Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

Just read a neat AI paper called SAMURAI -- it takes SAM 2 (Meta's "segment anything" model) and makes it way better at tracking objects in videos. Basic problem is SAM 2 gets confused when things move fast or there's a crowd of similar objects (big problem for VFX... and video intelligence alike). The fix? They basically gave it a sense of motion and smarter memory selection using our old friend the Kalman filter -- helps it remember where things were going instead of just where they are. No retraining needed and runs in realtime. Elegant solution that matches (and even beats!) a lot of purpose built tracking systems. Nice example of how sometimes you don't need fancy new architectures -- just smart ways to use what the model already knows, plus some tried & tested classical methods. Every time someone's like "we solved tracking!" I check the methods section and... yep, there's our boy Kalman, still crushing it after all these years. Sometimes the old school stuff just works.show more

Bilawal Sidhu

111,685 subscribers

47,291 görüntüleme • 1 yıl önce •via X (Twitter)

Sanat Bilim & Teknoloji Eğitim

Anya Rossi• Live Now

Private livecam show

9 Yorum

Bilawal Sidhu profil fotoğrafı

Bilawal Sidhu1 yıl önce

Project/paper here: If y'all enjoyed this post, you might also enjoy my writing on all things spatial ai and creative tech:

Min Choi profil fotoğrafı

Min Choi1 yıl önce

This is genius. I'm playing with SAM 2 and this was the exact limitation that I was running into for my use

Bilawal Sidhu profil fotoğrafı

Bilawal Sidhu1 yıl önce

Same! @BlacePlugins ya'll need to implement this paper for your after effects SAM 2 plugin

Kishen Patel profil fotoğrafı

Kishen Patel1 yıl önce

very cool! there’s definitely something satisfying when intuitive ways of solving ML problems work. like kalman filter to track noisy movement just makes sense - basically a signal processing problem. another might be SSMs for modeling a differential system.

Bilawal Sidhu profil fotoğrafı

Bilawal Sidhu1 yıl önce

Totally — sometimes the best solutions come from recognizing when an ML problem is actually a classical problem in disguise!

subnet Sam profil fotoğrafı

subnet Sam1 yıl önce

wow this is incredibly cool! i woudl use this for non nefarious purposes

Bilawal Sidhu profil fotoğrafı

Bilawal Sidhu1 yıl önce

Lol. VFX and surveillance about to get a nice boost 😂

Maskai profil fotoğrafı

Maskai1 yıl önce

Very interesting 🧐

Wit Of Gold profil fotoğrafı

Wit Of Gold1 yıl önce

Impressive

Benzer Videolar

HOLY CR*P 🚨 Minnesota Rep Krista Knudsen says they just found out the state TURNED OFF tracking for money being sent overseas They turned it on for 2 weeks. Saw all the money leaving, and Democrats quickly SHUT OFF TRACKING to allow it to keep happening READ THAT AGAIN “Today in the fraud committee, the Department of Human Services testified that they haven't been tracking the IP addresses of money that's being literally flown out of our state and where it's going. They did turn on that reporting for 2 weeks, and they found that yes, a lot of those IP addresses were tied to overseas computers, and then they quickly shut it off after just 2 weeks of seeing where this money is actually going” “Every time we turn around, it's just more and more fraud scandals in our state”

HOLY CR*P 🚨 Minnesota Rep Krista Knudsen says they just found out the state TURNED OFF tracking for money being sent overseas They turned it on for 2 weeks. Saw all the money leaving, and Democrats quickly SHUT OFF TRACKING to allow it to keep happening READ THAT AGAIN “Today in the fraud committee, the Department of Human Services testified that they haven't been tracking the IP addresses of money that's being literally flown out of our state and where it's going. They did turn on that reporting for 2 weeks, and they found that yes, a lot of those IP addresses were tied to overseas computers, and then they quickly shut it off after just 2 weeks of seeing where this money is actually going” “Every time we turn around, it's just more and more fraud scandals in our state”

Wall Street Apes

1,269,747 görüntüleme • 6 ay önce

Engram cofounder Jack Morris just raised $98M to build a new type of AI. He says models don't need to get smarter over time. Instead, they just need to know you better and better over time. Jack describes what he's building: "Our product is a new type of AI. We have a pretty different vision from a lot of the frontier labs, which are working on one model per lab, and trying to make that model smarter every month." "There's another way to think about it, which is that the model doesn't need to get smarter every month. It needs to know you better." "So we're working on a whole different stack, which is a way to train models that train themselves to know your world better and adjust to the things that you say." "So: new ways of training, new ways of running the models."

