Fork me on GitHub
#off-topic
<
2018-04-22
>
tbaldridge00:04:11

@qqq, we see this same effect in the video game industry. Pre-built tools and engines (Unity, UE4, Amazon's engine, etc). keep getting more advanced and are very cheap, asset stores are everywhere. But what we don't see is an explosion of games being created in a few months, at least not good ones.

tbaldridge00:04:24

As the tools increase in power, so do the demands of the public. And like startups it's not enough to say "it's like X but better". You can't just clone PUBG and get millions of players. Hearthstone clones are everywhere, and yet none of them do very well at all. In the end what sells games is how unique they are, how they scratch an itch that players have.

tbaldridge00:04:04

I think that applies to startups. Tools increase in power, but overall time-to-market doesn't, due to increased demands and the real problem: you have to "have a good idea" for the company to stick

tbaldridge00:04:18

Funny thing about that too: Hearthstone is insanely successful, netting about $400mil in sales in one year, it's mostly written in Unity, by a team of 5 people, ramping up to about 15 at the end of development. But what made it successful was A the marketing of Blizzard, and B some rock solid game design. Neither of those are tech problems.

tbaldridge00:04:21

Same applies to Uber, successful, but not because of their tech. AirBnB's tech is mostly a non-issue, finding people willing to rent their homes is the real problem. Instagram is the same, for what they built, you could almost built that on any infrastructure, what made them popular was social acceptance of their services.

qqq01:04:28

@tbaldridge: although this was not mentioned in my earlier examples, what interests me is things like: Minecraft, http://Slither.io, PlentyOfFish, Instagram, ... other apps that were built with very small teams.

qqq01:04:12

I'm not making the argument of "open source + cloud computing" lets one duplicate $UNICORN with half the people. What I am interested in is: does "open source + cloud computing" allow one to replicate the tech stack with half the headcount / lines of code.

qqq01:04:52

I also completely agree that the success of these startups are in large parts due to marketing / fund raising / non-tech issues.

benzap01:04:37

@qqq,if you're interested in lean startups, there's a good book on it

benzap01:04:48

i'd also recommend zero to one, and innovator's dilemma

benzap01:04:15

The show silicon valley uses a lot of the terminology in this book

benzap01:04:31

The Innovator's Dillemma: http://a.co/2kYi8D2

benzap01:04:27

Zero to One, here's a youtube video (not sure if it's good) https://www.youtube.com/watch?v=Y3vYPHKt7e0

benzap01:04:36

The book is well done though

qqq02:04:13

@benzap: Thanks for the resources. Lean Startups is definitely very interesting. However, it's a different question from "pick a fixed point in terms of what software should do; given current progress of cloud computing / open source, is it likely that in 18 months, we can implement it in half the lines of code"

benzap02:04:31

Oh, I guess I misunderstood

benzap02:04:50

I thought it wasn't about the software, but about the product that was made

benzap02:04:46

ex @tbaldridge's example with twitter is a good example of lean manufacturing, they only built at scale when it became a requirement

dpsutton03:04:03

It was a rails app to begin with I think

john03:04:09

How can it ever really be "half the lines of code"? Are we not counting dependencies? If we include deps, lines of code is increasing over time. If we aren't including deps, you could could say 1 line of code can depend on 10,000 lines written yesterday. Or are you saying that, per project, lines of code are getting smaller over time? I'm not sure what the metrics on that might be, but it obviously can't go on forever.

benzap04:04:30

On another note, Canada is going to allow people to grow their own marijuanas soon

benzap04:04:47

You can grow 4 marijuanas in your household

benzap04:04:22

I've been getting into hydroponics, so I a great business idea would be to develop a product that gets marijuanas to people

benzap04:04:32

because marijuanas.

sveri07:04:59

we had one in our shared appartment, back then during study times. It was groing well and big, good times before us when we would finally harvest it. Before we all went on vacation for two weeks one of our members had the glorious idea to use fertilizer on it, without reading the manual first. Of course he used way to much of it and back we came, the plant died a lonesome and gruel death. Good times 😄

😁 4
qqq18:04:09

@john: we are not counting open source libraries; we are not counting aws services advancements in aws/open source makes it easier to create software; is there a way to measure "how much easier" ?

john18:04:50

Sounds sort of like, "the same amount of code can leverage more code, over time"

john18:04:23

Leverage through indirection

john18:04:56

There's probably a growth rate that can be measured, as a consequence of human factors, but I doubt some hard rule exist that prevents that inderection leverage from growing faster or slower, for any given context

john19:04:41

Another question is, is functionality actually growing at the same speed? Or is the growth just an optimization over existing functionality?

john19:04:47

In that optimization, you're trading space for time, leveraging existing code and memory. Whereas some optimization trades time for space, eliminating code and memory and leveraging computed values.

john19:04:21

Perhaps there's a correlation there to inductive vs deductive reasoning. Going from generality to specificity and vice versa.

benzap20:04:45

@sveri That's hilarious! I never touched the stuff while I was in university, but my roommate was big into homebrewing. He owns a microbrewery now!

benzap20:04:20

He was a software engineer who went on to get his masters in software/electrical, and ended up following his dreams to become a brewer

benzap20:04:09

Wish I could do the same in hydroponics