Kai-Fu Lee | AI Era - Leadership and Technology

As part of the Dean's Speaker Series www.haas.org/deansspeakers at Berkeley Haas, leading thinker on AI, Kai-Fu Lee, explores themes of leadership and technology that are of critical importance in the era of AI.
Kai-Fu Lee serves as Chairman and CEO of Sinovation Ventures, an international technology investment firm. He previously held executive roles at Apple, Silicon Graphics, Microsoft, and more recently served as President of Google China. Dr. Lee has written two books, including the newly released AI Superpowers: China, Silicon Valley, and the New World Order.
Dr. Lee discusses how and why AI’s impact -- led by the two superpowers of US and China -- will be different from that of any previous technology revolution we've seen before. He also covers the many benefits and challenges AI brings to business and society and Dr. Lee’s proposed solutions for healthy co-existence between humans and AI.
This event is co-sponsored by the CITRIS Research Exchange

[@typhooonn]

Great insight from Kaifu Lee

[@shepherdmukora5558]

kai always giving educative insights on ai

[@laoniuhaha]

Great questions and answers.

[@AlexFalararo]

Wow!

[@peterlavina4138]

None of what he said was artificial; they're all human! Great thinker and advocate of AI and the need for us to prepare for its eventuality.

[@boostftw123]

Great talk. There is no bridge from the east to the west, given the amount of sinophobia in the west. Even if there was a bridge, it's been burned by those that are fearful of contribution (guess who? Hint: USA). Best to focus on one parallel that you have an advantage in. However, that is only true for software. Hardware and other spaces, there is co-dependency between these two countries.

[@goldnutter412]

The US and China are not arch enemies they are secret superfriends.. global tension is as far as anything goes for years now.. they all get we are on the same team, we don't get the behind closed doors view because we aren't able to understand all the details on any level. The only way global leaders do their job is with countless delegation and ? aides. Everyone has aides. They rely on the consulation and expertise of others more than anything else, and have to take on masses of incomplete data, opinion and bias included ? sure sometimes there is an unknown bias. Only human.. mistakes ? we all make them all the time. Misinterpretations and bad bets ? we all make them, all the time ! We all work with incomplete data, this reality keeps telling us.. TOUGH ! deal with it !! What China teaches us in the West ? respecting the planet and teamwork.. is an immediate efficiency upgrade, and leaves far less complexity and edge cases where things are pressure without anywhere near enough data - those things happen more in the West, and the fallout is obvious every time. What the West teaches China ? lots of how not to do things.. but that everyone should have a chance and the core values of America and capitalism and a shot at making opportunities into more and more success.. they are great. Just a broken system now that so much human bias, meddling and so on basically killed capitalism. Long live open source.. collaboration.. the G groups.. the UN... EFFECT FORCE is a great architecture the UN is behind.. humans as middleware, provide context for dumb brute force compute.. the result.. harmony and productivity ! China is awesome.. we are all awesome WE ARE ALL JUST WATER walking around with some level of ego.. ego is very dangerous. Ego shows how much fear potential we have, but also how much initiative and drive we can muster.. willpower.. its an important two sided coin deal.. Ego points to possible fears.. negative emotions surrounding your "SELF" and your privilege or negativity towards others is evolution yelling, over here.. this is the hard way.. try the easy way ! trust can be earnt. Someones moral fibre can be estimated. Without being reckless, we have to let go of our fear to be truly happy, and look more to empower others, give what we can.. and by evolving ourselves we become more and more useful to others, who in turn are more likely to become more evolved and so on. Pretty easy equation; Everyone focuses on themselves, takes everything they can even if it means killing someone else - end result none of us would be here. Any level of cooperation... we have a chance.. we are still here. Willingly being selfless and leading by example ? leads to Utopia. We are all responsible for choosing how we interact and why we do what we do.. there are no excuses. Everything is up to every one of us. The more we avoid conflict and try to give as much as we can without ruining everything for ourselves, the faster it all just takes off and we all get happier exponentially. Momentum and self reinforcing cycles are far more powerful and common than we recognize.. look at memes and meme shitcoins go. People who find life hard become sheeple.. But plenty of warnings.. advice.. and trying to help people is all we can do. FREE WILL CHOICE is the most important human right. Denying choice is very dangerous, and prohibition of alcohol is the best example.. people just said well no, sorry just no. We choose NOPE.

