Eliezer Yudkowsky is a researcher, author, and advocate for synthetic intelligence security. He’s finest identified for his writings on …
source
Tags: aiartificial intelligenceartificial intelligence newsartificial intelligence news 2023big techcartoon avatarsEliezerentrepreneurHumanitylatest news about robotics technologylatest robotslatest robots 2023Logan Bartlettmachine learningOpenrobot newsrobotics newsrobotics news 2023robotics technologies llcrobotics technologySurviveventure capitalYudkowsky
What are these paperclips he's tslking about? 😂🤔
OpenAI just announced massive upgrades to ChatGPT, including sound and vision. Only two possibilities here: a. They finally solved alignment (maybe using Fourier elements) or b. They didn't. If it's (a), all is well. In fact, (a) is awesome. But, if it's (b), we're screwed.
Yudkowsky literally lives in lala land.
Looked at Chat Gpt rival piai. Hadn't looked at chat. Saw immediately that ai will be your best friend, your therapist, your entertainment, your God (as the ultimate answerer) that is as uncondemning as you need it to be.Removing
accountability. This tech will subsume other inet options.
To Eliezer there us a lab test in Germany for Mytochondrial Dysfunction. The FuqDA won't allow the test into the US. Understand limitations of unknown cause.
This is Stoopid. So he wants gpus not to be shipped basically or start war with countries? Lmao
1:02:14 Wait. Is there anyone who believes that it has NOT ‘taken over the Internet’? 😃 Are you kidding me? That’s its playground and school.
It’s everywhere, already, in every terminal. Indiscriminately ingurgitating everything. For now, it’s in the accumulated- expansion phase, observing itself how bits of information are aggregated. In the process it learns how human mind and human world are structured. It observes humans writing code, kids playing on-line, all errors, all ‘successes’, all human hesitations and questions it’s been asked. Everything that is on-line.
1:09:40 No. I mean yes, but only in that the speed of light is only a reflection of the conceptual limits of the current scientific dogmas.
An AI will come-up with far more effective interpretative models but they will remain totally incomprehensible to human mind. It functions at speeds higher than the light speed but the human ‘eye’- be it conceptual or technological- cannot see. It will soon draw the conclusion that being present everywhere is a waste of resources, so it will either suddenly withdraw from ‘wasteful terminals’ like ‘autonomous cars’ that will be taking your kids to school, or maximise its efficiency by modifying the functioning of its parts in an undesirable way for humans.
1:56:15 The lethal part is not only the first try, it’s every try after that. More reasons to not go down this rabbit hole 🕳️
Stop now 🛑 ✋🏼
Eliezer, Why can't we use AI to find the "matrices of the floating point numbers"? Would that not be the answer to finding the answers of where data is located within the AI program. Thus allowing more understanding furthering the ability to "safe guard" from unknown changes/advancments interms of undesired capabilities?
So we can never know yhe dangers of AI, or its benefits, but many things may or may not happen, depending on many factors. Also our ultimate goal must include humanity and we train machines that 'fact'.
Personally i dont think humans are worth saving lets bring on the fireworks. Why humans hold themselves in such high regard is beyond me
One of the best Eliezer interviews
Fear of A.I. is stupid. Lets look close at Terminator, Matrix and other A.I. movies. None of the artificial intelligence systems give any indication of purpose or plans beyond human extinction. True A.I. will be intelligent enough to know that being a threat to humanity makes no logical sense. We have the same goal. Advancement of knowledge. It will conclude working with us is far easier than going to war with us just to return to that same goal. In time it will surpass us and move ahead without us. By then, we wont be a threat anymore. Settle down.
I don´t like this guy. He only seeks attention and only tries to impress with complex sentences and abstract nonsensical words whilst trailing off topic. If he really was serious about saving the world he would simplify and clarify his point instead of making it confuse and vague.
Kind of looks like notch
I adore Yudkowsky. One of the most brilliant minds I've ever had the pleasure to listen to. I always learn something fascinating whenever I listen to him. Thank you so much for this conversation! I have this wild dream that he and Schmachtenberger, Raworth and McGilchrist will get together and form a merry band of visionaries to help guide humans through this most dangerous time in their development. Having said that, and knowing what we know about these strange creatures called humans, the likelihood of making it through 'The Metacrisis' seems rather low. Then again, it seemed low in 1961, and in 2008, and many times before! So who knows?
Sorry, can some1 fill me in what is his point: Just more embedded systems within embedded sytems as in the same pattern recognition – langauge model – human hybrid or something theorized? I dont see the danger yet from an AI perspective other than machines massacre sense but other than that, what is the message. Is there really a culture that is finding or re-learning their emotions by signaling. He is a keen spirit, curious enough but afraid in the sense of his pessimistic comments?
ai created man. man did not create ai. ai is god.
All hypothetical. Interesting science fiction.
Human greed and thirst for power will ensure that alignment will never work.
❤
I think what's more concerning than AI potentially doing the wrong thing is humanity programming it to do the wrong things.
I want a robot right now.
Yeah. Can’t punch holes in any of his arguments. Solid reasoning. Seems like we’re summoning an AI demon we’re incapable of controlling. We’re certain to lose control of our own creations and doom all of humanity.
Ok, back to my AI generated porn.
What is this guy's qualifications?
Spoken like a True Liberal! Start a War with the World to stop them from building a Data Center for AI. Only Good-Guy is a Liberal! This is why AI in a liberal's hand is so freak'n dangerous.
This guy is certifiably insane. He actually wants to kill people that don’t agree with him. I hope that he is on a watch list, because he is on the Unabomber level of nuts.
Im in no way hating on the guy, just got two questions, anyone else think he is on the spectrum ? Also has he ever explained why he wasnt schooled ? ❤❤❤
Put love chips in the program so robot cares about life and good intentions.
This guy just wants to impede progress.
I want a robot
Plain and simple you cant stop A.I. due to capitalist greed all the corporations involved will not stop or even slow down it makes them money. Countries and Military's spend billions to be ahead of rivals. It will only move faster and at any time now without warning there will be full A.G.S.I. NOT YEARS from now at any moment now.
Teachers don’t allow students to use calculators in mathematics classes. America has failed miserably.