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

Software Disenchantment (2018)

tonsky.me

Post reply on HN
Print view View original

481–490 of 504 posts

Re: Software Disenchantment (2018)

#481

Post by johnwatson11218 »

ssivark wrote: post #353
johnwatson11218 wrote: post #119

Earlier quoted context omitted.

I'm suggesting that you take an existing system and build up a corpus of request/response pairs. Then you use the LSTM to build a prediction model so that given a request it will tell you that the current production system will produce the following sql statement and this http response. Once the LSTM's output is indistinguishable from your current production system , for all use cases, then you replace the production…

Since you seem sincere, I’d like to mention that neural networks (as of now) are a complete clusterfuck for a problem with as much structure as you’re describing. There are no known ways to impose the relevant constraints/invariant on neural network behavior — and stop them from producing junk — leave alone doing something useful. That Karpathy article is pure hype with very little substance (like most commentary on…

Thanks for the feedback. I don't consider myself a world-leading expert on neural networks by any means.

I do want to point out that I'm thinking of doing this on a very limited website, not a general purpose thing that replicates any possible website. When I imagine the complexity of a modest CMS or an online mortgage calculator I think that it is much less complex than translating human languages. The fact that web code has to be so much more precise than human language actually makes the task easier. But to be fair, I'm all talk at this point with no code to show for it. So I will keep these comments in mind, this thread has been helpful for helping me think through some of this stuff.

Re: Software Disenchantment (2018)

#482

Post by johnwatson11218 »

jodrellblank wrote: post #325
johnwatson11218 wrote: post #179

Earlier quoted context omitted.

I don't think a caching layer would work. One example would be an online mortgage estimator. You input the loan amount, interest rate, length of loan etc. all as http input parameters. I'm suggesting that the LSTM can eventually figure out that those variables are being used by the application code to go in to a formula. That application code and its formula would all be replaced by the LSTM. I just don't know how yo…

Seems that either the NN memoizes all the inputs and outputs until the function is totally mapped - then functions as a memoized lookup table, or the NN has discerned what the mortgage calculation is, and is doing exactly the calculation your {Python} backend does, but migrated into an NN middleware layer instead, which sounds like it would be slower. And then you're hoping that the NN would act like a JIT compiler/o…

If I ever get any actual code running I'll try and post it here.

Re: Software Disenchantment (2018)

#483

Post by jiofih »

ptx wrote: post #479
jeffmcmahan wrote: post #222

Earlier quoted context omitted.

I completely agree, here. React has replaced the DOM, and it's pretty fast, pretty efficient when you understand its limitations... but when you start rendering to the canvas or creating SVG animation from within react code, everything is utterly destroyed. Performance is 1/1000 of what the platform provides. I have completely stopped using frameworks in my day-to-day, and moved my company to a simple pattern for upd…

Maybe React has the advantage as the project grows? From what I understand it batches updates from all the different components on the page, avoiding unnecessary reflows that might easily creep in when you do things the old-fashioned way.

It makes it even slower, and you have to manually optimize component renders. Sure you get less reflow due to dom diffing, but tons higher cpu time.

Re: Software Disenchantment (2018)

#484

Post by jiofih »

engineeringwoke wrote: post #413
somatic wrote: post #337

Earlier quoted context omitted.

jQuery: 88KB, standard everywhere, one entity responsible for all of it, people know what it is and what it does, if it breaks you know what went wrong and who to blame. Literally anything built with NPM: megabytes? tens of megabytes? in size, totally inscrutable, code being pulled in from hundreds of megabytes of code in tens of thousands of packages from hundreds or thousands of people of unknown (and unknowable) c…

Write your own framework-like code with just jQuery and watch it turn into a pile of mush. React is many things, but it is absolutely better than jQuery or Backbone. People always mis-use new technology; that isn't React's fault.

My whole argument is that _it is_. I don’t know why we are comparing to jQuery though, they are not replacements for each other.

