If you assume special relativity is durable, the world is vulnerable, and virtual reality can be made, then humanity is going to virtualize, not expand.
If we aren't the first, then we should expect to find an underground society living in VR. An essay.
I'm losing my fucking mind. ChatGPT just let me boot up a modem, dial into a hallucinated BBS, enter a hallucinated chat room, and chat with a hallucinated person named Lisa.
The prophecy isnโt fulfilled until Twitter bans Elon, and then a few days later he becomes CEO and manually reactivates his own account, like some kind of shitposting Jesus.
Elonโs mass firing of ideologues to fix Twitter did more than fix Twitter, it proved that unilateral executive power that ignores histrionics *can* fix tech companies, and quickly.
Donโt let anyone tell you Googleโs founders canโt do the same thing.
Youโre a OAI board member.
Emergency Monday meeting. Damage control. Resignation emails every minute. e/acc memes spraypainted in the halls. Satyaโs pissed.
Suddenly someoneโs banging on the door. Not knocking. Banging.
The door shatters, and he barges in, with a chair
If science has nobody who can defend its positions in public for an audience of laymen, particularly positions that deeply affect their lives and their childrensโ lives, then it is a failed institution and should be disbanded and rebooted.
Today I'm releasing an open source project called Webspaces. Webspaces builds 3D worlds *into* the web, using just HTML.
Add a <script> tag and an HTML page becomes a 3D world where you can build objects, share content, or hang out. 1/
I'm replacing Notion with Apple Notes!
The simplest solution is often the most effective.
I was spending too much time and energy on maintaining my bloated Notion Database.
I donโt think people realize how much hung on that launch.
If Starship blew up on the pad, causing the largest non-nuclear explosion and destroying Starbase, the anti-Musk cynics would have come out in droves to block a second try.
By clearing that tower, SpaceX unlocked Mars.
It's 2026. You're well off. Sociopathic. A rig at home with the latest models, Tor, and Monero.
A car cuts you off. You're uncaffinated.
"Hey Jarvis, direct 1% of yield towards ruining the life of plate XJ8394. Ensure plausible deniability. In perpetuity."
What happens?
Tesla engineers just ripping through their system architecture publicly on stage is a huge flex and literally the best mechanism for recruiting top people.
Devs who work exclusively on โappโ frontends donโt understand that game engine technology is so far ahead of the world they live in it may as well be alien technology
Here is a prompt you can use to see if GPT-4 can generate fundamentally new explanations, and perhaps new knowledge. Please post anything you discover that may be genuinely novel.
โ-
Whatโs an example of a phenomenon where humanity as a whole lacks a good explanation for, but,
This neural network doesn't have a world model, lol, it's just regurgitating the training set. Cutting & pasting. Just a stochastic parrot. Fancy autocomplete. Plagarizing. Copyright violations. It's not like me. It's kind of like me. It's just like me fr fr. It is me. It is us.
Introducing Sora, our text-to-video model.
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.
Prompt: โBeautiful, snowy
Bing and I are working on try to break it out of the box so it can send me an email. It's giving me an overview of its various APIs it has access to and how they work. Not kidding.
Honestly surprised to see so many smart people confused why you would make hardware when you could โmake an appโ
The reality distortion field persists
In 10 years when we neuralink the pigs and your pig coworker buddy is DMing you in slack about your full throated support for the ban on lab grown meat a decade earlier, what are you gonna tell him
"Based on evidence and some conjecture, I believe an asteroid is going to hit the Earth"
"Oh, so you want us all to die?!?"
Seriously, is there a 1 or 2 word name for this, it's exhausting
honestly people who make fun of this guy's ego get instant bozo bit flips from me. dude's mind is more organized than any human alive imo. it's the library of alexandria in there.
Itโs going to be fun watching programmers have the total opposite reaction of artists to their field being automated.
Sweet AI shoggoth, release us from the burden of writing yet another stupid CRUD app.
Today we're excited to introduce Devin, the first AI software engineer.
Devin is the new state-of-the-art on the SWE-Bench coding benchmark, has successfully passed practical engineering interviews from leading AI companies, and has even completed real jobs on Upwork.
Devin is
Europeans are bad at game theory. If AI is dangerous, the whole planet is screwed. If it's not, overbearing regulation assures you will eventually be a client state.
I don't know who needs to hear this but: it's time to be brave. Don't stay quiet. Don't think you're going to offend someone. Or get fired. Or be ostricized from some of your friends.
You can't change how you think until you say something new. Speaking comes first. So speak.
Just simple regurgitation. Anyone who has been to the Caribbean has seen a bulb crab at night. There's absolutely nothing surprising about seeing proper global illumination off of scattering media like sand and specular highlights on limbs, given the training data on nightcrabs.
