‘Elon #Musk’s artificial intelligence start-up #xAI has closed a $6bn funding round at a valuation of $18bn, as investor fervour for new challengers to #OpenAI continues unabated…
X’s Grok AI generated a trending story stating Golden State Warriors guard Klay Thompson had been vandalizing local homes with bricks, when in fact, he’d just had a bad game.
After shooting 0 for 10 in a loss against Sacramento Kings, users took to the platform to say the player had been “shooting bricks”… and as @engadget reports, Grok took it literally.
The #Grok AI is so absolutely lousy that the other day it turned speculation on #Twitter intro a fake news story about a nonexistent military conflict.
This is just what AI does. It launders data unintentionally at best. It doesn't understand truth and can't really check facts.
After months of work and $10 million, Databricks has unveiled DBRX - the world's most potent publicly available open-source large language model.
DBRX outperforms open models like Meta's Llama 2 across benchmarks, even nearing the abilities of OpenAI's closed GPT-4. Novel architectural tweaks like a "mixture of experts" boosted DBRX's training efficiency by 30-50%.
Trying out the recently "open sourced" #Grok Falling foul of the terrible documentation ...worse than phd-ware. At least some people are trying to improve it.
A year ago he released the promised algorithm for determining content pushed upon you (if you use twitter).
Perhaps more surprisingly, He admitted that even internally, that the #spaghetti_code comprising it isn't even fully understood internally by their staff. That's more #transparency than was expected.
But now, just a few hours ago, he released the source for #xAI's #Grok chatbot too, under the #Apache_license:
GPUs have been good for a lot of things, but #ML is far past the point where specialized hardware will help disproportionately. OTOH #Nvidia benefits disproportionately from the status quo. i imagine it’ll be some other company, like #grok, that dominates in a few years
A big part of getting the world to embrace the open web was just getting people to understand it. That you had to experience it for a while, and maybe even do a little programming or fuck up some settings to wrap your head around it. Only then, could we consider the cultural repercussions. Didn't the term #grok evolve out of that experience?
It's the very early days of decentralized social networks. We have no idea what moderation will look like when everyone is here. #fediverse#fediblockmeta
My favorite part about the #grok disaster is it was advertised as "vulgar" and "offensive" but it never swears it just starts every response with "Well well well, first of all, oh boy!!!" it sounds like a fucking 6 year old trying to be edgy
Another piece on #Grok that I was quoted in! Always nice to see Ross Anderson pop up too, I started down this whole ethics and computing thing writing my ugrad dissertation on trusted computing after reading his paper on it back in the early 2000s!
Let's consider some of the factors connecting kids to this phenomenon. Is it cute? Is it influential? Are kids impressionable? Is it really just a toy? Does it boil down to kids laughing and having fun being the most valuable part of childhood..?
"Grimes teams up with startup on AI-equipped plush toy"
"...users noticed that #Grok is "woke"; it doesn't share its creator #Elon#Musk's right-wing political or cultural views. For instance, as #Mashable's #CecilyMauran pointed out, Grok isn't aligned with Elon Musk's anti-trans beliefs, responding to a question of whether trans women are women with "yes."