Comment by mswen | Hacker News Reader

Comment by mswen | Hacker News Reader

One thing that the author focuses on is trying to predict what will be big in the medium to longer time horizon and preparing oneself to be competent in that while few others realize that "it" even exists.

I am a bit older than the author and have had some experience with this strategy. One specific example is creating an approach to latent semantic analysis that worked really well back in 2000. My partner and I couldn't get people with money to pay attention. We ended up putting it on the shelf and moving on. Around 2007 semantic gets really hot with VCs but we are each engaged elsewhere.

My point is that this strategy can suffer from being too early. You may eventually be shown to be right but you were too early or not persuasive enough to gain the benefit from the insight and work.

Still, I find myself much more intrigued with creating something new, or exploring the leading edge of some field than just grinding out the next iteration of the same old thing.

Maybe one day I will get the timing just right.

Young people tend to severely underestimate how much luck and timing play into success.

When I read, "When I was in 9th grade...and Googled..." I realized why the article sounded like it did.

I had an identical reaction. I'd like to think I'm above simple agism but someone who is 29-31 years old dropping life lessons seems hollow.

I was curious about his background and started googling him: http://www.jameskoppel.com/bio.html. He was a Thiel fellow in 2012 so he was less than 20 yrs old then. He is at most 23 yrs old now.

I saw some discussion about corporate onboarding recently, apparently the best person to help the new hire get onboarded is the person who was most recently onboarded themselves.

Maybe something similar applies to adult (or almost-adult) life?

I don't think so. The reason a recent hire is well suited to onboarding a new hire is because process of onboarding is very dynamic and changing and its a one time or at least rare thing in the life of an employee. The onboarding process the new employee is going to go through is likely completely different from what a employee might have gone through 15 years ago but is probably almost identical to want someone hired 15 days ago went through. It seems silly to me to compare who is best at providing help to complete a single, short, well defined task to who may be best at advising how to live a happy productive adult life. How could a fresh 18 year old adult possible give advice on how to balance your work and family life?

Edit: Now if you were to narrow the question to: What age of adviser is best for assisting a young high school grad with their transition to college life such a comparison might be more meaningful.

Being around 30 gives more-or-less a decade of experience at being an adult. Which is plenty of time to get your 10k expertise hours in if you haven't been completely slacking off the whole time.

You know the whole 10,000 hours thing is made up bullshit right?

Yep. :)

Which is pretty much what people do in their twenties.

A 30 year old will probably give better life advice to a 15yo than a 60yo for example.

I think it's about relevant experience at your level.

In fact, the secret of the not-as-young people is to manufacture the luck as much as they can and increase exposure to serendipity through consistent - well - showing up. And behind most lucky youngsters are diligent parents and support networks that know this lesson and apply it to their offspring, even if the offspring stay unaware.

I've heard this quote before in various forms but I don't like it.

The luck applies to those who are already doing everything right. The other people are "failures" because you can pinpoint where they were deficient. When I think about luck and success I picture a pool of highly educated, ambitious, type A people. Why does one of them become an industry tycoon billionaire while the others become merely wealthy? You can manufacture luck for the latter but not the former.

   "manufacture the luck"

If you can manufacture it, it isn't luck, period. And you need some luck to succeed.

Your broader point stands though, you miss 100% of the opportunities you didn't show up for.

I think the idea is more like that saying "Opportunity is a meeting of luck and preparation". So maybe it's more like 'manufacture opportunity' by constantly showing up prepared.

I like that much better.

Luck is when things fall out better than you could reasonably have expected. Opportunities are something you can work for.

manufacturing the conditions for luck. Sheesh.

I think the distinction is important.

There is a pernicious lie we tell, culturally, about success as if it were a matter of worth and effort. The truth is most success has a strong element of luck, and many failures as well.

you're not looking at the right examples. Sure luck. But my colleagues who started companies who are 40 - 300 employes strong now, they did not do it by luck. They wanted something, knew how to do things better, took risks and made it. Then 15 years later somebody could say something about luck but it's almost insulting. However my colleauges would acknowledge luck, but in reality it was just a ton of the right decisions and service to others creates the momentum.

Luck is just probability, and the Law of Large Numbers says that low-probability events are much more likely to happen if you sample repeatedly.

No, luck is when your manage to get a sample that is much better than the expectation value. Of course this has to happen for some people, some of the time - but counting on it is irrational.

There is nothing lucky about merely resampling until in expectation you should see the result you want. That's just doing the work.

>There is nothing lucky about merely resampling until in expectation you should see the result you want. That's just doing the work.

That's just showing up.

You could also consider luck to be our sampling bias of only taking into account the successful outcomes.

The population as a whole that is attempting to determine the 'next big thing (tm)' does the majority of the sampling (with a small amount of resampling from those who manage a second or third try). We then read the stories of the successful outcomes and call it luck.

> If you can manufacture it, it isn't luck, period.

I'm not sure I agree with the statement.

It is possible that a certain outcome is attached to a certain action, but the link between the two is sufficiently hidden that the outcome seems like luck. Is it luck? It surely looks like luck, and behaves like luck. I would call it luck.

But of course for some people, that link might be a little more obvious. Those are the people that can "manufacture luck."

I would like to note, it's usually not a single action that generate luck, it's a set of complex actions, which is why it's hard to manufacture, and stays luck.

This is really insightful, there's a pretty large library of algorithms and approaches that have been sitting on the shelf for decades (or used in very few places) that become far more interesting with modern hardware. The recent resurgence with Bloom filters for example, a technology form the 1970s, used in a few database products, is now everywhere.

But there's also the curious problem of expensive and difficult to produce technology that doesn't have a market large enough to sustain it. Example, the crop of now dead or dying enterprise entity extraction companies. It turns out it's expensive to build those systems, but the market isn't that large, which forces the price up to try and survive, which then further reduces the market and so on into a death spiral. Nowadays people just try to get by with OpenNLP...which was easily bested by tools like ThingFinder and ClearForest a decade earlier.

The bloom filter thing is interesting - it was just useful enough to stay occasionally used, till what a decade ago when simple lookup needs exploded around key-value stores.

But I would suggest the "occasionally used" is as much a signal as anything - I'm wondering how to discover that signal.

Yeah,

Unique technical skills generally aren't a great path to riches. The most uniquely skilled people in the world might be famous but they often aren't rich or at least not extremely rich.

Some combination of knowing what people want, being able to (directly or indirectly) organize people to create the thing, and knowing rich people willing to invest in your plan for doing it, is pretty the formula for most people gaining wealth in this society. There are other formulas but this is by far the most common and why fight the average if you're after success?

With LSA, I believe it suddenly became hot in 2007 because the belcor patent was set to expire.

> With LSA, I believe it suddenly became hot in 2007 because the belcor patent was set to expire.

Which has interesting implications about the extent to which patents help or hinder innovation…

Reply on news.ycombinator.com