I see N+1 everywhere
There's no reason for software to be slow anymore
431–440 of 530 posts
Re: There's no reason for software to be slow anymore
#432Yes, Seriously, where is the native versions of LLM created slack, vscode and other memory sucking Electron apps. No one should need Electron now as well..
Re: There's no reason for software to be slow anymore
#433It could very well be argued that a move towards more optimized, yet still human-readable languages is appropriate. In a _very_ generic sense, LLM's can pick up the slack of those languages being harder to write, AND hard to debug.
There is less of a need for super convenient, yet bloated languages now. For software that matters, and burns the most joules in the world, I hope we can also spend a few more tokens instead of doing it in interpreted python just because it's a bit easier for the LLM.
Re: There's no reason for software to be slow anymore
#434One of the biggest causes of slowness is just waiting for web requests. The fact that so much software is either online or built using the same stack even if it isn't, puts all that software in this blocked/waiting state constantly while using it. Anyone not in the US feels this even more since so much online is US hosted, 300ms for every little interaction adds up quick. If your software has the affordance of a wait…
Unfortunately, you either put all the commercially interesting bits on your own server and let your customers eat the latency, or you ship it to the edge and let piracy decimate your profits. I don't think there is any technical way out of this, and probably not any reasonable legal ways.
I feel like you're hitting at the real reason why so much of this happens. Not necessarily the piracy, but centralized control of the data and data flow. Even if it's slow as hell to send off that data to proprietary company servers (or rented cloud servers) to verify it, you're still verifying it. You, as in you the company, can't do that if it's local compute only. You can't control whether the user installed your paid plugin or some free alternative. You can't control someone stripping out libraries or code to remove intentional friction points designed to annoy them into a higher tier of the software. You can't control whether or not they update, or whether or not you can force the software into end-of-life with an update despite it still functioning.
If there's a connection to your services outside of the user's machine you can control all of that.
Re: There's no reason for software to be slow anymore
#435The peak of fast software was definitely Windows XP, Windows 7, and OS X Snow Leopard. I don't see us returning to that glorious era. Recently I was frustrated by Windows 11's seeming inability to open a context menu with acceptable speed - right click an item in the taskbar and there is nearly a 1000ms delay before the menu appears. That is unacceptable. When I need to run old software, I now try to the "minimum via…
As much as i dislike Windows, i agree that the versions of the OS from that era were really fast (except for maybe Vista)!! As far as OS X, I had used OS X back around ~2007 - 2010, but can't recall what versions it was...and it performed fine back then too. I've been on one or another linux distro since around 2003 or maybe 2004, and have been running linux as my primary driver for laptops and desktops since maybe 2010 or so(right after i uncoincidentally abandoned OS X). While any issues that linux runs into tend to mostly related to proprietary drives, the majority of the time, things run fast...just like Windows did of that older, golden period of performance - and many times much better! I don't say this to sway anyone to move over/start using linux...and, in fact, linux is still far from perfect! Rather, its to show that there is joy in computing that still exists somewhere in the world. I have found it in linux, but i'm sure others have found it elsewhere as well.
Re: There's no reason for software to be slow anymore
#436Earlier quoted context omitted.
> What I’m wondering now is can we reliably evolve python and have codex act as an extremely unreliable transpiler to the rust. Why you even start with Python at this point? Just write the Rust version straight up instead of porting things? Personally I used to use dynamic languages for most things, because development and maintenance is so much faster and easier, particularly for larger projects (granted you know ho…
Because it’s a port of a scientific process and that’s what the scientists work in.
Ah, sorry, I didn't understood you weren't the original author of the program you're porting :) Cheers for the additional context.
Re: There's no reason for software to be slow anymore
#437Earlier quoted context omitted.
> What I’m wondering now is can we reliably evolve python and have codex act as an extremely unreliable transpiler to the rust. Why you even start with Python at this point? Just write the Rust version straight up instead of porting things? Personally I used to use dynamic languages for most things, because development and maintenance is so much faster and easier, particularly for larger projects (granted you know ho…
What happens when LLMs will become either too expensive or unavailable?
I guess lots of Rust developers will find a lot of employment?
I mean what happens with network engineers when the biggest network of them all goes down?
Re: There's no reason for software to be slow anymore
#438Earlier quoted context omitted.
>but if the web were content addressed instead of server addressed you could then be serving that cache to your municipality even after it became disconnected from the rest of the internet. This is already possible without content addressing with CDNs. They can serve content from a local cache even when the host is disconnected from the internet.
Without content addressing how do I know that whoever holds the cache hasn't tampered with the content?
The integrity attribute of the element lets you provide a hash to ensure the content has not been tampered with.
Re: There's no reason for software to be slow anymore
#439The peak of fast software was definitely Windows XP, Windows 7, and OS X Snow Leopard. I don't see us returning to that glorious era. Recently I was frustrated by Windows 11's seeming inability to open a context menu with acceptable speed - right click an item in the taskbar and there is nearly a 1000ms delay before the menu appears. That is unacceptable. When I need to run old software, I now try to the "minimum via…
As much as i dislike Windows, i agree that the versions of the OS from that era were really fast (except for maybe Vista)!! As far as OS X, I had used OS X back around ~2007 - 2010, but can't recall what versions it was...and it performed fine back then too. I've been on one or another linux distro since around 2003 or maybe 2004, and have been running linux as my primary driver for laptops and desktops since maybe 2…
I dunno, Ubuntu got bloated. I tried to put 24 on an old Chromebook at it was atrocious. Going back to MX Linux made the UI relatively snappy.
Re: There's no reason for software to be slow anymore
#440Earlier quoted context omitted.
That's fair for a well-scoped subroutine: what I meant is that if you ask an agent to write a compiler and let it rip for a few days, you are going to be spending a few more days correcting the default behaviors in the distribution, which often do not tend towards hardware-oriented design. To correct those behaviors, you're going to write tools and skills, and that's going to help, but it is still clear that you are…
Yes, say "Build a compiler" will require you to clean up stuff if you leave the agent for days, but not because of the LLM or the quality of the tool, but because you hardly specified anything, so of course it's gonna make assumptions you need to correct. If you instead spend a day writing a proper specification, then ask the agent to spend a week implementing that, you'll need zero tools and skills afterwards to cle…
Bro, what the fuck do you think I’m doing? Do you think I don’t know about spec driven development?
What is the most complicated thing you’ve built with LM agents? Have you done it with a single spec? How novel was it?
This comment is so laughably “you’re holding it wrong” I can’t respond to you seriously.
> If you instead spend a day writing a proper specification, then ask the agent to spend a week implementing that, you'll need zero tools and skills afterwards to clean it up, because there won't be any misunderstandings, assumptions or other things
The set of software that has followed this process is measure zero.