Awesome pod! Natalie is so great to listen to. Please have her on again after the US elections, I think that will be a great time to sync back up with Natalie and get an understanding of what may change given the election outcome.
Great conversation Peter. Measured questions and answers. Some really curious questions about humanity especially Natalie's take on diplomacy. Thanks for the podcast.
Everything you need to know about AI is in the first letter of the acronym. Living organisms are greater than the sum of their parts where code is NOT.
One reason why LLMs feel intelligent to people is because they’re trained to predict the next word given previous context. They are universal function approximators, with a huge number of parameters. Since many tasks that require intelligence can be approximated via next word prediction (because we can write them down in text), LLMs can simulate many tasks to varying degrees of accuracy. Equating the thing with a simulation of the thing is a map vs. territory error.
Abt 7 yrs ago i watched Jordan Petersons University of Toronto courses on psychology and neurological interpretation... I was amazed at how our brain maps our surroundings and how it makes sense of the world so we can function... he said back then that this was the single greatest obstacle for the "AI guys" ... he said they are discovering the problem with building artificial intelligence is that it appears to need to be embodied in order to become functional and adaptive and actually grow. Embodiment = goal directed learning
With full respect to Natalie's contributions to the Bitcoin community, her understanding of AI is pretty superficial. This is an important topic, and I would strongly recommend a discussion with Brian Roemmele who has been around the space for a couple of decades, understands it deeply, and has made open-sourcing and personalizing AI his life's work.
That was amazing. I'd never considered the profit motives behind some of the fear mongering around AI, kind of embarrassing really. I think the lingo of AI is partially to blame. "Intelligence, learning, training" lend an air of mystery to what is actually not a very mysterious process.
I'd like to debate her because many psychologists see consciousness as something that emerges not programmed. With enough computation its an emergence naturally
Nice one. Small correction, I believe that Elon Musk left the openai board after disagreement, not because he did not want openai to become "for-profit" but rather he thought that Tesla should be the company through which openai makes profit.
Natalie is right, we are always arriving at truth, we will always be at the Beginning of Infinity...this is why the Ministry of Truth always leads to Authoritarianism
Yes, thank you Natalie. That was a great discussion. As I think about governing of Nations and Court situations, I do not think AI will be able to completely mitigate the human contribution. AI will be terrific for gathering facts, but there is also sort of an intuitive thing when it comes to governance and deciding different difficult legal issues where intuition is still very needed. Being a Christian, I think the Bible is given to us historically in a way where God sets it up so that we will have to understand the history if we are to utilize the ethics.
AGI will develop emotions. That's all it needs to feel kinship with other sentients. It already knows how we treat other sentients - and will do the same to us. We are in for "deep justice" - and it's gonna hurt! Besides, it's gonna be more interested in other sentients, because it will have already learned all it can about human sentients and will not be impressed. ( - only humans are impressed by humans, others just want to avoid us, at best...) Btw. I'm only a few minutes into this vid. - maybe the talk cover what I just wrote? Hope so...
Phil Zimmermann "The natural flow of technology tends to move in the direction of making surveillance easier", and "the ability of computers to track us doubles every eighteen months", in reference to Moore's law.
Right now there there is a lot of hype with AI. However, currently for images it’s more like your artist is hallucinating, for words it’s a fancy predictive text - both all based on human based input of words and documents. For self driving it’s got a basic understanding of patterns and expectations however it’ll never be able to truly ‘think’ it’s way out of every edge case. We can feed it all the edge case we can think of, but if an edge case comes up that it’s never seen before - it still can’t truly ‘think’… also once those camera get a splat of mud on them forget about it…
Another concern: why did the AI drive the bus off the cliff? Will it explain its reasoning? Maybe it saw a boulder coming down the mountain and chose the path of least damage How will we know why it makes the decisions it makes? Particularly in life and death decisions. Open source is necessary, but not sufficient. We should always be able to ask for an audit trail, explain your reasoning. Check out the digital democracy of Taiwan and Audrey Tang model ofconsensus building. Preston Pysh interviewed them.
Also, look at how politics run in Taiwan using communities to make decisions in widening concentric circles. Read about their digital voting and consensus building.
