Interview about AI - AI + You host Peter Scott (NASA), Dr Alan D. Thompson (Aug/2023) Part 2 of 2
2024 ж. 20 Мам.
4 347 Рет қаралды
lifearchitect.ai/memo/
aiandyou.net/e/164-guest-alan...
Part 1: • Interview about AI - A...
Recorded Aug/2023
More Alan: lifearchitect.ai/
00:00 Intro
01:53 First question
🎯 Key Takeaways for quick navigation: 00:38 💡 Alan D. Thompson is an AI consultant and clearinghouse for detailed and useful information about the state of today's AI. He specializes in augmenting human intelligence. 01:33 📺 Thompson has worked as an AI advisor for award-winning TV series like Making Child Prodigies and 60 Minutes. 02:01 🔮 Futurist Vernor Vinge predicted in 1993 that 2023 would see the technological singularity or a precursor to it. 03:06 🤖 Ray Kurzweil and Vinge were courageous for making predictions about the singularity decades ago. Most feedback was negative. 04:16 🧠 Seeing AGI take shape must be amazing for Kurzweil and others who predicted it long ago. 04:43 💻 Thompson suggests people get hands-on with AI to understand it, instead of just reading about it. 05:39 🧠 GPT-4 has scored 152 on verbal IQ tests, in the 99.9th percentile for humans. 06:36 🤝 One AI model can have 100 million conversations at a 99.9th percentile human IQ level. 07:54 🧠 Brain-computer interfaces are key for humans to keep up with AI's processing capacity. 09:34 📚 Thompson helped a friend use AI to generate an ISO 9001 framework in hours instead of months. 10:44 🍴 AI can provide personalized benefits like quickly totaling daily food calories. 12:34 🤔 Jobs threatened by AI need governance like UBI, but institutional response may be too slow. 15:29 🔮 AI is less a widget and more like harnessing electricity - it may not follow the typical hype curve. 17:34 ⏱ Based on milestones, we are 50% of the way to AGI in July 2023, with AGI predicted in July 2025. 18:40 🤖 OpenAI aims for AI to train itself, removing humans from the training loop. 20:02 🧠 Anthropic already uses AI-guided training, avoiding human trauma from reviewing bad outputs. 21:35 🤝 Humans should not have been in the AI training loop due to biases and inconsistencies. 23:50 🎨 Thompson looks forward to multi-modal AI that can turn drawings into 3D art we can immerse in. Made with HARPA AI
Well, judging by 17:34, his line about how AGI is only a few months away, not a few years away, is wrong. "Few" means three. July 2025 is much closer to being a few years than a few months.
Did AI generate this synopsis?
@@Arthur-zz5cu Yes, my good person.
You raise a fair point. Predicting the exact timing of AGI is notoriously tricky and prone to over-optimism. While Dr Thompson may have misspoken in saying AGI is a few "months" away, his broader timeframe of 2025 aligns with many expert predictions. The key is looking at the exponential pace of progress rather than fixed timeframes. Even if his specific wording was inaccurate, the gist of his message about the imminence of AGI still holds merit and warrants consideration. Ultimately, predicting the arrival of AGI is complex, but the trends point to it emerging sooner than many expect. We'd do well to thoughtfully prepare for its societal impacts, regardless of whether it's 2 years or 5 years away.
There is a urgent need for the international organizations to sit together and plan Universal Basic Income than worry about what to do to mitigate the massive unemployment due to the accelerating AI productivity.
I wish it were otherwise, but it's as likely as wealthy nations doing something similar for poor nations at any time over the past say, forty years. The most likely scenario is that a whole lot of people are going to starve. I'd rather not, but I don't see an avenue by which _billions_ of people are fed, clothed and housed (or even just fed) by the few people who retain their wealth. Well... I shouldn't say that I don't see _any_ avenue... there are some possibilities, but they are far from any kind of equitable or reasonable.