Engram cofounder Jack Morris just raised $98M to build a new type of AI. He says models don't need to get smarter over time. Instead, they just need to know you better and better over time. Jack describes what he's building: "Our product is a new type of AI. We have a pretty different vision from a lot of the frontier labs, which are working on one model per lab, and trying to make that model smarter every month." "There's another way to think about it, which is that the model doesn't need to get smarter every month. It needs to know you better." "So we're working on a whole different stack, which is a way to train models that train themselves to know your world better and adjust to the things that you say." "So: new ways of training, new ways of running the models."

TBPN

70,704 görüntüleme • 3 ay önce

Chamath: Two terms you need to pay attention to in AI are Prefill and Decode “There's two terms that I think you're going to hear a ton about over these next few years.” “The first term is prefill, and the next is decode.” “What prefill and decode are, are two very distinct ways of how models think, and how a model goes through the process of answering a question that you ask it.” “And so when you send a prompt to AI, what happens is that the model processes it. This is called the reading phase or prefill.” “It reads your entire prompt all at once. And then it does a bunch of math, calculates all these relationships between all the words, and it stores them in temporary memory.” “The problem is that this is really compute bound. So it requires massive brute force. And Nvidia GPUs crush here.” “And their architecture is designed for massive parallel processing, which makes them really amazing at digesting these long prompts.” “So the problem just gets bigger and bigger, Nvidia just completely dominates.” “But the next phase though, this critical phase, the decode phase, is the writing phase, right?” “So the model starts to generate a response, you ask it a question and its response, one token at a time.” “And then to pick the next token to pick the next word, it has to look back at everything it has said already so that it doesn't hallucinate.” “The problem is that this is incredibly memory bandwidth constrained.” “And in our architecture, a long time ago, we made these design decisions from day one.” “And so what we did was we took a very different architectural approach, we took a very conservative process technology. We weren't pushing the boundaries of physics.” “And we used a lot of what's called SRAM. So memory on the chip so that we could do this decode thing as well or better than everybody else.” “And so now when you put these two things together, I just think it's going to create a huge acceleration in the ability for this entire infrastructure layer to get much cheaper and much more valuable, which I suspect then it'll have a lot more developer pull, you'll get a lot more applications being built, billions and billions of more people using it.”

Chamath: Two terms you need to pay attention to in AI are Prefill and Decode “There's two terms that I think you're going to hear a ton about over these next few years.” “The first term is prefill, and the next is decode.” “What prefill and decode are, are two very distinct ways of how models think, and how a model goes through the process of answering a question that you ask it.” “And so when you send a prompt to AI, what happens is that the model processes it. This is called the reading phase or prefill.” “It reads your entire prompt all at once. And then it does a bunch of math, calculates all these relationships between all the words, and it stores them in temporary memory.” “The problem is that this is really compute bound. So it requires massive brute force. And Nvidia GPUs crush here.” “And their architecture is designed for massive parallel processing, which makes them really amazing at digesting these long prompts.” “So the problem just gets bigger and bigger, Nvidia just completely dominates.” “But the next phase though, this critical phase, the decode phase, is the writing phase, right?” “So the model starts to generate a response, you ask it a question and its response, one token at a time.” “And then to pick the next token to pick the next word, it has to look back at everything it has said already so that it doesn't hallucinate.” “The problem is that this is incredibly memory bandwidth constrained.” “And in our architecture, a long time ago, we made these design decisions from day one.” “And so what we did was we took a very different architectural approach, we took a very conservative process technology. We weren't pushing the boundaries of physics.” “And we used a lot of what's called SRAM. So memory on the chip so that we could do this decode thing as well or better than everybody else.” “And so now when you put these two things together, I just think it's going to create a huge acceleration in the ability for this entire infrastructure layer to get much cheaper and much more valuable, which I suspect then it'll have a lot more developer pull, you'll get a lot more applications being built, billions and billions of more people using it.”

