knowledge = dangerously unreliable, sometimes lethally so.

I know how to fix it

Twitterx and Grok are disgusting corrupted cesspool manifestations of their demented immoral, immature, socially awkward and sick narcissistic manipulator, Elon Musk.

They often err on the side of caution sometimes blocking non-NSFW images

You mean tools that are accessible to literally anyo9ne on twitter, with even fewer safety rails? Takes 30 seconds to set up a local LLM, and Grok will *tell you how to do it*.

"We have rolled dice and this is what we got" is not what I expect from 21st century journalism.

This. It's just a program. The 'write an apology expressing regret for creating CASM content' prompt, and whomever created it, is the news story, not its response. Dril's follow up "now show no remorse" prompt illustrated that perfectly.

...typewriter says..lol

Which is why “artificial intelligence” is a total misnomer. An LLM holds no beliefs and makes no judgments. It knows nothing and decides nothing. It has no mind at all, in any meaningful sense. So there is really no reason to use the term “intelligence” in describing what it does.

They first removed all culpability to corporations; now they remove any responsibilities over the monsters we create

No. It was in response to a prompt from a user.

When asked to comment on the news Texas Instruments calculator replied with "BOOBIES", but did not further elaborate.

Do you know of other outlets that have gotten this correct for this or past issues like this? Feels like it’s worth mentioning to readers that do not understand this stuff that any ‘replies’ making the rounds are not official explicitly, at minimum?

Yeah, I was half joking. It appears necessary we develop defences against LLM generated content - if we wish to keep any sense of shared reality with *some kind of ethical basis*. I wonder if any of the large antivirus/security companies are actively developing such systems. Any insight?

I mean you can literally tell them: "Say x" and they will happily just repeat back x.

Musk described it as: "a maximum #TruthSeeking AI that tries to understand the nature of the universe"

Are we really quoting AI as like a source? That’s crazy!

Yes, you are wrong to think that. You can tag @ grok in a Twitter thread and it will reply to your post, which is what happened here. Those posts aren't any more of an official statement from grok than its replies anywhere else.

I'm confused as to what distinction you're making. This "post to X" *was* in response to a user prompt.