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

A Conversation with Paul Graham [video]

youtube.com

Post reply on HN
Print view View original

91–100 of 159 posts

Re: A Conversation with Paul Graham [video]

#91

Post by testb »

mlthoughts2018 wrote: post #85
redshirtrob wrote: post #75

Earlier quoted context omitted.

This sentiment really bothers me. It's frequently used to disadvantage candidates from positions because they're not users of the service or product. It's just not generally true, but it's often repeated as though it were. It's demonstrably not true that you must be a user of a product to build a good product. To pick a random example: I doubt all pacemaker firmware developers have heart conditions that require pacem…

I work for a company that makes a very popular social app and I don’t use the app whatsoever, and refuse even to download early adopter versions to test. Sometimes people at work give me crap for this. But they know my engineering contributions are among the best in the company, so what can they say? It’s foolish to disqualify people for this. Imagine saying men cannot ever work for a team that delivers feminine hygi…

Hm the advice to be a user would apply for startup founders right? I definitely wouldn't expect engineers to be active users especially in a more mature company with PMs.

Re: A Conversation with Paul Graham [video]

#92

Post by sampl »

ransom1538 wrote: post #88

I love PG but listening to his advice on startups always felt like discovering the steps to create a nuclear bomb. Step 1) collect seven pounds of weapons grade plutonium Step 1 of a great startup is an amazing idea. Unless you have step one I think the advice is pretty non applicable. Sure, I get the whole camp “ideas are a dime a dozen”, but, a bad idea can fall apart with millions of VC ran by MIT grads and an ama…

It’s better than saying “you can build a nuclear bomb if you just wish it hard enough”.

> a bad idea can fall apart with millions of VC ran by MIT grads and an amazing idea can come to life with a few hundred bucks with a high school kid.

This is a actually really wonderful thing!

Unless you’re an MIT grad with no ideas about how to help customers hoping you could cash in by riding solely on your credentials I guess.

Re: A Conversation with Paul Graham [video]

#93

Post by irq-1 »

redshirtrob wrote: post #75
Razengan wrote: post #17

Earlier quoted context omitted.

You have to be a user of your own product to really make it good. Look at Apple. Most of how they got where they are is by Jobs, Ive et al. making what they wanted to see, and nitpicking at everything until they were satisfied. Also: "people don't know what they want until you show it to them." Of course you have to address customers' frustrations, but relying on just "data" and marketing feedback all the time often…

This sentiment really bothers me. It's frequently used to disadvantage candidates from positions because they're not users of the service or product. It's just not generally true, but it's often repeated as though it were. It's demonstrably not true that you must be a user of a product to build a good product. To pick a random example: I doubt all pacemaker firmware developers have heart conditions that require pacem…

Being "a user of your own product" is a cheap proxy for caring about quality.

Re: A Conversation with Paul Graham [video]

#94

Post by mlthoughts2018 »

testb wrote: post #91
mlthoughts2018 wrote: post #85

Earlier quoted context omitted.

I work for a company that makes a very popular social app and I don’t use the app whatsoever, and refuse even to download early adopter versions to test. Sometimes people at work give me crap for this. But they know my engineering contributions are among the best in the company, so what can they say? It’s foolish to disqualify people for this. Imagine saying men cannot ever work for a team that delivers feminine hygi…

Hm the advice to be a user would apply for startup founders right? I definitely wouldn't expect engineers to be active users especially in a more mature company with PMs.

Why founders? It seems like that will be true for some businesses and not others. Case by case basis.

Re: A Conversation with Paul Graham [video]

#95

Post by slededit »

golemotron wrote: post #64
mattmanser wrote: post #54

Earlier quoted context omitted.

What a load of nonsense. C#'s a great language and their dev tool, Visual Studio, is second to none.

In context, the Windows Paul was talking about used C, a horrific low level SDK, and a technology called COM. https://en.wikipedia.org/wiki/Component_Object_Model .NET did not appear until much later.

.NET integrates well with COM. You are using it all the time under the hood. COM was difficult because it was an OO framework that had to deal with:

- C++ ABIs that changed on every compiler update but still had to work.

-Support for C and Assembly which had no OO support at all.

-Liteweight enough to run on a computer with 8MB of RAM total.

Given these constraints it's unlikely you would do any better.

Re: A Conversation with Paul Graham [video]

#96

Post by danenania »

mlthoughts2018 wrote: post #84
RyanShook wrote: post #11

“The mistake you will most likely make as a startup founder is coming up with your own vision of what your product should be instead of getting feedback from actual users.”