The All-In Podcast

567,660 görüntüleme • 9 ay önce

"I think among a lot of Democrats there's this kind of what psychologists would call 'learned helplessness' where you just think, 'Well we can't do anything about it. We have to wait till 2029' ... In New York with some people it's just like, 'Well we just have to do our things in New York and not worry about the rest of the country.' It does not work that way. We are one country. Our democracy can rise or fall on what this man is doing. And, it's not just our democracy. ... It directly affects our lives. The stability of our economy depends on the belief in the rule of law that people think we have a stable legal system, and it's hard to believe that today"

"I think among a lot of Democrats there's this kind of what psychologists would call 'learned helplessness' where you just think, 'Well we can't do anything about it. We have to wait till 2029' ... In New York with some people it's just like, 'Well we just have to do our things in New York and not worry about the rest of the country.' It does not work that way. We are one country. Our democracy can rise or fall on what this man is doing. And, it's not just our democracy. ... It directly affects our lives. The stability of our economy depends on the belief in the rule of law that people think we have a stable legal system, and it's hard to believe that today"

George Conway ⚖️🇺🇸

80,959 görüntüleme • 4 ay önce

🚨David Friedberg: AI is starting to identify and solve problems on its own “I'll give you a science corner example: there's this Evo 2 model that they publish at the Arc Institute, which Patrick Collison, you know, is the main funder and chairman.” “So that Evo 2 model, they just ingested all the DNA data they could find in the world.” “Trillions and trillions of base paired data that they ingested and then they looked at patterns in DNA. And that's it.” “They had no context for what the DNA represented, they had no context for the concept of genes, none of the structured understanding of what that DNA does, what it is, and you know what it did?” “They fed in the BRCA gene variant and the thing output a warning saying, ‘I think that this is a pathogenic variant to DNA,’ without having any context.” “This is the breast cancer allele.” “And it didn't have any knowledge and it wasn't trained on that at all.” “It had no knowledge that there are pathogenic variants for cancer, and it identified that this was a genetic variant that can cause some sort of pathogenic outcome in the organism.” “That's a great example where there's a lack of understanding at the human level on what really drives some of the patterns in nature, the patterns in society, the patterns in behavior that are kind of emergent phenomena perhaps, that these AI models are starting to identify.”

🚨David Friedberg: AI is starting to identify and solve problems on its own “I'll give you a science corner example: there's this Evo 2 model that they publish at the Arc Institute, which Patrick Collison, you know, is the main funder and chairman.” “So that Evo 2 model, they just ingested all the DNA data they could find in the world.” “Trillions and trillions of base paired data that they ingested and then they looked at patterns in DNA. And that's it.” “They had no context for what the DNA represented, they had no context for the concept of genes, none of the structured understanding of what that DNA does, what it is, and you know what it did?” “They fed in the BRCA gene variant and the thing output a warning saying, ‘I think that this is a pathogenic variant to DNA,’ without having any context.” “This is the breast cancer allele.” “And it didn't have any knowledge and it wasn't trained on that at all.” “It had no knowledge that there are pathogenic variants for cancer, and it identified that this was a genetic variant that can cause some sort of pathogenic outcome in the organism.” “That's a great example where there's a lack of understanding at the human level on what really drives some of the patterns in nature, the patterns in society, the patterns in behavior that are kind of emergent phenomena perhaps, that these AI models are starting to identify.”