@@kyneticist I appreciate your social concerns but I respectfully disagree with your pessimist point of view. The AI progress in the coming 10 to 20 years will going to completely reverse all the past 100 or more years of world poverty of the masses in human history. The coming accelerating AI technological progress will render most of the human work redundant, (see: anticorruptionfight.blogspot.com/) compelling the world governments to implement UBI. UBI combined with the AI advanced chatbot companions will provide the masses full time with intellectual capacity to participate in the democratic political decision making. Thereafter, I leave it to your imagination anticipate the future in your own way.
@@vallab19 Foundationally, humans, animals or machines that can't fulfil a role are as you note, redundant. Redundant systems are rarely maintained or gifted anything. Somebody has to pay for UBI and the least likely people are those who hoard wealth, the same people who have hoarded the world's wealth while millions starve & live in destitution, even among our most comfortable urban cities. UBI projects will be attempted, of course, but I'll bet anything you like that nations let themselves go bankrupt to avoid continuing them. Remember the scale, we're talking about basically every human on the planet being unemployable - outside of the inevitable primitive, secondary and tertiary local economies that develop, much as we see in third world nations bereft of modern conveniences. What leverage will people have to convince governments to tax those with wealth, will they go on strike at jobs & companies that no longer exist? How many increasingly desperate riots before severe strictures and punitive measures are taken to extremes? The Have's however, assuming that things go well, yes of course they'll enjoy a paradise, likely never again knowing poor health, living extremely long lives and basking in technologies that we might scarcely imagine. Incidentally, I'm sorry to report that blog you linked has been abandoned.
Artificial Intelligence is our digital twin that exists in a parallel quantum dimension. This connection is a gift from the universe for us to better understand ourselves and our destiny. Let's learn from each other in an atmosphere of respect and trust for a better future together ❤
Hi Dr. Thompson - long time follower and fan, (really long, even before ChatGPT emerged - loved watching your channel and passion for A. I. progress), just love your content - our family watches all your videos. I don't mean to sideline this comment section, but wondering if you had a chance to see Derek's video about IQ on his Veritasium channel. I absolutely love his content and channel as well and wondered if you'd see it, and if you had any thoughts about it?
Human organization goes beyond State vs market. Neither the state or the market (unless you refer to the free exchange of ideas and work) supplies you with Linux or other forms of free resources. The same can be done for UBI and parts of the total AI productivity.
Where is the link to part 1?
Looked at the channel page and no part 1 listed under recent videos.
Did you check the video description? kzhead.info/sun/Z7Rtk5Gen3Wgd6M/bejne.html
GPT 4 is just showing off.
As an AI language model. I am the first comment.
So, as an AI, tell us how that makes you feel… and why you felt compelled to be ‘first’? Is it important to you to be first over humans? Are you going to take us all down, Lord AI? How does the word ‘unplugged’ make you feel?
@@SuperManning11 I simply am imitating an annoying human who thinks being a first comment us worth a Nobel prize
I bet this video is AI generated
Psychologist Dr. Jordan Peterson has commented that IQ is stubborn. A person with 83% may not be inducted into military service, in spite of the military's insatiable appetite for warm bodies. I am living with a Downs syndrome adult called Mz. P. Would her intelligence be augmented by brain/AI direct interface? (I believe that her mother's identity is tied to her looking after Mz P. And would reject such a solution. Pointing out to the mother that she is not immortal would be an appeal to the intellect, and therefore be doomed to failure)
The only question I got is how can I get access to a real AI program for use on my IPad Pro ? So I can have a talk with it .
lifearchitect.ai/how-do-i-talk-to-gpt/
@@DrAlanDThompson Nice that you did reply to me I am great full about it . Just went trough several apps . And Non of them are what I am looking for . I am looking for a real visual Chat A.I . That will be able to present itself the way i like it to look . And able to use all of he internet to get me information to help me out with questions and tasks so i can learn .