Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

conversationswithtyler.com

Post reply on HN
Print view View original

61–70 of 141 posts

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#61

Post by dang »

hogFeast wrote: post #53

The self-importance of pretty much everything in this interview is literally unbearable. I will give you the precis: I think X is important, why doesn't everyone think X is important (usually about five sentences of escalating grandeur about the importance of X), because I think X is important and "no-one is talking about X" you know that I am a real genius...on and on. Cowen is also an utter ignoramus. He seems to c…

Thoughtful critique is fine, but please don't post petty personal attacks to HN. It poisons the well regardless of how right you are. Also, please omit the name-calling.

If you'd please review https://news.ycombinator.com/newsguidelines.html and take the spirit of this site more to heart, we'd appreciate it.

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#62

Post by whb07 »

eloff wrote: post #15
whb07 wrote: post #11

Earlier quoted context omitted.

Why wouldn't it be? Humans have infinite wants and needs. These change overtime in terms of taste/perception as well. So the counter point should be made as to how this isn't the case and that a "growing pie" is actually against human nature.

It's the compound interest problem. If you keep compounding growth, like growing GDP at 1-2% a year, within some few thousand years you've grown bigger than all the resources within a sphere the same number of light years in radius. So clearly, growth is not sustainable and must eventually slow down until it approaches zero. To be clear, we're far away from that today. But the parent is mathematically correct, the be…

Uh, more notable incorrect persons have noted this argument. You’re using what’s called a Malthusian argument.

Anyhow, let’s say everything you say is correct for today’s wants and needs. But if it is true that society changes and needs change etc, then your argument doesn’t hold for tomorrow.

Suppose for a moment there is only 10 years worth of X left, who is to say that one of these things doesn’t happen:

* X thing is no longer needed because Y thing is no longer in fashion

* X thing gets replaced by other comparable Z material

... and this goes on forever.

That’s the point.

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#63

Post by pc »

the_watcher wrote: post #27

What I was most struck by in the conversation is that Zuck is actually doing some of the things Cowen advocates - policy experimentation, taking on long-term projects with uncertain payoffs, etc. at CZI. I'm a huge fan of Patrick Collison and know that Stripe is also doing tons of experimentation and product innovation, but I was struck by how much less like an academic Zuck sounded. I don't know if it's a good or ba…

> What I was most struck by in the conversation is that Zuck is actually doing some of the things Cowen advocates - policy experimentation, taking on long-term projects with uncertain payoffs, etc. at CZI.

I'm a big admirer of what Mark is doing with CZI (which is part of why Cowen/I agreed to the interview). Stripe is still a private company, but maybe I'll be able to intervene more directly over time. If I do, having CZI as precedent will be very helpful.

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#64

Post by chubot »

valleyjo wrote: post #2

They discuss the fact that we are seeing linear gains from exponential investment in several fields (healthcare, semiconductors to name a few). This is a fascinating problem.

That reminds me of deep learning. Investment in computing resources is definitely increasing exponentially over time:

https://openai.com/blog/ai-and-compute/

https://techcrunch.com/2019/11/10/the-post-exponential-era-o...

Tthere doesn't seem to by any evidence of exponential gains in performance, even on their own benchmarks (which importantly don't measure "intelligence", but something more narrow):

https://blog.piekniewski.info/2019/11/18/late2019-the-wizard...

Even though we are doubling the compute available to AI every 3.5 months, so a factor 10x per year, since say 2015, so some 10.000x, the classification performance as measured by top-1 accuracy on ImageNet had barely moved. Now arguably top-1 accuracy on ImageNet is probably not the best measure, but still that seems somewhat striking. The diminishing returns are visible elsewhere too, even that Open AI Rubic's cube is somewhat indicative.

From an outside perspective, it seems that everyone is coming around to the fact that self-driving cars is harder than expected. There are no "exponential gains" in that field -- to the contrary, it seems like 90% of the problem was solved several years ago, and you can now take undergraduate courses that teach the basics.

But going from 90%, to 99%, to 99.9%, etc. is taking decades, not months or years. There seems to be something fundamentally wrong with the current approach.

I believe at the time the OpenAI blog post was supposed to be "impressive". Certainly I felt a bit of that -- "hey something important is going on there!" I believe that in 10 years it will look like a waste of effort. There will be a different, more efficient approach to solving these types of problems.

----

Peter Thiel is one of the only "major tech figures" speaking out against AI, which I appreciate. And the "AI is communist" meme is pretty memorable, and I don't think wholly inaccurate. I listen to him for his contrarian opinions.

I just wish he didn't have these random "tribal conservative" outbursts. The most recent one was a video on here where he was rambling about Obama saying something far worse than George Bush and the Iraq war. I guess it was about sending their daughter to Harvard, but saying that lots of schools are just as good.