The All-In Podcast

79,718 görüntüleme • 1 yıl önce

I've been obsessing over Runway Gen-4.5 this past week. I tested it quite extensively, and I noticed a few things that should be fundamental to any video model: 1. Amazing spatial awareness. It just knows what comes next in the scene and renders it perfectly, with the correct distance, geometry, and style. 2. It understands the space and time. For example, how a character interacts with the environment, even if you don't prompt for it: High altitude > Windy. 3. This spatial sense of motion and place makes it look natural and cinematic. Regarding this last point, I've said it many times here on X that Runway needs to get out of the slow-motion era, and in the end, they delivered massively on it. It's just beautiful to watch scenes come to life naturally. Now, we just need sound and 1080p. Cristóbal Valenzuela take your time, we're here.

I've been obsessing over Runway Gen-4.5 this past week. I tested it quite extensively, and I noticed a few things that should be fundamental to any video model: 1. Amazing spatial awareness. It just knows what comes next in the scene and renders it perfectly, with the correct distance, geometry, and style. 2. It understands the space and time. For example, how a character interacts with the environment, even if you don't prompt for it: High altitude > Windy. 3. This spatial sense of motion and place makes it look natural and cinematic. Regarding this last point, I've said it many times here on X that Runway needs to get out of the slow-motion era, and in the end, they delivered massively on it. It's just beautiful to watch scenes come to life naturally. Now, we just need sound and 1080p. Cristóbal Valenzuela take your time, we're here.

Alex Patrascu

16,141 görüntüleme • 8 ay önce

Creator of C#, Anders Hejlsberg(Anders Hejlsberg), on the design goals that shaped the language and why standardisation was a goal from the beginning: “The overarching thing was this power and productivity of C++ with the ease of use of Visual Basic. We knew we wanted to build an object oriented language. We wanted managed code or byte code so we could target different runtime environments. We wanted garbage collection and exception handling, but also things like a unified object system where, and that's true in C#, anything can be assigned to an object. If it's a value type, we box it. And it's a self-describing object. So, reflection: you can ask an object, ‘what are you’? And you can get all of the facts about it at runtime, and you can dynamically manipulate it in ways that just don't exist in a lot of other environments. We wanted a language that made this new model of properties, methods and events first class because that was how components were built as opposed to just sort of functions and procedures and even objects. And we wanted to create a language that was standardised. We wanted to give this language to a standardisation committee and try to level the playing field there. And all of those things were sort of like what was rolled up in C#”

Creator of C#, Anders Hejlsberg(Anders Hejlsberg), on the design goals that shaped the language and why standardisation was a goal from the beginning: “The overarching thing was this power and productivity of C++ with the ease of use of Visual Basic. We knew we wanted to build an object oriented language. We wanted managed code or byte code so we could target different runtime environments. We wanted garbage collection and exception handling, but also things like a unified object system where, and that's true in C#, anything can be assigned to an object. If it's a value type, we box it. And it's a self-describing object. So, reflection: you can ask an object, ‘what are you’? And you can get all of the facts about it at runtime, and you can dynamically manipulate it in ways that just don't exist in a lot of other environments. We wanted a language that made this new model of properties, methods and events first class because that was how components were built as opposed to just sort of functions and procedures and even objects. And we wanted to create a language that was standardised. We wanted to give this language to a standardisation committee and try to level the playing field there. And all of those things were sort of like what was rolled up in C#”

The Pragmatic Engineer

18,332 görüntüleme • 4 ay önce

🦮: we know that there's people, there's a huge fandom around vampires itself and we are very lucky that we build the space around that world, that we're able to kind of talk about it in our album as well. 🦮: and overall, i think it's just new ways or new like doors that can be opened, into this big kind of thing that we built, not just our album, not just us, but also just ENHYPEN itself. 🦮: and i think a lot of people can just, you know, some people might just listen to our music and they might just, you know, like listening to music, that might be it. but some people might really enjoy the vampires aspects of the album ❤️‍🔥

