Ben, insight"s on robots needing new physics models hits home. It mirrors how humans develop sensorimotor skills through real-world interaction. Imagine robots learning like newborns, gaining 'street smarts' neuron by neuron!
Mark correctly listed all the non maker traits the cost-disease society demands (diplomat, lawyer, social networker, etc.). Ben lists the one thing that is needed, and the tragedy of not putting it first. This is why founder -controlled companies are outperforming 100x. (And Andy Grove’s high performance management is a masterclass in high performance goal setting)
On the energy/electricity front, it will be interesting to watch the Poland deal with Rolls Royce to install factory-built mini nukes: willingness to scale up kWh like this will strongly influence where AI and industry get located. On the “boiling the sea” comment - combined-cycle generators can consume that heat and recycle it as electric power.
Ben and Marc Great stuff as usual! Demonstrating you have the right people on board who can wear multiple hats does signal to investors that you can effectively utilize their capital. Yet from a Founders perspective having a seed stage investor willing to help shape the team and vision is 💯, especially for founders from diverse backgrounds.
Absolutely! Embracing advanced manufacturing with fully robotic and AI-enabled factories is key to revitalizing US manufacturing. CEOs with backgrounds in product creation can bring invaluable insights to companies aiming to innovate and build complex products. 🏭💡
In all caps as if it's not a society in massive downfall and the only reason it still exists is cause US committed mistake of a lifetime by helping them build nukes.
Airbus CEO Guillaume Faury or French big enterprises bosses (e.g TotalEnergies - Patrick Pouyanné, LVMH - Bernard Arnault) usually are graduates of Ecole Polytechnique, the best engineering and most competitive grande ecole in France - no wonder Airbus kicks Boeing's ass.
Maybe LLM can help pick the next CEO, if one trains LLM on all the emails, meetings and chats and the board asks LLM who are the top guys and why. Of course it should be especially trained on filtering out fake mutual recognition that is always there, especially on the management level, that would probably be hard part to figure out. LLM should focus more on output than words , like thanks for helping with the problem is more worth then statement you have done truly amazing job .
I think emerging technology specialists are going to help most businesses and communities transition over the next 5 to 15 years or so. The changes are going to be insane at a daily rate by the point though. By then it should just sort of flow state in a way that simply wasn't possible until then. We'll need local, national and international human specialists. Everybody else just gets to get help having life be better. It's going to be great.
ben's comments on CEO selection are really on point - but (judging from my oiwn finite Board experiences avoiding Board internecine warfare is pretty hard to navigate. I'd be interested how you assemble an effective coalition of Board members to support a candidate without completyely pissing off everyone elkse on the Board. Is there a general structure for this process that's been successful for uoi?
I'm really curious if anyone thinks that the most efficient route is to have mechanized robots running around larping as humans? I always assumed the best approach is a completely enclosed system that does things at it's pace, not bottle necking any independent systems.
Hey Guys, given the incredible potential for AI enabled robotics one area of new future work (jobs) is (or very much should be) discrete task trainers. Where I am going with this is AI capability is limited on two fronts power and compute at the individual robot level (imagine a Optimus in your home). You can either try to get a robot to learn a new task through trial and error (with some base capably) OR you could employ a trainer that can specifically help your robot learn the task both quickly and accurately for your environment. This would reduce the compute needed for training and ultimately the power as well. In addition provide avenues for incremental employment of current labourers. Do you have any thoughts on this?
How may our startup raise capital to fulfill requests from the nations largest military, law enforcement and emergency response fleets for a public demonstration of our hardware product.
There won't be meaningful job cuts. Just some temporary stirs. We're chronically understaffed across industries and we'll be using AI as a complement technology because the decision making technology isn't and won't be capable of human level capacity for a long time. Population is aging and heading into decline and we can't serve the needs. Currently you can't have any service delivered on personalized capacity level because we don't have enough humans.
Listen to me Marc. 15 Minutes in, and I have a comment to make. EVERY COMAPNY IS IN THE BUSINESS OF MAKING MONEY. and I have a point to prove it, when I say that a stock went from $10 to $20 nobody asks me what the company does, because it does not matter. The only thing that matters in business is money, the product is the supply to the demand. Study Economics. Macroeconomics to be precise.
How many hotels AI factory needs? I think its overblown how many "workers" factory like that will require. The incentive is to put it to 0, or at least scale expensive labor by remote access. Amazons dream is fully robotic warehouse - how many hotels needs that around it? Zero.. so why factories would be any different? AI factory will create enormous value, but not for people being close to it geographically.
