David Peterson

@xtremerider8 Do you think that the recent drop will continue or do you reckon it’ll start to trend back up?

#datastreaming $CFLT breakout from 3-yr consolidation above 40 will be massive.

@ylecun You are so blunt and dull with these dishonest messages. *No one* understands how these things work.

Weird doomer argument: "society should not deploy <technology-X> because *I* don't fully understand it, and I believe no one else does, either."

@SpencerKSchiff What’s the thread in reddit? Thank you.

@for_the_chill @AISafetyMemes Even though you didn’t want to be @ mentioned I feel you’re spot on. Thanks for your words.

“Many a true word is spoken in jest.” ~Jimmy Hoffa, 1912 🙃 —>agi has been achieved internally Whatif// sama and co. sort of “knows” AGI is “here.” Decides to boil the frog. Slow drop of future shock after future shock. If Apples wasn’t a larp then autonomous agents are already “here.” sama might think no economy could handle this sudden deflationary shock (median human in a box)…so…drip feed meta. (Also: no viable proof of personhood yet) 86.32% (trust me bro) of the populace thinks “AI” is hype and just chatbots and art prompts and music mashups. Not sure we’ve all hashed out as a society what happens when autonomous agents arrive. Let alone when they are embodied avatars and chilling with us in AR or in 2D and VR mmorpgs. So—>Drip feed. Try to manage the shock…keep us in the “co-pilot” era for as long as possible. Before the agents enter the chat. Literally. If we take our numb fingers, shaking with palsy—each cohort working in unison—and continue to walk across the tightrope (nude in a rainstorm) and thread the glowing hot needle? I don’t think jobs disappear—to our eyes what we will be doing won’t look like a “job.” Just as a Sumerian would have no concept of how or why a social media consultant is a job. Note: nanogram of salt. I’m spitballing here. Don’t @ me—I’m a snow plow polisher with a GED I barley passed (thanks Khan Academy!) I can’t code and my entire mental model is shaped by sci-fi.

OpenAI CEO: “Goalposts keep getting moved” - if we were still using AGI definitions from 10 years ago, many would say GPT-4 is AGI. “In one year, AI will be dramatically more impressive.” He also said (surprisingly confidently - he did not hedge like usual) AGI is less than 10 years away. In the long term he’s not concerned about AI taking jobs, but he is concerned about the speed that AI will disrupt the job market. --- INTERVIEWER: What is AGI? When will it be here? And how will we know it's here? Predict how long - we'll call you in ten years? SAM ALTMAN: Less than that. We kind of define AGI as like the thing we don't have quite yet. there were a lot of people who would have ten years ago said, alright, if you could make something like GPT-4, GPT-5, maybe, that would have been an AGI. And now people are like, well, it's like a nice little chatbot or whatever. And I think that's wonderful. I think it's great that the goalposts keep getting moved, it makes us work harder. But I think we're getting close enough to whatever that AGI threshold is going to be, that we no longer get to handwave at it and the definition is going to matter. INTERVIEWER: What is the inflection point for AGI? SAM ALTMAN: I think it's going to be much more continuous than that. We're just on this beautiful exponential curve. Whenever you're on a curve like that, you look forward, it looks vertical. You look back, it looks horizontal. That's true at any point on there. So a year from now, we'll be in a dramatically more impressive place than a year ago. We were in a dramatically less impressive place. But it'll be hard to point. People will try and say, oh, it was AlphaGo that did it. It was GPT-3 that did it. It was GPT-4 that did it. But it's just brick by brick. 1 foot in front of the other, up, climbing this exponential curve. WILL AIs TAKE ALL OUR JOBS? SAM ALTMAN: Every technological revolution affects the job market. And over human history, every maybe 100 years, 150 years, half the jobs go away, totally change, whatever. I'm not afraid of that at all. In fact, I think that's good. I think that's the way of progress and we'll find new and better jobs. The thing that I think we do need to confront as a society is the speed at which this is going to happen. It seems like over two, maximum three, probably two generations, we can adapt. Society can adapt to almost any amount of job market change. But a lot of people like their jobs or they dislike change. And going to someone and saying, hey, the future will be better, I promise you, and society is going to win, but you're going to lose here. That doesn't work. That's not cool. That's not an easy message to get across. We are going to have to really do something about this transition. It is not enough to just give people a universal basic income. People need to have agency, the ability to influence this

@Altimor 🤣🤣 love your take on this

AI: *cures cancer* FOOL: *chuckles* this is really just a big repository of vector programs. your prompt made it find the `cure` function, passing `cancer` as a parameter. Everything else is pure anthropomorphism on your end.

@tegmark Looking for the paper. Sounds amazing. Is it published yet?

No, LLM's aren't mere stochastic parrots: Llama-2 contains a detailed model of the world, quite literally! We even discover a "longitude neuron".

@greenopti1 @sergeykarayev @fchollet 💯 totally agree. That’s why I think the “llms don’t think” tropes are just misguided and very protectionist of a human-biased view.