🦮: we know that there's people, there's a huge fandom around vampires itself and we are very lucky that we build the space around that world, that we're able to kind of talk about it in our album as well. 🦮: and overall, i think it's just new ways or new like doors that can be opened, into this big kind of thing that we built, not just our album, not just us, but also just ENHYPEN itself. 🦮: and i think a lot of people can just, you know, some people might just listen to our music and they might just, you know, like listening to music, that might be it. but some people might really enjoy the vampires aspects of the album ❤️‍🔥

샐마•𝗘𝗡╸

19,393 görüntüleme • 2 ay önce

Sam Altman answered Elon’s “money won’t matter in 2036” statement Sam Altman: “Clearly I think money will still matter in 10 years.” “One of the greatest concerns I have with AI is the concentration of power and concentration of wealth.” “There are all of these ways where I think it will make a lot of things much cheaper and better, and you don’t have to buy a car, and that’ll be great.” “But yeah, I think that there are many reasons to think that it won’t just magically solve wealth inequality.”

Sam Altman answered Elon’s “money won’t matter in 2036” statement Sam Altman: “Clearly I think money will still matter in 10 years.” “One of the greatest concerns I have with AI is the concentration of power and concentration of wealth.” “There are all of these ways where I think it will make a lot of things much cheaper and better, and you don’t have to buy a car, and that’ll be great.” “But yeah, I think that there are many reasons to think that it won’t just magically solve wealth inequality.”

painn

315,037 görüntüleme • 3 gün önce

Today marks the six month anniversary of Pete's death. It feels like forever ago and it feels like yesterday. Time ticks by and we know there's a lot of uncertainty about what the future will hold. Many things are out of our control, but you know what isn't? Our actions. We - all of us - get to wake up each morning and decide how we're going to spend the day. We get to decide how we react to things. Choose to help Ukraine. There's a lot of noise online. Drown it out, stay focused. There are lives that rely on the fundraising done by ordinary people like us - like you. And if you get tired and wonder if it's worth the effort, just remember: Even if just one life is saved, it was all worth it. Even if just one soldier gets to return home because of your efforts, it was worth it. Razom do peremohy

Today marks the six month anniversary of Pete's death. It feels like forever ago and it feels like yesterday. Time ticks by and we know there's a lot of uncertainty about what the future will hold. Many things are out of our control, but you know what isn't? Our actions. We - all of us - get to wake up each morning and decide how we're going to spend the day. We get to decide how we react to things. Choose to help Ukraine. There's a lot of noise online. Drown it out, stay focused. There are lives that rely on the fundraising done by ordinary people like us - like you. And if you get tired and wonder if it's worth the effort, just remember: Even if just one life is saved, it was all worth it. Even if just one soldier gets to return home because of your efforts, it was worth it. Razom do peremohy

Project Konstantin

76,478 görüntüleme • 1 yıl önce

Important reality check from WashPost's Josh Rogin about the age/condition of govt planes during after AF1 had electrical issues and turned around instead of continuing its flight to Davos.... "Over the course of 20 years of traveling on government aircraft, this happens all the time. You know, it's a really old fleet and there's a constant state of disrepair. It's much more rare for Air Force One than it is for like the planes that the vice president and the secretary of defense, the secretary of state, might fly on. But these things happen, and it's because the planes are very, very old. And so, I think the fact that they caught it early and they turned back, it shouldn't really be a big problem. But it is kind of shocking to people who don't know, but it's kind of an open secret amongst the reporters who travel with senior officials that the U.S. fleet is very old and the planes are not in good condition, and that doesn't justify taking a plane from Qatar for free or anything like that. But it's just kind of a thing that everybody knows that these are really old planes and they malfunction all the time. So, they caught it. They're going to go back, they're going to get on a couple other planes. They'll be a little late. It'll probably be okay, and they'll fix the plane. But these things happen because the U.S. government fleet is very, very old and these planes are very, very old."