The problem with this thinking is that the fancy AI factories are still going to have to compete with overseas human factories and supply chains. And they’ll lose, at least at building the products that are currently available. The only way for this to work is for the AI factory to build products that do not currently exist, either because they haven’t been invented yet, or ideally making products that are currently impossible to make by human teams.
Awesome awesome awesome. But why? Like, who cares? Nerds? There's no money in selling machines to other machines. The wave of boredom that is coming week be unprecedented. These people think Elon musk is NOT a scammer. It's incredible.
I had so many questions answered from this single video, loved the no nonsense approach😂
Massive respect for this amazing lessons for free. Your friend from Afghanistan 🇦🇫
I can see you guys have a decade of flow together. Just found this channel... awesome
Appreciate your optimism for robotics
This is better than Harvard business school Holy shit for free
I feel the same.
Ben, insight"s on robots needing new physics models hits home. It mirrors how humans develop sensorimotor skills through real-world interaction. Imagine robots learning like newborns, gaining 'street smarts' neuron by neuron!
Thank you so much guys for educating us
I loved it. So tactical. Would love more like this.
Love these deep dives. Ben and/or Marc, come on AI For Humans to talk about this.
Mark correctly listed all the non maker traits the cost-disease society demands (diplomat, lawyer, social networker, etc.). Ben lists the one thing that is needed, and the tragedy of not putting it first. This is why founder
-controlled companies are outperforming 100x. (And Andy Grove’s high performance management is a masterclass in high performance goal setting)
On the energy/electricity front, it will be interesting to watch the Poland deal with Rolls Royce to install factory-built mini nukes: willingness to scale up kWh like this will strongly influence where AI and industry get located. On the “boiling the sea” comment - combined-cycle generators can consume that heat and recycle it as electric power.
House of Harkonnen has spoken 😂 great insights as always from Ben and Marc 🤌
Make more! These are amazing!
Charlie & Buffet bro-mance. Beautiful to witness chaps who go through the hard things, about hard things, and still be making memes, today.
Love these keep em coming!
Ben and Marc
Great stuff as usual! Demonstrating you have the right people on board who can wear multiple hats does signal to investors that you can effectively utilize their capital.
Yet from a Founders perspective having a seed stage investor willing to help shape the team and vision is 💯, especially for founders from diverse backgrounds.
Absolutely! Embracing advanced manufacturing with fully robotic and AI-enabled factories is key to revitalizing US manufacturing. CEOs with backgrounds in product creation can bring invaluable insights to companies aiming to innovate and build complex products. 🏭💡
Shut up Carl, this is not Linkedin.
The transformation from advanced ML / DL to AI is also a HCI & hardware revolution with or without A16 - 43:02
Robotics will be perfect decoupling from China , dominating speed of innovation for USA
Another fire episode 🔥🤖
Marc's new favorite word is steelman
Chamath ruined that word for me.
@@GiantsOnTheHorizon 100%
@@GiantsOnTheHorizon lex fridman ruined that word for me
LOL, fun fact is steelman in Russian is Stalin.
I like you more than all in already!
Premium content. High calibre insights. Awesome mind share
Great deep dive. Also the Brown shoes black socks tidbit had me crying with laughter 😂
Fast becoming my favourite podcast
I wish Marc would tell us what books he's got behind him.
You can’t do consensus hire of CEOs. Ben you nailed it!
Accounting is the core competency for such complex companies.🎉
Aka Financial Engineering
great episode
Great discussion
😂 Ben on the output of an Airplane Company CEO: "The airplane doesn't fkin fall out of the sky"
the first 30 seconds ! thats the pod
🦅🦅
Massive Respect for Both from PAKISTAN.
Love Pakistan!
Lol
Pigs
as their government sponsors terrorism@@moisesdelcastillo6703
In all caps as if it's not a society in massive downfall and the only reason it still exists is cause US committed mistake of a lifetime by helping them build nukes.
Airbus CEO Guillaume Faury or French big enterprises bosses (e.g TotalEnergies - Patrick Pouyanné, LVMH - Bernard Arnault) usually are graduates of Ecole Polytechnique, the best engineering and most competitive grande ecole in France - no wonder Airbus kicks Boeing's ass.
I’m not surprised. I feed safe in an airbus
Great discussion on the future of manufacturing! What do you think is the biggest hurdle for widespread adoption?
(sips tea) 🤣
Great one again, thanks for doing it
How can I contact Marc? or send him something?
Who else immediately thought of Iran when they talked about low cost data centers at 49:15
Maybe LLM can help pick the next CEO, if one trains LLM on all the emails, meetings and chats and the board asks LLM who are the top guys and why. Of course it should be especially trained on filtering out fake mutual recognition that is always there, especially on the management level, that would probably be hard part to figure out. LLM should focus more on output than words , like thanks for helping with the problem is more worth then statement you have done truly amazing job .