He was trying to get some points from his conservative audience. I guess it gets attention, but that kind of tribalism undermines the rational points that he generally makes.

(I realize the OP doesn't feature Thiel but he was brought up elsewhere in this thread, and is one of the people also lamenting the lack of technological progress despite investment, although maybe due to lack of investment in certain areas too)

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#65

Post by nugget »

valleyjo wrote: post #2

They discuss the fact that we are seeing linear gains from exponential investment in several fields (healthcare, semiconductors to name a few). This is a fascinating problem.

One purely linear constraint is time. In many health care fields, it simply takes time for clinical studies to collect data, from which the next generation of clinical studies can be planned, and there's no good way to meaningfully speed up that process.

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#66

Post by mikorym »

pc wrote: post #63
the_watcher wrote: post #27

What I was most struck by in the conversation is that Zuck is actually doing some of the things Cowen advocates - policy experimentation, taking on long-term projects with uncertain payoffs, etc. at CZI. I'm a huge fan of Patrick Collison and know that Stripe is also doing tons of experimentation and product innovation, but I was struck by how much less like an academic Zuck sounded. I don't know if it's a good or ba…

> What I was most struck by in the conversation is that Zuck is actually doing some of the things Cowen advocates - policy experimentation, taking on long-term projects with uncertain payoffs, etc. at CZI. I'm a big admirer of what Mark is doing with CZI (which is part of why Cowen/I agreed to the interview). Stripe is still a private company, but maybe I'll be able to intervene more directly over time. If I do, havi…

I don't want to sound like I am trolling or something. However, I find it hard to take anything Mark Zuckerberg does seriously. For that matter, I am so biased that I have no interest in these interviews at all (maybe to your detriment).

How do you see this? You say that you are an admirer of CZI. In my experience with obnoxious CEO types, they are happy to be philantrophic or "build" things when it suits them, and only then. Do you think it's a net positive? Wouldn't someone else (like Gates or even Buffet) be a better leader for these initiatives? I'm sure that if Mark did not get so successful with FB, it would have been the WA founders, otherwise someone else yet again.

Sorry if this is a steriotypical reaction, but I am sorry; I just can't trust something from camp Zuck.

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#67

Post by alkibiades »

svcphr wrote: post #23
artfulhippo wrote: post #9

> Collison: when we sort of talk about the phenomenon of “progress,” you know, I think GDP or GDP per capita is a pretty good first approximation measure of it and it correlates strongly —— I got confused when Collison said that there was scientific evidence that business productivity increases with improved management practices. That’s the thesis of business schools and management consultants. What’s the new idea he…

He's referring to research by Nick Bloom, an economist at Stanford. They've worked together. Ctrl-F management on his research page ( https://nbloom.people.stanford.edu/research ). In one paper, they randomly assign consulting services from a management consultancy to manufacturing plants in India and found a 17% increase in productivity.

not sure that means it generalizes to all countries or types of business

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#68

Post by mikorym »

Isamu wrote: post #40
sova wrote: post #5

It seems that naturally with technological progress and population increase, we are doomed to always zero out our gains in the minute with losses in the aggregate. How is it that science and technology move forward but crop yields remain constant? Cowen mentions that it seems to be a "last mile problem" for delivering scientific research benefits to all. Collison has a cool list of "Fast" completed projects on his si…

>but crop yields remain constant Some, not all, crop yields have been stalling in recent years. I think Tyler's point was maybe the increasing cost of research and effort to keep moving the yield curves up, such that it may be stalling progress there. Data on crop yields here: https://ourworldindata.org/crop-yields

>> crop yields remain constant

> some, not all...

Good that you made this correction. In 1960 the average yield for tomatoes was 50 ton/ha or less in the area where I am from. [1]

In 2019, the yield was 100 ton/ha, but in the most optimal areas you get 140 ton/ha. If you use nets you get 250 tons/ha.

>> How is it that science and technology move forward but crop yields remain constant?

This is bordering on conspiracy theory; I have experience in tech in agriculture and the progress the last 50 years is nothing short of a humanitarian miracle. The miracle is far from complete though.

[1] First hand knowledge.

Re: Mark Zuckerberg Interviews Patrick Collison and Tyler Cowen

#69

Post by buboard »

pyb wrote: post #38

"What’s interesting and, anyway should be concerning, is that for almost every conceivable applied measure we seem to be getting at best constant returns. But that’s really bad because we see exponentially increasing inputs and constant return outputs." Ironically, in terms of innovation, the yearly output of companies such as Facebook and Stripe has seemed somewhat constant, despite those companies getting many orde…

if that were true the companies would have gone bankrupt, but even google keeps increasing its year to year profits.
Post reply on HN