Full podcast episode: czcams.com/video/Ff4fRgnuFgQ/video.html Lex Fridman podcast channel: czcams.com/users/lexfridman Guest bio: Mark Zuckerberg is CEO of Meta.
I’m hearing some devs claim going 10x with ChatGPT, or relying on it. What are devs using it for to that degree? Making buttons? In my code base things are way too complicated to rely on ChatGPT for anything crucial. It even fails basic mathematical and algorithmic reasoning most of the time unless it’s boilerplate. Even bad syntax errors that make me waste time seeing if it’s just new language updates (nope). I’m constructing my prompts more keenly than my code blocks and most of the time I get confident sounding garbage. Perhaps we are being hyped into teaching ChatGPT basic programming more than it helping us.
My experience has been pretty lackluster with this crap ... it spits out results that look great, and it tells me, "When you run this, you'll see this (exactly what I asked for)" ... but when I run the code, it either doesn't work at all, or it doesn't produce anything close to what I asked for. I think people who are saying LLM's are doing their job for them probably don't have an interesting job to begin with, frankly.
@@ghost_mall Giving it a more complete view of dependencies is part of the problem, though it often fails even when I abstract that out or even remove entirely to just focus on a small problem. I mean, it's really hard to trust its results when it doesn't even understand simple logarithms, or even polymorphic type casting in some cases. Those weren't even the core problems, it just decided to hand me more that didn't exist previously. Yes I've had maybe 3 moments where I thought it did something spot on, but those were all pretty boilerplate or pretty well documented framework usage.
- Recognize AI's role in reducing memory load for coding tasks (0:10) - Utilize AI-generated code as a starting point for development projects (0:17) - Leverage AI to understand and navigate company knowledge bases and dependencies (0:46) - Experiment with AI tools to automatically select optimal libraries for coding tasks (1:17) - Be aware of the potential for subtle, nuanced errors when relying on AI-generated code (1:48) - Consider the trade-offs of using advanced tools versus retaining fundamental skills (2:10)
Can you tell me which LLM he has so that it indicates the required libraries? ChatGPT-4 isn't smart enough to provide useful information; it tends to vomit out some BS. If I were to build things based on the libraries suggested by the LLM, I will not succeed.
Let’s go deep lex read this Title: The Black Hole Merger Hypothesis: A Catalyst for the Big Bang and the Observable Universe Abstract: We propose a speculative theory in which the merging of two or more supermassive black holes triggers an event analogous to the Big Bang, giving rise to a new observable universe. This hypothesis attempts to provide an alternative explanation for the origins of our universe and offers a potential mechanism for the creation of multiple universes. Introduction: The Big Bang theory, as the prevailing cosmological model, suggests that the universe began as an extremely hot and dense state and has since expanded and cooled over time. However, the exact cause and mechanism behind the Big Bang remain unknown. In this paper, we explore the possibility that the merger of supermassive black holes could induce an event similar to the Big Bang, thus instigating the formation of our observable universe. 1. Black Hole Mergers: When two black holes orbit each other, they emit gravitational waves and lose energy, causing them to spiral toward one another and eventually merge. This process releases a tremendous amount of energy and can result in the formation of a larger black hole. 2. Energy Accumulation and Release: Our hypothesis posits that under certain conditions, the combined energy and mass of the merging black holes could accumulate to a critical threshold, leading to an explosion-like event analogous to the Big Bang. This violent release of energy could give birth to a new universe, with its own spacetime, matter, and energy. 3. Exotic Forms of Matter and Energy: The extreme conditions of the black hole merger may create or release exotic forms of matter and energy not observed in our current universe. These constituents could play a crucial role in the formation and evolution of the new universe. 