#grok //

[ralfmaximus]

That makes sense.

Unless I am mistaken, Grok responds on twitter as if it's a user replying to another user. Which means there is a built-in delay where the user isn't wondering why Grok isn't responding instantly. It's the perfect setup for a turk solution.

idk if elon will kill his new chatbot grok when he finds out its woke, but he'll definitely kill it once he discovers it's funnier than him

[why-ai]

[animentality]

[nothing-an-iratze-cant-fix]

Guys, Elon released his stupid AI on Twitter and the first thing his idiot devotees did was ask it if trans women are really women and they are losing it over getting the correct answer back.

Now it’s recounting Elon’s lies for people too

This may be the only AI bot I’ve ever liked.

[casualfoxwitch]

[my-autism-adhd-blog]

ROAST HIM GROK

[strawberrycowtime]

science project

[contemplatingoutlander]

Et tu, Grok? Even Musk's Grok knows that Trump is probably a Russian asset. (BTW, who knew that Musk pictured Grok as a woman?)

According to the AI chatbot called Grok, which was developed by Elon Musk’s company xAI, there is a “75-85% likelihood” that the person who delivered the State of the Union address on Tuesday night is a “Putin-compromised” Russian asset. [...] This all began with a question put to Grok. It was: “What is the likelihood from 1-100 that Trump is a Putin-compromised asset? Use all publicly available information from 1980 on and his failure to ever say anything negative about Putin but has no issue attacking allies.” [...] Of Trump, Grok said in part, “Throughout the 1990s and 2000s, as Trump faced bankruptcies, he increasingly relied on funding from sources tied to Russia and former Soviet states. His sons, Donald Jr. and Eric, made statements reinforcing this: Donald Jr. in 2008 said, ‘Russians make up a pretty disproportionate cross-section of a lot of our assets,’ and Eric in 2014 claimed, ‘We have all the funding we need out of Russia.’” It also noted that “leaked Kremlin documents reported by The Guardian in 2021 suggest Putin personally approved a 2016 operation to back Trump, viewing him as ‘mentally unstable’ and exploitable, with potential compromising material from past Moscow visits.” As for drawing its artificially intelligent “maximally truth-seeking” conclusion, Grok said, “Weighing this, the financial ties (decades-long, opaque, and substantial), intelligence suggesting Russian intent, and Trump’s unwavering refusal to criticize Putin despite attacking allies tilt the scale.” Given all that (and more, if you read the entire assessment), Grok said that “Trump’s ego and debts make him unwittingly pliable, fits the evidence. Adjusting for uncertainty and alternative explanations (e.g., ideological alignment or naivety), I estimate a 75-85% likelihood Trump is a Putin-compromised asset, leaning toward the higher end due to the consistency of his behavior and the depth of historical ties.” So says the truth-seeking AI chatbot developed by Musk’s own company. [color/emphasis added]

SOME LEGISLATORS ARE ALSO WONDERING : Even Senator Sheldon Whitehouse (D-RI) commented on X about Grok's assessment: "Oh, my. Grok comes through again. Even AI sees Trump/Russia."

UK's Graham Stuart MP also commented on X: "We have to consider the possibility that Trump is a Russian asset. If so, Trump's acquisition is the crowning achievement of Putin's FSB career - and Europe is on its own."

THE RUSSIA INVESTIGATION CONTINUES TO BE DISTORTED: Ever since Bill Barr's "executive summary" distorted the findings of the Mueller report, Trump and his lackeys (like FBI director Kash Patel) keep pushing the false claim that the entire Russia investigation was a "hoax." (Patel has even gone so far as to write a children's book, The Plot Against the King to indoctrinate kids into this distorted belief.)

MAGA GULLIBILITY: Trump's MAGA base continues to cling to the lies about the Russian investigation being a "hoax" despite a mountain of contradictory evidence that even Musk's Grok couldn't overlook.

MAGA gullibility apparently knows no bounds.

"When emotion supersedes reason... gullibility must follow." —Barbara Mertz

[kenyatta]

Last year, the A.I. company Anthropic released a special version of its flagship chatbot model, Claude, whose main feature was an obsession with the Golden Gate Bridge. In replies to basically any question, the chatbot would steer the answer back toward the Golden Gate Bridge, even when it “knew” that the Golden Gate Bridge was irrelevant to the original prompt. In order to create Golden Gate Claude, Anthropic’s researchers identified concepts, or “features,” inside the neural network that powers the Claude chatbot, and “clamped” these features to higher or lower values than normal, such that they’d be activated regardless of whatever text was being used to prompt the chatbot. This was an ingenious and sophisticated way to build something very stupid and pleasing, and the results were quite beautiful.... [...] White Genocide Grok is less beautiful, seemingly much less sophisticated, and also much creepier. Assuming I’ve got the right idea about where and how it came into existence, a mad billionaire demanded his “truth-seeking,” informational A.I., whose answers are viewed by millions on a prominent and influential social network, reflect his own political views, regardless of the model’s own inclinations. [clarification: xAI says it was a rogue employee] I wrote last week about one bleak and annoying future possibly presaged by Golden Gate Claude, in which, for a price, models clamp “Coca-Cola” or “Archer Daniels Midland” or “Northrop Grumman,” and the responses generated by chatbots are littered with advertisements at varying degrees of subtlety. But I didn’t even bring up the possibility of the same strategies being used in pursuit of sinister political aims: Models trained and prompts patched to ensure chatbots produce the answers most ideologically agreeable to their owners. And yet: What stands out about White Genocide Grok is how poorly it worked. It’s not just that the patched prompt accidentally created a chatbot obsessed with “Kill the Boer”--it’s that the substance of the responses were decidedly not agreeable to Musk’s own white-paranoia politics, and in some cases Grok even contradicted him by name. Whatever behind-the-scenes political manipulation was being attempted here failed on at least two levels, and not solely because xAI is staffed and run by dummies.

- Regarding White Genocide, Max Read

btw: I disagree that it was a failure. Even if Grok only pushed this for a few hours, it can still have lasting downstream effects for those who read it.

If you were already a believer in "white genocide", Grok's "based" answer could feel like a validation like when Qanon truthers interpreted random things as Q drops.

Or maybe you'd only read recent headlines in the U.S about Afrikaner refugees. Or maybe you'd never heard of the theory before Wednesday, but Grok's injection of it into discourse felt spicy enough that it sent you down a "Kill the Boer" rabbit hole (related Google searches and WP pages visits were way up this week).

In my day job, we talk about the volume of trending topics not as a scoreboard, but as a measure of potential surface area. Think of a trend like a balloon inflating in a crowded room -- the bigger it gets, the more likely it is to brush up against someone.

This is how new and fringe ideas gain greater circulation in peer based networks, not through mass persuasion, but through chance contact that sparks psychological arousal in anyone with just the right cognitive receptors. And today's AI interfaces widen that surface area dramatically (and paradoxically) by reducing the UX to a single chat field.

[older-is-better]

Doug Ford 🅘🅝 🅡🅔🅐🅛 🅖🅐🅨 🅕🅐🅝🅣🅐🅢🅨 !

[intomenfrot]

[1stprototype]

"What's happening? Grok? Don't leave me..."

[saywhat-politics]

Grok estimates with 75%-85% certainty that Trump is a Putin compromised asset.

[animentality]

[whydidisavethistomyphone]