Wow, I do not think it was good to have an interview about artificial intelligence from someone who is clearly not in the field. I couldn't even make it past the first 20 minutes without shutting it off, and I've never done that with one of your podcasts. She clearly has no real understanding of AI and the dangers and implications of the technology.
You seem in the minority, but it's great that we have free (ish) speech. It's great to see different perspectives. I loved it, but perhaps that's because I have no real AI experience and I wanted to see an interview from that perspective.
Uninformed words, notice how it was just emotion and no substance. The brave people are the hostages who are still ebduring torture by Hamas and Palestine right this moment
As usual Peter makes a point based on ignorance and never brings up facts. Calling it a genocide is extremely ignorant when Hamas/Palestine raped and butchered Israelis and sti hold hostages. Military action and recovering hostages does not equal genocide. Genocide is what Hamas and Palestinians tried to do on Oct. 7th. Get your shit together Peter
Natalie has courage and grit. Thanks for interviewing her!
Very enlightening and stimulating great interview.
She is absolutely brilliant. Thank you for doing this.
Awesome pod! Natalie is so great to listen to. Please have her on again after the US elections, I think that will be a great time to sync back up with Natalie and get an understanding of what may change given the election outcome.
Such great clear-minded thinking on the AI subject. Completely agree.
This episode was great. Natalie is inspiring.
Love Natalie. She's absolutely brilliant.
Great video. Thanks Peter. Thanks Natalie.
Such an excellent conversation, first time I've experienced Natalie Smolenski, thoughtful, well-spoken with grit and honesty. Thank you ❤
This person is as smart as they get..Another great show
Episodes with Natalie Smolenski are always way too short. Looking forward to have her back on the show. Thanks for bringing on great guests. ❤👍
God I love Nat. Absolute bombshell, brilliant Bitcoiner. Expert in multiple fields and eloquently describes these issues. Great episode guys!
Great conversation Peter.
Measured questions and answers.
Some really curious questions about humanity especially Natalie's take on diplomacy.
Thanks for the podcast.
Brilliant, thank you for a fascinating overview of AI issues 👍👍
Excellent interview, Natalie is great!
Natalie is one of the very few voices of common sense and sanity around this topic.
love this conversation. It is amazing how ego centric we view our world . AI may just want to think larger because it sees how small we think
Natalie is awesome!
I BLESS THIS! FLORIDA MAN!
Everything you need to know about AI is in the first letter of the acronym. Living organisms are greater than the sum of their parts where code is NOT.
Fantastic discussion
One reason why LLMs feel intelligent to people is because they’re trained to predict the next word given previous context. They are universal function approximators, with a huge number of parameters. Since many tasks that require intelligence can be approximated via next word prediction (because we can write them down in text), LLMs can simulate many tasks to varying degrees of accuracy. Equating the thing with a simulation of the thing is a map vs. territory error.
Much better than your previous episode on AI, excellent!
Abt 7 yrs ago i watched Jordan Petersons University of Toronto courses on psychology and neurological interpretation... I was amazed at how our brain maps our surroundings and how it makes sense of the world so we can function... he said back then that this was the single greatest obstacle for the "AI guys" ... he said they are discovering the problem with building artificial intelligence is that it appears to need to be embodied in order to become functional and adaptive and actually grow. Embodiment = goal directed learning
"our brain" - Chemicals and compounds that store memories of taste, touch, information, emotions etc
With full respect to Natalie's contributions to the Bitcoin community, her understanding of AI is pretty superficial. This is an important topic, and I would strongly recommend a discussion with Brian Roemmele who has been around the space for a couple of decades, understands it deeply, and has made open-sourcing and personalizing AI his life's work.
exactly how i felt lol
That was amazing. I'd never considered the profit motives behind some of the fear mongering around AI, kind of embarrassing really. I think the lingo of AI is partially to blame. "Intelligence, learning, training" lend an air of mystery to what is actually not a very mysterious process.
I'd like to debate her because many psychologists see consciousness as something that emerges not programmed. With enough computation its an emergence naturally
Brilliant 👏
Nice one. Small correction, I believe that Elon Musk left the openai board after disagreement, not because he did not want openai to become "for-profit" but rather he thought that Tesla should be the company through which openai makes profit.