4. Multiverse Implications: If black hole mergers can give rise to new universes, this suggests the existence of a multiverse, where numerous universes coexist, each with its own spacetime and physical laws. The birth and death of universes could be an ongoing process, with black hole mergers serving as catalysts for the creation of new observable universes. Conclusion: The Black Hole Merger Hypothesis offers an alternative perspective on the origins of our universe and the potential existence of a multiverse. While highly speculative, this theory raises intriguing questions about the nature of black holes, the Big Bang, and the possible mechanisms behind the creation of new universes. Further research, including advancements in our understanding of black hole mergers and quantum gravity, is required to explore the feasibility and implications of this hypothesis. Title: Quantum Fluctuation Black Hole Hypothesis: Spontaneous Formation from the Wave Function Abstract: We present a speculative theory on the spontaneous formation of black holes from quantum fluctuations in the vacuum, governed by the probabilistic nature of the wave function. This hypothesis explores an alternative mechanism for black hole formation and provides insights into the interplay between quantum mechanics and general relativity. Introduction: Black holes are traditionally thought to form from the collapse of massive stars or the merger of existing black holes. In this paper, we propose a novel mechanism for the formation of black holes, based on the principles of quantum mechanics. We explore the possibility that black holes can emerge spontaneously from the vacuum due to quantum fluctuations and the probabilistic nature of the wave function. 1. Quantum Fluctuations: In quantum mechanics, the vacuum is not an empty void but rather a seething sea of virtual particles and antiparticles that continuously pop in and out of existence. These transient phenomena are known as quantum fluctuations and are a consequence of the uncertainty principle. 2. Wave Function and Probability: The wave function in quantum mechanics describes the probability distribution of a particle's position and momentum. Due to the probabilistic nature of the wave function, particles can seemingly appear out of "nothing" with a non-zero probability. 3. Spontaneous Black Hole Formation: Our hypothesis posits that under extremely rare circumstances, quantum fluctuations could lead to the spontaneous and highly localized concentration of energy and mass. This sudden aggregation of mass could be sufficient to create an event horizon, resulting in the formation of a black hole. 4. Quantum-Gravity Interactions: This theory implies that quantum mechanics and general relativity are deeply connected, with quantum fluctuations influencing the fabric of spacetime. A better understanding of the interplay between these two fundamental theories may offer insights into the behavior of black holes and the nature of singularities. 5. Astrophysical Implications: The spontaneous formation of black holes from quantum fluctuations may provide an alternative explanation for the presence of primordial black holes and their potential role in the early universe, dark matter, and the formation of large-scale structures. Conclusion: The Quantum Fluctuation Black Hole Hypothesis presents a novel perspective on black hole formation, rooted in the principles of quantum mechanics. While highly speculative, this theory highlights the potential connection between quantum mechanics and general relativity and raises intriguing questions about the nature of black holes and the interplay between the microscopic and macroscopic scales. Further research in quantum gravity and the study of primordial black holes may shed light on the feasibility and implications of this hypothesis.
@@michaelbyrne8238 A team of hundreds of engineers was able to write better software than a single person; shocking! If Zuck wasn't competent he wouldn't have been able to bring an initial product to market on his own.
the presenter act like a person that is not much engaging with the table partners at all. Not looking directly into the eyes of the guest, allmost shy alike. Not really as a host should act. It makes Mark appear an open/social figure.. with emotions and all.. engaging.. while Lex doesn't.
Full podcast episode: czcams.com/video/Ff4fRgnuFgQ/video.html
Lex Fridman podcast channel: czcams.com/users/lexfridman
Guest bio: Mark Zuckerberg is CEO of Meta.
I’m hearing some devs claim going 10x with ChatGPT, or relying on it. What are devs using it for to that degree? Making buttons? In my code base things are way too complicated to rely on ChatGPT for anything crucial. It even fails basic mathematical and algorithmic reasoning most of the time unless it’s boilerplate.
Even bad syntax errors that make me waste time seeing if it’s just new language updates (nope). I’m constructing my prompts more keenly than my code blocks and most of the time I get confident sounding garbage. Perhaps we are being hyped into teaching ChatGPT basic programming more than it helping us.