Important reality check from WashPost's Josh Rogin about the age/condition of govt planes during after AF1 had electrical issues and turned around instead of continuing its flight to Davos.... "Over the course of 20 years of traveling on government aircraft, this happens all the time. You know, it's a really old fleet and there's a constant state of disrepair. It's much more rare for Air Force One than it is for like the planes that the vice president and the secretary of defense, the secretary of state, might fly on. But these things happen, and it's because the planes are very, very old. And so, I think the fact that they caught it early and they turned back, it shouldn't really be a big problem. But it is kind of shocking to people who don't know, but it's kind of an open secret amongst the reporters who travel with senior officials that the U.S. fleet is very old and the planes are not in good condition, and that doesn't justify taking a plane from Qatar for free or anything like that. But it's just kind of a thing that everybody knows that these are really old planes and they malfunction all the time. So, they caught it. They're going to go back, they're going to get on a couple other planes. They'll be a little late. It'll probably be okay, and they'll fix the plane. But these things happen because the U.S. government fleet is very, very old and these planes are very, very old."

Curtis Houck

47,464 görüntüleme • 8 ay önce

Whose Line Is It Anyway? was one of those shows where you never knew what was coming next, and that was half the fun. 😂 The quick thinking, the improv, the looks they gave each other when somebody went completely off the rails… and somehow they just kept it going. Wayne Brady could take just about anything they threw at him and run with it. They just don’t make shows like this anymore. Sometimes you need to forget all the serious stuff for a few minutes and just LAUGH. 😂❤️

Whose Line Is It Anyway? was one of those shows where you never knew what was coming next, and that was half the fun. 😂 The quick thinking, the improv, the looks they gave each other when somebody went completely off the rails… and somehow they just kept it going. Wayne Brady could take just about anything they threw at him and run with it. They just don’t make shows like this anymore. Sometimes you need to forget all the serious stuff for a few minutes and just LAUGH. 😂❤️

✨ Lady Vet ✨

57,561 görüntüleme • 4 gün önce

From Eric Vishria on how the top AI founders are building products completely opposite of the SaaS era: "One of the things that is really different in the AI world versus the SaaS world, is that in the SaaS world, over and over again, you had people who really understood the customer. And the problem. And then they understood a domain. They understood what the technology was more or less capable of. But it wasn't a real question of if you could build something or not. For example, take Salesforce, Workday, and ServiceNow. CRM existed before Salesforce. HR management existed before Workday. Same thing with ServiceNow. So in every case, Salesforce followed Siebel. Workday followed Peoplesoft. ServiceNow followed Peregrine and Remedy, and others. So they were just kind of, cloud SaaS versions of the prior generation product. They just understood the customers. They understood the problem. And they were just like, here's a better version. And that evolved a little bit over time in SaaS land. But that's what it is. And so product development in that way was done by people who really understood the customer and the problems. And then just took advantage of the next wave. And this is almost diametrically opposite of product development in the AI era. When I look at the teams that are having the most success today, they have intimate knowledge of the models. They are right on the frontier of understanding which models are better at what, and why, and when. And what they're going to be good at and what they're not going to be good at. And what they're spending their time on, is figuring out how do I apply this capability of this model to this domain or to this user. So they're actually working inside out or technology out, versus customer problem in. And of course, they understand the customer problem. And a lot of times they have firsthand knowledge of it. But they're really close to the metal and capability, and they're applying it. And I think this is a really different way to develop products than in SaaS. I started my career as a product manager a long time ago, and it's almost the complete opposite of everything you learned. "Listen to the customer, understand it, then bring it back to the engineering and product teams." If you did that right now, ask a bunch of customers what they want out of AI, and you brought it back, for the most part, it may not be possible today with today's technology. Whereas the teams that are winning right now really understand the technology and are applying it out. And so I think this reversal matters. I think it's a big difference in terms of how companies are getting built. And maybe even the types of entrepreneurs that will be successful. I'm not sure. You're seeing some real change there. Look at the Bret Taylor's at Sierra. That's a super, super technical founder who really gets it. Brett and Clay really get it. You look at Michael and his co-founders at Cursor. They're super technical founders and they get it. They all really understand what these things can and can't do. And that's a pretty different dynamic relative to the way the best SaaS companies got built." Link in bio for the full conversation going deep on the current class of startups going from zero to $100m+ in ARR within 12 months.