I think emerging technology specialists are going to help most businesses and communities transition over the next 5 to 15 years or so. The changes are going to be insane at a daily rate by the point though. By then it should just sort of flow state in a way that simply wasn't possible until then.
We'll need local, national and international human specialists. Everybody else just gets to get help having life be better.
It's going to be great.
I suck at talking to investors but I am working on it , to get better
Drink every time he says “um”
ben's comments on CEO selection are really on point - but (judging from my oiwn finite Board experiences avoiding Board internecine warfare is pretty hard to navigate. I'd be interested how you assemble an effective coalition of Board members to support a candidate without completyely pissing off everyone elkse on the Board. Is there a general structure for this process that's been successful for uoi?
I'm really curious if anyone thinks that the most efficient route is to have mechanized robots running around larping as humans? I always assumed the best approach is a completely enclosed system that does things at it's pace, not bottle necking any independent systems.
Hey Guys, given the incredible potential for AI enabled robotics one area of new future work (jobs) is (or very much should be) discrete task trainers. Where I am going with this is AI capability is limited on two fronts power and compute at the individual robot level (imagine a Optimus in your home). You can either try to get a robot to learn a new task through trial and error (with some base capably) OR you could employ a trainer that can specifically help your robot learn the task both quickly and accurately for your environment. This would reduce the compute needed for training and ultimately the power as well. In addition provide avenues for incremental employment of current labourers. Do you have any thoughts on this?
💯
This 2 insight are more powerful than current US President.
Elon spells it out pretty clearly: you're the CEO, make your product the best it can possibly be.
Marc + Ben Shapiro + Destiny real-time translation software. NVidia's secret weapon when bench-testing?
Shenzhen China ? 42:00
I can't even get a hold of a venture capitalist and I've been trying for 3 years
How may our startup raise capital to fulfill requests from the nations largest military, law enforcement and emergency response fleets for a public demonstration of our hardware product.
First! Cookie?
seed round
People don't care right now because AI is not affecting them as of yet. When AI affects them and starts cutting their job then they will care
There won't be meaningful job cuts. Just some temporary stirs. We're chronically understaffed across industries and we'll be using AI as a complement technology because the decision making technology isn't and won't be capable of human level capacity for a long time. Population is aging and heading into decline and we can't serve the needs. Currently you can't have any service delivered on personalized capacity level because we don't have enough humans.
Does Apple actually build their own computers though?
Everything is relative / in context. In relation to the others, yes.
A16z should look into EV play Bumblebee
Listen to me Marc. 15 Minutes in, and I have a comment to make. EVERY COMAPNY IS IN THE BUSINESS OF MAKING MONEY. and I have a point to prove it, when I say that a stock went from $10 to $20 nobody asks me what the company does, because it does not matter. The only thing that matters in business is money, the product is the supply to the demand. Study Economics. Macroeconomics to be precise.
Guys lets group together to discuss our ways to get global, we could share some experiences. Somebody up to? Lets create discord or sth
How many hotels AI factory needs? I think its overblown how many "workers" factory like that will require. The incentive is to put it to 0, or at least scale expensive labor by remote access. Amazons dream is fully robotic warehouse - how many hotels needs that around it? Zero.. so why factories would be any different? AI factory will create enormous value, but not for people being close to it geographically.
The only super tech proficient manufacturers in the AI Auto Space and Robotics sector is Tesla/Space X
Fourth!
I started business in 2024.
The problem with this thinking is that the fancy AI factories are still going to have to compete with overseas human factories and supply chains. And they’ll lose, at least at building the products that are currently available.
The only way for this to work is for the AI factory to build products that do not currently exist, either because they haven’t been invented yet, or ideally making products that are currently impossible to make by human teams.
No. They won't. Humans don't live very long and population is heading into decline.
Argentina IA Data Centers, just watch !!!!
The dollar is so strong and labor cost is very high. Who can afford to buy our stuff?
Manufacturing is the hard part. Plus China is very good at copying hardware.
why the hell is the episode titled AI, Robotics yet half of it is about board and ceo politics. False advertisment
um uh uh uh um uh uh uh um uh
Awesome awesome awesome. But why? Like, who cares? Nerds? There's no money in selling machines to other machines. The wave of boredom that is coming week be unprecedented. These people think Elon musk is NOT a scammer. It's incredible.
Elon is the obvious choice as winner in AI on FSD but just not popular with the idiots
Cut out the humans.. that is what this dude is talking about
Stop saying steel Man
great episode