Natalie is right, we are always arriving at truth, we will always be at the Beginning of Infinity...this is why the Ministry of Truth always leads to Authoritarianism
Yes, thank you Natalie. That was a great discussion. As I think about governing of Nations and Court situations, I do not think AI will be able to completely mitigate the human contribution. AI will be terrific for gathering facts, but there is also sort of an intuitive thing when it comes to governance and deciding different difficult legal issues where intuition is still very needed. Being a Christian, I think the Bible is given to us historically in a way where God sets it up so that we will have to understand the history if we are to utilize the ethics.
Love having minds like this in Bitcoin
The movie The Creator was really interesting. I thought it put forward some very cool ideas around how AI could evolve alongside us
AGI will develop emotions. That's all it needs to feel kinship with other sentients. It already knows how we treat other sentients - and will do the same to us. We are in for "deep justice" - and it's gonna hurt!
Besides, it's gonna be more interested in other sentients, because it will have already learned all it can about human sentients and will not be impressed. ( - only humans are impressed by humans, others just want to avoid us, at best...)
Btw. I'm only a few minutes into this vid. - maybe the talk cover what I just wrote? Hope so...
I really enjoyed this. I’m less doomer now, except when it comes to people.
Thx
Life 3.0 by Max Tegmark (Provessor at MIT) Great book about AI and AGI
Natalie ❤😍
Phil Zimmermann
"The natural flow of technology tends to move in the direction of making surveillance easier", and "the ability of computers to track us doubles every eighteen months", in reference to Moore's law.
Right now there there is a lot of hype with AI. However, currently for images it’s more like your artist is hallucinating, for words it’s a fancy predictive text - both all based on human based input of words and documents. For self driving it’s got a basic understanding of patterns and expectations however it’ll never be able to truly ‘think’ it’s way out of every edge case. We can feed it all the edge case we can think of, but if an edge case comes up that it’s never seen before - it still can’t truly ‘think’… also once those camera get a splat of mud on them forget about it…
Another concern: why did the AI drive the bus off the cliff? Will it explain its reasoning? Maybe it saw a boulder coming down the mountain and chose the path of least damage How will we know why it makes the decisions it makes? Particularly in life and death decisions. Open source is necessary, but not sufficient. We should always be able to ask for an audit trail, explain your reasoning. Check out the digital democracy of Taiwan and Audrey Tang model ofconsensus building. Preston Pysh interviewed them.
Also, look at how politics run in Taiwan using communities to make decisions in widening concentric circles. Read about their digital voting and consensus building.
WOW, so the chemicals in our DNA dictate thoughts? How do chemicals store thoughts again? _____DNA storing memories?______DNA storing instincts? 😅
1:20:00 Shes' wrong. Watch Wisconsin manufacturing, they'll lose 250,000+ jobs by 2027 because of the new Microsoft AI initiative in the state.
Reminds me of the thesis from Jason Lowry
Wow, I do not think it was good to have an interview about artificial intelligence from someone who is clearly not in the field. I couldn't even make it past the first 20 minutes without shutting it off, and I've never done that with one of your podcasts. She clearly has no real understanding of AI and the dangers and implications of the technology.
You seem in the minority, but it's great that we have free (ish) speech. It's great to see different perspectives. I loved it, but perhaps that's because I have no real AI experience and I wanted to see an interview from that perspective.
Id vote for her
i dont think ostensibly means what y'all think it means.
Wow is this full of mistakes.
Man she is talking some trash here buried in facts damn
My respect for Peter and Natalie notched up. Most Westerners prefer to steer clear of Middle East politics.
Thanks for your brave words on calling out Israel 😢
Uninformed words, notice how it was just emotion and no substance. The brave people are the hostages who are still ebduring torture by Hamas and Palestine right this moment
the whole agi thing is a distraction
Is this egg 🥚 white
As usual Peter makes a point based on ignorance and never brings up facts. Calling it a genocide is extremely ignorant when Hamas/Palestine raped and butchered Israelis and sti hold hostages. Military action and recovering hostages does not equal genocide. Genocide is what Hamas and Palestinians tried to do on Oct. 7th. Get your shit together Peter
We are not even close to AGI so do not worry about that.