From Eric Vishria on how the top AI founders are building products completely opposite of the SaaS era: "One of the things that is really different in the AI world versus the SaaS world, is that in the SaaS world, over and over again, you had people who really understood the customer. And the problem. And then they understood a domain. They understood what the technology was more or less capable of. But it wasn't a real question of if you could build something or not. For example, take Salesforce, Workday, and ServiceNow. CRM existed before Salesforce. HR management existed before Workday. Same thing with ServiceNow. So in every case, Salesforce followed Siebel. Workday followed Peoplesoft. ServiceNow followed Peregrine and Remedy, and others. So they were just kind of, cloud SaaS versions of the prior generation product. They just understood the customers. They understood the problem. And they were just like, here's a better version. And that evolved a little bit over time in SaaS land. But that's what it is. And so product development in that way was done by people who really understood the customer and the problems. And then just took advantage of the next wave. And this is almost diametrically opposite of product development in the AI era. When I look at the teams that are having the most success today, they have intimate knowledge of the models. They are right on the frontier of understanding which models are better at what, and why, and when. And what they're going to be good at and what they're not going to be good at. And what they're spending their time on, is figuring out how do I apply this capability of this model to this domain or to this user. So they're actually working inside out or technology out, versus customer problem in. And of course, they understand the customer problem. And a lot of times they have firsthand knowledge of it. But they're really close to the metal and capability, and they're applying it. And I think this is a really different way to develop products than in SaaS. I started my career as a product manager a long time ago, and it's almost the complete opposite of everything you learned. "Listen to the customer, understand it, then bring it back to the engineering and product teams." If you did that right now, ask a bunch of customers what they want out of AI, and you brought it back, for the most part, it may not be possible today with today's technology. Whereas the teams that are winning right now really understand the technology and are applying it out. And so I think this reversal matters. I think it's a big difference in terms of how companies are getting built. And maybe even the types of entrepreneurs that will be successful. I'm not sure. You're seeing some real change there. Look at the Bret Taylor's at Sierra. That's a super, super technical founder who really gets it. Brett and Clay really get it. You look at Michael and his co-founders at Cursor. They're super technical founders and they get it. They all really understand what these things can and can't do. And that's a pretty different dynamic relative to the way the best SaaS companies got built." Link in bio for the full conversation going deep on the current class of startups going from zero to $100m+ in ARR within 12 months.

The Peel

209,752 görüntüleme • 1 yıl önce

Sam Altman on one of the highest-return years of his career: no job, exploring his interests, and a lot of people who were nice enough to teach him things. "I read many dozens of textbooks. I learned about fields that I had been interested in. I didn't have any idea that they were all gonna come together in the way they did." "I learned a lot about nuclear engineering. AI was starting to work, so I learned a lot about AI. I learned about synthetic biology. I learned about investing." " I met people that were working on all sorts of different things who were nice enough to talk to me." "If I met someone interesting and it seemed good and needed help, I would just help them. And they would teach me stuff, or they would offer me the chance to invest in their startup later." "Out of all of it, almost all of it didn't work out. But the seeds were planted for things that worked in deep ways later." Sam Altman w/ Craig Cannon (Nov 2018)

Sam Altman on one of the highest-return years of his career: no job, exploring his interests, and a lot of people who were nice enough to teach him things. "I read many dozens of textbooks. I learned about fields that I had been interested in. I didn't have any idea that they were all gonna come together in the way they did." "I learned a lot about nuclear engineering. AI was starting to work, so I learned a lot about AI. I learned about synthetic biology. I learned about investing." " I met people that were working on all sorts of different things who were nice enough to talk to me." "If I met someone interesting and it seemed good and needed help, I would just help them. And they would teach me stuff, or they would offer me the chance to invest in their startup later." "Out of all of it, almost all of it didn't work out. But the seeds were planted for things that worked in deep ways later." Sam Altman w/ Craig Cannon (Nov 2018)