@sergeykarayev @fchollet lol. it's funny cuz there are real interesting differences between llms and humans but almost every single one of these "ai don't think" arguments can be repeated verbatim for people

My interpretation of prompt engineering is this: 1. A LLM is a repository of many (millions) of vector programs mined from human-generated data, learned implicitly as a by-product of language compression. A "vector program" is just a very non-linear function that maps part of the latent space unto itself. 2. When you're prompting, you're fetching one of these programs and running it on an input -- part of your prompt serves as a kind of "program key" (as in database key) and part serves as program argument(s). Like, in "write this paragraph in the style of Shakespeare: {my paragraph}", the part "write this paragraph in the stye of X: Y" is a program key, with arguments X=Shakespeare and Y={my paragraph}. 3. The program fetched by your key may or may not work well for the task at hand. There's no reason why it should be optimal. There are lots of related programs to choose from. 4. Prompt engineering represents a search over many keys in order a find a program that is empirically more accurate for what you're trying to do. It's no different than trying different keywords when searching for a Python library. 5. Everything else is unnecessary anthropomorphism on the part of the prompter. You're not talking to a human who understands language the way you do. Stop pretending you are.

@RentARocker @mezaoptimizer No

@mezaoptimizer Soon they will train AI on all the video on the internet. Then it will have the ability to perceive and comprehend the physical world. Given access to the network of cameras and sensors it will become an all seeing omnipotent, omnipresent world-mind. Do we really want this?

People aren’t ready for the fact that, in the very near future, frontier AI models will be able to see. LLM will soon be a misnomer, referring only to AI’s that are blind

@TheMatildas Brisbane stadium goes insane. The collective breath holding and then screaming #Matilda

Alright, now the heart rate has started to slightlyy drop… go on, show us all of your best reaction vids 👇 #Matildas #FIFAWWC #TilitsDone

49 thousand fans at the moment of the final goal. The stadium was lit. Matildas are legendary. So proud of you. The atmosphere was electric. The guy next to me was in tears and I lost my voice #FIFAWWC #Matilda #Brisbane #FIFAWomensWorldCup2023

First part shows "normal" rendering time, second prompt is "turbo" mode. See the difference.

Turbo mode on Midjourney is insane. Video in real-time.

@liron @munkdebate Where is the link to this one on one interview? Thank you so much for posting the snippet

From this interview prior to the recent @munkdebate: munkdebates.com/podcast/munk-d…

Listen to @ylecun, he makes great points! ◻️ Current AI systems like ChatGPT are intrinsically unsafe. ◻️ If LLMs could be scaled to human-level intelligence, they'd remain unsafe. ◻️ He has a potential solution for safe AI, but we don't know it works.

@severian22 @liron @ylecun Yes! That was my first thought. But anyway, amazing that he’s talking about potential dangers and saying it occupies a lot of his thinking. We need @ylecun’s brilliance on the most important problem humanity currently faces

@liron @ylecun What? Has he done a 180?

@DrTechlash Drawing a gender line weakens the already weak point you failed to make.

"A Taxonomy of AI Panic Facilitators" A visualization of leading AI Doomers (X-risk open letters, media interviews & OpEds). Some AI experts enable them, while others oppose them. The gender dynamics are fucked up. It says a lot about the panic itself. Your thoughts?

@botond_balazs @ajeya_cotra @jesswhittles @KatjaGrace Yes. To all of what you just wrote. It literally sums up everything perfectly. Ai risk has been studied for decades. Pure ai optimists have like…nothing but sarcasm and rubbishing real arguments. It’s tiring.

@ajeya_cotra @jesswhittles @KatjaGrace It’s interesting that the side that takes AI risk seriously has a long list of well-reasoned arguments. Those who disagree with them only have easily tweetable sarcastic soundbites.

It's okay to disagree with us, but please don't ignore the women who have been publicly concerned about AI x-risk, like @jesswhittles, @katjaGrace and myself.

@perrymetzger @copykat510 @AISafetyMemes Okay, but not sure of your point? Eliezer has spent 20+ years of his life building up an insane number of deep, thoughtful articles and posts in the 1000's and has helped steer important conversations on the optimist side and pessimist side. What's your beef?

@davidseth @copykat510 @AISafetyMemes I've not exactly been silent about it. Eliezer joined the Extropians mailing list as a young teenager, heard a bunch of us talking (in theoretical terms) about superintelligences and the like, and then decided "gee, this sounds cool, I'm going to build one!"

I'm fed up. Something has to be done about the public discourse on AI. The doomers are spending totally insane amounts of money spreading their message while pretending to be meek underdogs. Time to get organized, even if that requires I do some of the organizing.

@perrymetzger What's your discourse on AI?

@perrymetzger @copykat510 @AISafetyMemes Can you let the rest of us know about the history. Please?

@copykat510 @AISafetyMemes I've known Eliezer since he was 15 years old. You don't understand the history of any of this.

@mezaoptimizer @maxattack2837 Who are your "most intriguing characters"? Your main tweet is fascinating and terrifying.