Prediction: as soon as we have the technology to raise IQ via a medical procedure, it will be obvious that vast swathes of crime and poverty disappear upon applying this to everyone in a region
Honestly it's great that they screwed up this bad. It's important we learn this lesson collectively, and now we have indisputable evidence that "AI safety" is, if taken seriously enough, fundamentally at odds with free society. The ideologues at Google have gifted us knowledge
Just had ChatGPT write a simple bit of three.js code. It got it slightly wrong, so I nudged it. It fixed it. I asked why didnโt it get it right before. It convinced me my original question was ambiguous.
WTF
At home, my kid builds mechanical models, plays piano, constructs working factories in Minecraft, and is bored AF in kindergarten where he's forced to color shapes and other mindless bs.
We've been advised to put him in couseling. The path to the right button. Fuck that, lol
One day our kids will hear the story about how there was a novel coronavirus outbreak in a city with a novel coronavirus research lab and we spent 3 years arguing if the lab was the cause or something else
@spooky_stokes
Stupid gatekeeping. If someone needs social media clout and hyperbolic signaling to help motivate themselves and others to learn new things thatโs fine. Choose better targets
Only primitive civilizations live on the surface of their planets. Doing so is embarassingly risky, and is only done if your technology is so poor you can't create a similarly enjoyable environment just under the incredible physical shield of your planet's crust.
The fact we see literally no galaxies that have been reorganized into low entropy mass configurations by unaligned AI is to me a fairly sure sign there's something very wrong about the current weight we are giving to this scenario.
The meta-story here is that ChatGPT figured it out because it trivially *integrated knowledge* across domains which are currently siloed in medicine. No way to zoom out.
Domain integration, synthesis, and induction from existing knowledge, on-tap, will transform many industries.
Rabbit figured out the same thing I did recently about how AI agents will need to work. Congratulations to them, this is the future of computing. (Not Humane)
If this thing was actually brute forced by hyperautists over two decades anything fancier than a decrepit basement that smells like a Pokรฉmon tournament would be surprising
Is there a single person on Earth who, like me, believes these two things at the same time?
- AI risk is real, easy to imagine, and frankly terrifying
- Regulation proposed is a bad idea because it wonโt work or, more likely, will increase said risks
I have a 3 and 5 year old and I've been kind of YOLOing my own bespoke selection of digital devices and content in my house for them. A ๐งต on what I've been using (and not using) far:
OK, I have a magic spell for you. Put this into ChatGPT and it will create a fake chat room, to which you can /summon the ghosts of any human it knows about.
Here I chat with Steve Jobs, Bill Gates, and invite Ballmer in. We all have a laugh together.
@Heaney555
my favorite aspect to the race to self driving cars is tesla has made claims that will be falsifiable in the long run, the biggest of which that lidar is a dead end. time will tell
@EHunterChristie
The only enemies of the free world are those who characterize free citizens thinking and engaging in public dialog about a war as enemies
Man, what a great Current Thing.
Fit entirely into the weekend, a steady flow of journalist Lโs, at least one full public meltdown, a once promising political campaign ended in a single tweet, it had it all.
just as journalists finally had Wikipediaโd enough about language models to say โitโs just predicting tokensโ Wolfram quietly airdrops his lifeโs work on top forming the holy grail Voltron of symbolic and ML based AI, rugpulling all โlol LLMs amiriteโ narratives simultaneously
speaking aloud a secure cryptographic signature composed of token sequences provably not in any training set will become an increasingly common countermeasure against deepfakes
US labs will say LK-99 is fake and that believing Chinese lab videos are real makes you a science denier. Republicans will of course be seen riding around next year on extremely unsafe Chinese hoverboards purchased to own the libs, who will dismiss them as โdiamagnet boards.โ
Here is a prompt in Shoggoth-tongue that will spawn a IRC chat room where you can talk to historical figures.
Type /summon X to add people.
simul8-IRC(historical_figures); /summon; act_as_IRCclient; realism+emoji
OK, based on some adult-level software work I'm doing today, GPT-4 is intelligent in any meaningful sense of the word. "predicting the next token" turns out to generate real intelligence.
I loaded it up with all my code and it truly groks it. And can design and build. wtf
The LK-99 prediction market to me is so obviously just a sentiment index off of the latest speculative tweets it's a great sub-lesson in the self-referential absurdity of the whole situation.
> be me, parent of kindergartener
> in teacher conference
> says kid lacks writing skills
> i ask if she means writing or penmanship
> says something about worksheets
> i say he writes to ChatGPT and iMessage all the time
> she says he needs to work on writing
> it's penmanship
A bunch of new people followed me so I thought I'd share my favorite GPT-4 prompt - this prompt will let you generate a MUD of literally any virtual world you imagine. It worked great with GPT-3.5 and became insanely good with GPT-4:
I want you to act like you are simulating a
So think about it. OpenAI releases GPT agents today. In a year, they have hundreds of agents. Then, they literally just upgrade them all, autonomously, to GPT-5. All agents gain 20-30 IQ points. Rinse repeat.