a16z

1,434,607 görüntüleme • 12 gün önce

.Marc Andreessen 🇺🇸 on why God models don't eat all AI usage: "There's one version of the world in which you have a small number of what we sometimes call God models... millions of processors, with billions or hundreds of billions or trillions of dollars of CapEx and training, investment behind it." "That's the equivalent of the world of search that we've been in for the last basically 20 years." "There's another scenario, what they sometimes call a race to the bottom, which is no, actually what happens is intelligence commoditizes." "The analogy there would be maybe what's happened in the microchip market... A really good one costs a few thousand dollars, there's companies like Arm that sell them for pennies. And you can put a microchip in anything, and it's super cheap, and everybody does it." "My guess basically is what we're about to see is something a lot like the microchip or the internet." "You'll be able to basically summon intelligence at some level into any application you need for basically very cheap or ultimately basically the same as free. And then you'll have the God models in the background when you need the super genius, but for most things, you don't actually need a super genius."

.Marc Andreessen 🇺🇸 on why God models don't eat all AI usage: "There's one version of the world in which you have a small number of what we sometimes call God models... millions of processors, with billions or hundreds of billions or trillions of dollars of CapEx and training, investment behind it." "That's the equivalent of the world of search that we've been in for the last basically 20 years." "There's another scenario, what they sometimes call a race to the bottom, which is no, actually what happens is intelligence commoditizes." "The analogy there would be maybe what's happened in the microchip market... A really good one costs a few thousand dollars, there's companies like Arm that sell them for pennies. And you can put a microchip in anything, and it's super cheap, and everybody does it." "My guess basically is what we're about to see is something a lot like the microchip or the internet." "You'll be able to basically summon intelligence at some level into any application you need for basically very cheap or ultimately basically the same as free. And then you'll have the God models in the background when you need the super genius, but for most things, you don't actually need a super genius."

a16z

104,537 görüntüleme • 3 ay önce

Palantir co-founder Joe Lonsdale reveals exact chain of events after 9/11 that made him and Peter Thiel realize they had to build Palantir "At PayPal we had to stop the Chinese and Russian mafia from stealing all of our money. So we got to know all the guys who were helping us arrest the bad guys, the Secret Service and the FBI. And right after this happened, it was 9/11" "These guys were spending billions of dollars on stuff that we thought didn't make any sense. They kept asking us for advice. They were confused about how to do things" "When you create a new department, people can't really fire people in government easily, so sometimes they'll have a lot of people they wanted to fire and then instead they just pushed them into the new department. So it was a bit of a mess" "It became really obvious to us that Silicon Valley, Google, PayPal, all these things were just way ahead technically of where the government was at that point" "The government was spending $38 billion a year gathering data, looking at the data, failing to stop the terrorists, but also abusing our civil liberties" "We said, there's actually a really important problem here to solve. I'd like to stop the bad guys from attacking us again. And I don't want everyone in the government seeing all of my data without any controls. That's crazy"

Palantir co-founder Joe Lonsdale reveals exact chain of events after 9/11 that made him and Peter Thiel realize they had to build Palantir "At PayPal we had to stop the Chinese and Russian mafia from stealing all of our money. So we got to know all the guys who were helping us arrest the bad guys, the Secret Service and the FBI. And right after this happened, it was 9/11" "These guys were spending billions of dollars on stuff that we thought didn't make any sense. They kept asking us for advice. They were confused about how to do things" "When you create a new department, people can't really fire people in government easily, so sometimes they'll have a lot of people they wanted to fire and then instead they just pushed them into the new department. So it was a bit of a mess" "It became really obvious to us that Silicon Valley, Google, PayPal, all these things were just way ahead technically of where the government was at that point" "The government was spending $38 billion a year gathering data, looking at the data, failing to stop the terrorists, but also abusing our civil liberties" "We said, there's actually a really important problem here to solve. I'd like to stop the bad guys from attacking us again. And I don't want everyone in the government seeing all of my data without any controls. That's crazy"

Jaynit

92,768 görüntüleme • 2 ay önce