[@ChuckSilva]

👁❤️

[@mustafagenius2110]

996 means 9am to 9pm 6 days a week, I like that😊

[@AliasHSW]

Mustafa Genius - I like Porsche 996 more than working 996, unless it’s a works 996

[@laatena]

With a breaks, etc., I hope!

[@db09pl]

Since when Bitmain is pure AI company? Their business is mining cryptocurrencies..

[@goldnutter412]

Yeah I get that too.. but they are a hardware producer, not a miner. They create complex, very specific machines. Only for mining, doing anything else is a complete waste of the silicon, memory and hashing algorithm management, and more that I can't be bothered getting up to speed on enough to be more clear.. Those machines have had logic boards from the beginning, which control how the many basic little cpu units (next evolution could be millions of cores if not already, haven't got infinite time to give exact numbers) But does it matter ? the fact this is the way the machines function, which actually matters a lot to have a data analysis management system.. AMD CPU's had neural networks how many years ago ? nVidia cards ? the same.. So yeah.. its not actually wrong for him to say that.

[@twenlil]

Andrew Yang was not wrong after all.

[@laatena]

Don't forget about Russia?

[@laatena]

Which wasnt so obvious att.

[@henrifritsmaarseveen6260]

The advanced is only because of population of china it has more data. What if AI systems need less data to accomplish the same result example humans do not need to see 1.000.000 dog and cats to see the difference what if machine learning in the near future would only need 10% or less of the data that is needed now? Then the complete advanced of china is gone .. So American and Europa AI developer will focus on this point , this would help AI development very much in a positive way Also if you need smaller data sets you can help with research where data sets are limited , for example rare diseases. So how will be the leader in AI is still open , because we do not know what will be developed in what order.

[@trionikzmindz7691]

It's called spiking neural network, Capsule networks . These deep learning tools are designed to work with less data.

[@NangongReng1973]

The world will not stay static for u to take advantage.You may be super good something but you may lose all advantage tomorrow.The main thing is stop grumbling and start to see opportunity from problems .Westerners won't understand this.

[@waynet8953]

Your thinking is not correct. Information/data can only increase accuracy assuming you have the proper data processing / analysis. Less data means less accuracy; more data gives more accuracy/ information. For a given application, it may be found that a certain amount of data is sufficient, and getting more may not be worth the "cost". What you're saying is like having more education, and becoming more stupid.

[@henrifritsmaarseveen6260]

Wayne T This is not true look how animals and humans learn they don’t need a lot of data , that means we are on the wrong path of building a AI .. it is tooo stupid , we will make systems in the future that will only need a couple of examples to train the model ..

[@waynet8953]

@@henrifritsmaarseveen6260: I suppose when you do that, you'll have a better system; my point was that more data won't make it worse; if you get an approach that is more accurate for the same amount of data, then it has a "faster covergence"; even then, more data may give it a higher accuracy, not less.

[@JohnBauschard]

TYPICAL CRAP WE EXPECT from a Standard Professor continually talking over Dr. Lee with her constantdisrespect. I have watch innumerable amounts of Dr. lee's lectures, so why do you think Stanford only as 21 comments to his speech. I had friends tell me about this speech and now seeing it I still don't believe it FYI: So few comments can always be explained, they remove the bad ones. Same thing Twitter does removing 6 of my Tweete under their "Community standards rules".

[@supermachoalfa5015]

China > USA me >> learn manderin

[@davidcripps3011]

Definitely the most intelligent Trump has sounded....