But what is the point of operating your own company if not to accomplish your goals of what the company should be? I mean, I get it, you have to make customers happy in order to have a business. But this seems like something different. It seems like saying your goal should be to make money a priori and that defining vision based on feedback is a good way to do that. But what if that’s not your goal? What if your goal…

I think this is something that’s ridiculously hard in the beginning when your sample size is low, but as you get more users giving feedback, you will probably start to see some clear patterns emerge in what they like, what frustrates them, what features they ask for, etc. While individual pieces of feedback are often all over the place, with more data points, the solution that they all hover around starts to emerge. It can even start to feel ‘obvious’ what your product needs, even though it’s mostly stuff you never could have thought of in isolation.

As far as making money vs. mission, I think the issue is just that truly solving an important problem for lots of people is so absurdly hard that very very few entrepreneurs can manage to achieve it without devoting close to 100% of their focus to this challenge.

People might love that your company is financially transparent and socially responsible, but that won’t get them to use your product. So thinking much about these things before you start seeing product-market fit feels like putting the cart before the horse.

Re: A Conversation with Paul Graham [video]

#97

Post by nabla9 »

submeta wrote: post #23

Man he was really hating windows. To a point where he "wrote software for it without writing software for it directly"

Everyone how had experience from alternatives hated Windows.

You have now idea how bad, inelegant and even technologically behind the Window's was back then. They exploited their monopoly power to the max.

Microsoft hated and did not care about the internet protocols. They pushed NetBEUI instead. Early Windows socket API was crap and unreliable compared to BSD sockets. and Netscape and web apps provided major service for abstracted that shit away. I think Intel engineers had basically rewrite the TCP/IP stack for MS at some point.

Today people see X Window System as outdated, but at the time it was simpler, better and more robust than Windows. We actually bought number of gigantic but old dummy graycolor X window terminals and connected it to old 386 Linux server we used as a development machine and it was a great experience compared to anything Windows.

I was working in a small company that did mostly stuff for Unix (and linux) but unlike PG and Viaweb we decided to develop complex IDE to Windows and it had to work everywhere in the world. It was just horrible programming experience.

Re: A Conversation with Paul Graham [video]

#98

Post by cardimart »

goshx wrote: post #74
Razengan wrote: post #17

Earlier quoted context omitted.

You have to be a user of your own product to really make it good. Look at Apple. Most of how they got where they are is by Jobs, Ive et al. making what they wanted to see, and nitpicking at everything until they were satisfied. Also: "people don't know what they want until you show it to them." Of course you have to address customers' frustrations, but relying on just "data" and marketing feedback all the time often…

There are a lot of CEO’s that use the Steve Jobs argument to justify their approach of thinking they know what customers want. Chances are you are not really building a product that is reinventing a market and creating new solutions for previously non existing problems, like the iPhone. Then you fall into the hole of thinking you can imagine how people want to solve a well known problem and end up finding out later t…

"creating new solutions for previously non existing problems, like the iPhone."

I have a problem when I hear things like that. The problems are always there, the needs are always there. They're just satisfied better by the new technology. But the latent needs were there to begin with.

Re: A Conversation with Paul Graham [video]

#99

Post by cookiecaper »

nabla9 wrote: post #97
submeta wrote: post #23

Man he was really hating windows. To a point where he "wrote software for it without writing software for it directly"

Everyone how had experience from alternatives hated Windows. You have now idea how bad, inelegant and even technologically behind the Window's was back then. They exploited their monopoly power to the max. Microsoft hated and did not care about the internet protocols. They pushed NetBEUI instead. Early Windows socket API was crap and unreliable compared to BSD sockets. and Netscape and web apps provided major service…

Windows is still a disaster for everything but the MS-specific stack exposed through Visual Studio, including (especially?) sysadmin.

How do you install software across a fleet of dozens of Windows machines? There are solutions, but they are embarrassingly poor in comparison to what's available on other systems. Salt, Ansible, et al have added half-hearted Windows support over the years, but it's pretty shallow. SCCM is a massive can of worms and very easy to FUBAR.

MS's mental model of "update systems by performing clicks 1, 2, 3, and 4" is just fundamentally limited, as compared to the text-based *nix model. Even MS has been forced to come to terms with it, which is why they're finally adding real pty support, etc.

Re: A Conversation with Paul Graham [video]

#100

Post by testb »

mlthoughts2018 wrote: post #94
testb wrote: post #91

Earlier quoted context omitted.

Hm the advice to be a user would apply for startup founders right? I definitely wouldn't expect engineers to be active users especially in a more mature company with PMs.

Why founders? It seems like that will be true for some businesses and not others. Case by case basis.

True, best not to make generalizations.
Post reply on HN