Re: Software Disenchantment (2018)

#487

Post by shiftless »

uncle_j wrote: post #27
shiftless wrote: post #2

To put a finer point on it, software today is absolute garbage. I've been screaming about this for decades. All of this bloated 'shitware' today is the result of it having been written by people who a) have no deeper understanding of what the computer is actually doing; also known as typical Python/Java/etc/etc/etc programmers, and/or b) simply not giving a damn about conservation of resources--as further evidenced b…

> To put a finer point on it, software today is absolute garbage. I've been screaming about this for decades. That statement contradicts itself. Is this satire? > All of this bloated 'shitware' today is the result of it having been written by people who a) have no deeper understanding of what the computer is actually doing; also known as typical Python/Java/etc/etc/etc programmers, and/or b) simply not giving a damn…

> 'That statement contradicts itself. Is this satire?'

My comment was the first one on this article. 473 comments later, it looks like over 90% of the other commenters agree with my observations, considering 90% of the other posts could have been written by me also. So the real question is, what do we know that you don't?

I started programming in 1991. Since the 90s I've watched as software has been steadily circling the drain. Many others have noticed it too, as shown by this discussion. The growing discontent is palpable. Where have you been during this time?

> 'Everything has a cost / benefit associated with it. Pretending otherwise show how little you know about engineering.'

Typical excuse from a lazy, talentless hack, excusing his shit code. What you are doing is not engineering.

> 'It not like most developers these days work on large software projects that are normally poorly costed, estimated, planned usually with stifling restrictions because certain "enterprise" technology is mandated by some architect who hasn't written a line of code in a couple of decades. Most people like a consistent regular income which allows them to support themselves and their family. I suspect these concerns are more important then incoherent ramblings of some guy on HN.'

Typical, common, ordinary, everyday selfishness. "Well I thought about taking a stand for what's right, but decided against it Because Reasons." Pathetic.

What are you doing that is actively making the world of computers, or even the world in general, a better place? Is that something you care about at all? Or do you prefer to just cash that check while making bullshit excuses?

I live in a cabin in the woods just because I'm not going to swim in the shark pit that you call home. Nor am I going to actively fuck up the world just to make a paycheck. It's called having principles.

Re: Software Disenchantment (2018)

#488

Post by shiftless »

mc3 wrote: post #347
shiftless wrote: post #2

To put a finer point on it, software today is absolute garbage. I've been screaming about this for decades. All of this bloated 'shitware' today is the result of it having been written by people who a) have no deeper understanding of what the computer is actually doing; also known as typical Python/Java/etc/etc/etc programmers, and/or b) simply not giving a damn about conservation of resources--as further evidenced b…

Did you notice the subtle contradiction in your post? > they're more about the foundations of our entire Western 'civilization.' > talentless 'developers' It's not talentless developers, it's the foundations of civilization. Capitalism, chasing the $ not the writing of beautiful software. If an Electron app makes me more money (e.g. by getting to market quicker) than a well crafted native app, then the Electron app w…

It is quite simple. Either you are helping the world, or you are hurting it. Either you do what's right, or you make excuses and continue in the course of wrongdoing. The choice is yours. And that is the existential problem we face--getting people to admit their own wrongdoing and culpability and to repent. I don't want to hear any fucking excuses from anyone about how they were just trying to make a buck or whatever. You don't have to do a damn thing, and sure as hell shouldn't, if it's fucking up the world! "Keeping up with the Joneses" is NOT a valid excuse; nor is "just following orders."

Re: Software Disenchantment (2018)

#490

Post by LargoLasskhyfv »

clouddrover wrote: post #326
thanhhaimai wrote: post #45

Earlier quoted context omitted.

> "how much would you prioritize a car that used only 0.05 liters per 100km over one that used 0.5? What about one that used only 0.005L?". I'd say that at that point, other factors like comfort, performance, base price, etc. become (relatively) much more important. I'll prioritize the 0.005L per 100km car for sure. That means the car can be driven for all its expected lifetime (500k km) in a single tank of gas, fill…

> I'll prioritize the 0.005L per 100km car for sure. That means the car can be driven for all its expected lifetime (500k km) in a single tank of gas, filled up at the time of purchase! It's a nice idea but it wouldn't work. The gasoline would go bad before you could use it all. Plug-in hybrids already have this problem. Their fuel management systems try to keep the average age of the fuel in the tank under 1 year. T…

There is [1] https://duckduckgo.com/?q=alkylate+petrol which is said to last anything between two years and up to ten years, depending on the mixture while burning rather clean.
Post reply on HN