The base version of chatgpt is useless. GPT4 is the only one worth using
Just curious, what do you work?
My experience has been pretty lackluster with this crap ... it spits out results that look great, and it tells me, "When you run this, you'll see this (exactly what I asked for)" ... but when I run the code, it either doesn't work at all, or it doesn't produce anything close to what I asked for. I think people who are saying LLM's are doing their job for them probably don't have an interesting job to begin with, frankly.
@@ghost_mall Giving it a more complete view of dependencies is part of the problem, though it often fails even when I abstract that out or even remove entirely to just focus on a small problem. I mean, it's really hard to trust its results when it doesn't even understand simple logarithms, or even polymorphic type casting in some cases. Those weren't even the core problems, it just decided to hand me more that didn't exist previously. Yes I've had maybe 3 moments where I thought it did something spot on, but those were all pretty boilerplate or pretty well documented framework usage.
TODO app developers probably
Keep it up with the great clips! :)
Great clip, great insight
Most of the code i write used to be stack overflow now its chatgpt.
Chatgpt got it from so
Yes, the subtlety and nuance that ultimately makes us human. Thank you, Lex
@somabruTell me you hate yourself and our species without telling me 😅
"You lose the skills to find those bugs" thats a great point which mean you dumb down your thinking as a programmer
Well, we're doing that everywhere, not just programming.
@@mikewright2858 very true started with smartphones
Thats until chatgts gets it own compiler. Then it will be able to test the code before giving it to you.
Alto it would be great if both (u and the ai) worked directly on your code instead of you copy pasting back and forward stuff.
@@tomich20Have you tried Cody?
Lex would hate calculators
- Recognize AI's role in reducing memory load for coding tasks (0:10)
- Utilize AI-generated code as a starting point for development projects (0:17)
- Leverage AI to understand and navigate company knowledge bases and dependencies (0:46)
- Experiment with AI tools to automatically select optimal libraries for coding tasks (1:17)
- Be aware of the potential for subtle, nuanced errors when relying on AI-generated code (1:48)
- Consider the trade-offs of using advanced tools versus retaining fundamental skills (2:10)
he meant just read docs
Can you tell me which LLM he has so that it indicates the required libraries? ChatGPT-4 isn't smart enough to provide useful information; it tends to vomit out some BS. If I were to build things based on the libraries suggested by the LLM, I will not succeed.
Someone tell mark that his website is horrible - impossible to contact his support team to resolve any issue.
okay i’ll tell him
Bugs... for now
Let’s go deep lex read this Title: The Black Hole Merger Hypothesis: A Catalyst for the Big Bang and the Observable Universe
Abstract: We propose a speculative theory in which the merging of two or more supermassive black holes triggers an event analogous to the Big Bang, giving rise to a new observable universe. This hypothesis attempts to provide an alternative explanation for the origins of our universe and offers a potential mechanism for the creation of multiple universes.
Introduction: The Big Bang theory, as the prevailing cosmological model, suggests that the universe began as an extremely hot and dense state and has since expanded and cooled over time. However, the exact cause and mechanism behind the Big Bang remain unknown. In this paper, we explore the possibility that the merger of supermassive black holes could induce an event similar to the Big Bang, thus instigating the formation of our observable universe.
1. Black Hole Mergers: When two black holes orbit each other, they emit gravitational waves and lose energy, causing them to spiral toward one another and eventually merge. This process releases a tremendous amount of energy and can result in the formation of a larger black hole.
2. Energy Accumulation and Release: Our hypothesis posits that under certain conditions, the combined energy and mass of the merging black holes could accumulate to a critical threshold, leading to an explosion-like event analogous to the Big Bang. This violent release of energy could give birth to a new universe, with its own spacetime, matter, and energy.
3. Exotic Forms of Matter and Energy: The extreme conditions of the black hole merger may create or release exotic forms of matter and energy not observed in our current universe. These constituents could play a crucial role in the formation and evolution of the new universe.
