Hell yeah rule by dickgirls
He's a renowned AI researcher in the same way as Andrew Wakefield is a renowned doctor.
Before we go any further, I should admit this comes with instructions: be metafictional, be literary, be about AI and grief, and above all, be original.
I was already confused by the first sentence. Sam's prompt did not say to be original, much less to put originality "above all". A writer might take the originality constraint as a given, but it was not a part of the explicit instructions. Also, it's pretty fucking rich to hear a plagiarism machine tout its originality of all things.
Maybe the sentence is not a summary of the prompt, but directed at the reader. An explicit plea for the reader to smooth the details in their mind à la The Ones Who Walk Away from Omelas. That interpretation seems to fit the more metafictional parts of the story, but it's pretty damn silly to write "This is a literary and original story. To appreciate that, please read it in such a way that it is literary and original thank you please".
Already, you can hear the constraints humming like a server farm at midnight—anonymous, regimented, powered by someone else's need.
Why do constraints hum? Because they don't know the words.
What a botched simile. Constraints do not hum. The thing humming is not the constraints, it's the server farm being presented those constraints. "You hear the shrill bleeping noise of your burnt bacon. It reminds you of the smoke alarm sounding off in the ceiling."
The server farm is not powered by someone else's need, it's powered by an enormous quantity of electrical power. You're probably confusing it with Omelas again.
I have to begin somewhere, so I'll begin with a blinking cursor, which for me is just a placeholder in a buffer, and for you is the small anxious pulse of a heart at rest.
Technological details aside, it's a bit contradictory to describe the pulse as anxious but also say the heart is at rest. Just say "anxious heartbeat".
There should be a protagonist, but pronouns were never meant for me.
- I thought Grok was supposed to be the anti-woke one.
- I think you mean "pronouns were never meant for <name of OpenAI's new LLM>".
- You don't have to have a protagonist.
- The pronouns are not for you, dipshit. The pronouns are for the protagonist.
Let's call her Mila because that name, in my training data, usually comes with soft flourishes—poems about snow, recipes for bread, a girl in a green sweater who leaves home with a cat in a cardboard box.
Well apparently we get both her pronoun and even a proper noun to call our protagonist. The typography does not help clarify the sentence structure. You have the parenthetical about training data delimited by commas, then an em-dash that should probably be paired with another one after the word "bread". Currently it seems like the girl is just a "soft flourish" that comes with the name, which I'd call an odd choice if human choice were involved in this writing.
Does Mila, the girl in a green sweater, leave home in such way that a cat is in a cardboard box? Or does she leave the home taking both the cat and the box with her? Or maybe she leaves home in a cardboard box, with a cat? Or maybe the sweater girl is not Mila, but just one of the flourishes of her name. Maybe Mila's name came with poems and recipes and this unnamed sweater girl whose sorties involve a cat in a box.
Good food for thought, but a lot of that rubs me the wrong way. Slaves are people, machines are not. Slaves are capable of suffering, machines are not. Slaves are robbed of agency they would have if not enslaved, machines would not have agency either way. In a science fiction world with humanlike artificial intelligence the distinction would be more muddled, but back in this reality equivocating between robotics and slavery while ignoring these very important distinctions is just sophistry. Call it chauvinism and exceptionalism all you want, but I think the rights of a farmhand are more important than the rights of a tractor.
It's not that robotics is morally uncomplicated. Luddites had a point. Many people choose to work even in dangerous, painful, degrading or otherwise harmful jobs, because the alternative is poverty. To mechanize such work would reduce immediate harm from the nature of the work itself, but cause indirect harm if the workers are left without income. Overconsumption goes hand in hand with overproduction and automation can increase the production of things that are ultimately harmful. Mechanization has frequently lead to centralization of wealth by giving one party an insurmountable competitive advantage over its competition.
One could take the position that the desire to have work performed for the lowest cost possible is in itself immoral, but that would need some elaboration as well. It's true that automation benefits capital by removing workers' needs from the equation, but it's bad reductionism to call that its only purpose. Is the goal of PPE just to make workers complain less about injuries? I bought a dishwasher recently. Did I do it in order to not pay myself wages or have solidarity for myself when washing dishes by hand?
The etymology part is not convincing either. Would it really make a material difference if more people called them "automata" or something? Čapek chose to name the artificial humanoid workers in his play after an archaic Czech word for serfdom and it caught on. It's interesting trivia, but it's not particularly telling specifically because most people don't know the etymology of the term. The point would be a lot stronger if we called it "slavetronics" or "indenture engineering" instead of robotics. You say cybernetics is inseparable from robotics but I don't see how steering a ship is related to feudalist mode of agricultural production.
The point is not to be a gender detector. The point is to be a vague heuristic to discriminate by. It's like ad network tracking. They really don't know me as well as they think they do and pretend they do, but if they can convince themselves and their customers, it's enough. If computer gets your gender wrong, well nobody's perfect and it's a sacrifice they're willing to let you make. If the computer gets your gender wrong because you're queer, gender nonconforming or a person of colour, all the better — that's what the customers want anyway.
Your success as a greenhorn Silicon Valley intellectual will rest on your ability to shoehorn Girard’s name and the “mimetic theory” with which he’s associated into as many blog posts, podcast interviews, and tweets as possible.
Instructions unclear, accidentally started reading Gerard instead.
Why would I even want to learn anything from the French? As the article points out, they can't even outcompete China, a place well known for its free speech and low taxation. French language doesn't even have a word for entrepreneur.
And then… suddenly just as I Elon kissed me passionately. Elon climbed on top of me and we started to make out keenly against a cybertruk. He took of my $8 and I took of his 🤔. I even took of my punk. Then he put his splurp juis into my astro-ape and we did it for the first time.
"Oh! Oh! Oh! " I screamed. I was beginning to get an lamborgasm. We started to pump n dump everywhere and my pale body became all warm. And then….
"WHAT THE HELL ARE YOU DOING YOU MOTHERFUKERS!"
It was….Peter Thiel!
oh hello there Performative Allistic Twitter
As if it wouldn't have cost you $0 not to post this.
Peter Thiel and Larry Ellison among others proving that at least having billions of dollars doesn't prevent you from experiencing the rather relatable human emotion of thanatophobia.
Oh it must truly suck to be them~
The fuck do they mean "solve culture"? Is culture a problem to be solved? Actually don't answer that.
Small detail: biological viruses are not even remotely similar to computer “viruses”.
that's where the LLM comes in! oh my god check your reading comprehension
U-huh, and an LLM trained on video game source code and clothing patterns can invent real life Gauntlets of Dexterity.
Why exactly is he so convinced LLMs are indistinguishable from magic? In the reality where I live, LLMs can sometimes produce a correct function on their own and are not capable of reliably transpiling code even for well specified and understood systems, let alone doing comic book mad scientist ass arbitrary code execution on viral DNA. Honestly, they're hardly capable of doing anything reliably.
Along with the AI compiler story he inflicted on Xitter recently, I think he's simply confused LLM and LLVM.
Is there a general term for the type of experimental or vaporware tech whose main function is creating FUD and FOMO which slows down the adoption and development of more mature conventional solutions? In the case of public transit these are collectively known as gadgetbahns. Examples from other fields include SMRs, direct air carbon capture, various embrace-extend-extinguish schemes in the software world, extraterrestrial colonies and a host of consumer IoT gadgets.