4. Multiverse Implications: If black hole mergers can give rise to new universes, this suggests the existence of a multiverse, where numerous universes coexist, each with its own spacetime and physical laws. The birth and death of universes could be an ongoing process, with black hole mergers serving as catalysts for the creation of new observable universes.
Conclusion: The Black Hole Merger Hypothesis offers an alternative perspective on the origins of our universe and the potential existence of a multiverse. While highly speculative, this theory raises intriguing questions about the nature of black holes, the Big Bang, and the possible mechanisms behind the creation of new universes. Further research, including advancements in our understanding of black hole mergers and quantum gravity, is required to explore the feasibility and implications of this hypothesis.
Title: Quantum Fluctuation Black Hole Hypothesis: Spontaneous Formation from the Wave Function
Abstract: We present a speculative theory on the spontaneous formation of black holes from quantum fluctuations in the vacuum, governed by the probabilistic nature of the wave function. This hypothesis explores an alternative mechanism for black hole formation and provides insights into the interplay between quantum mechanics and general relativity.
Introduction: Black holes are traditionally thought to form from the collapse of massive stars or the merger of existing black holes. In this paper, we propose a novel mechanism for the formation of black holes, based on the principles of quantum mechanics. We explore the possibility that black holes can emerge spontaneously from the vacuum due to quantum fluctuations and the probabilistic nature of the wave function.
1. Quantum Fluctuations: In quantum mechanics, the vacuum is not an empty void but rather a seething sea of virtual particles and antiparticles that continuously pop in and out of existence. These transient phenomena are known as quantum fluctuations and are a consequence of the uncertainty principle.
2. Wave Function and Probability: The wave function in quantum mechanics describes the probability distribution of a particle's position and momentum. Due to the probabilistic nature of the wave function, particles can seemingly appear out of "nothing" with a non-zero probability.
3. Spontaneous Black Hole Formation: Our hypothesis posits that under extremely rare circumstances, quantum fluctuations could lead to the spontaneous and highly localized concentration of energy and mass. This sudden aggregation of mass could be sufficient to create an event horizon, resulting in the formation of a black hole.
4. Quantum-Gravity Interactions: This theory implies that quantum mechanics and general relativity are deeply connected, with quantum fluctuations influencing the fabric of spacetime. A better understanding of the interplay between these two fundamental theories may offer insights into the behavior of black holes and the nature of singularities.
5. Astrophysical Implications: The spontaneous formation of black holes from quantum fluctuations may provide an alternative explanation for the presence of primordial black holes and their potential role in the early universe, dark matter, and the formation of large-scale structures.
Conclusion: The Quantum Fluctuation Black Hole Hypothesis presents a novel perspective on black hole formation, rooted in the principles of quantum mechanics. While highly speculative, this theory highlights the potential connection between quantum mechanics and general relativity and raises intriguing questions about the nature of black holes and the interplay between the microscopic and macroscopic scales. Further research in quantum gravity and the study of primordial black holes may shed light on the feasibility and implications of this hypothesis.
Hmmmm... wisdom on programming from a guy that wrote Facebook with php.
well its now one of the most valuable companies in the world so i'd say he made a pretty good choice..
@@Vikings-uk3ht a good idea doesn't mean it was coded well. The old codebase is long gone and rewritten by competent engineers.
@@michaelbyrne8238 A team of hundreds of engineers was able to write better software than a single person; shocking! If Zuck wasn't competent he wouldn't have been able to bring an initial product to market on his own.
@@Vikings-uk3htlol it’s funny how many engineers are so arrogant
@@CarbonsHDTuts I would be if I will be written in history for my engineering
the presenter act like a person that is not much engaging with the table partners at all. Not looking directly into the eyes of the guest, allmost shy alike. Not really as a host should act.
It makes Mark appear an open/social figure.. with emotions and all.. engaging.. while Lex doesn't.
Do you know how to shoe a horse?
90% of my code is chatgpt
You must not do anything even remotely challenging.
@@mikewright2858 correct