Post #2065649

2026-05-05 20:43 UTC

When I was young, I learned, and was taught, how to make the computer to work efficiently and correctly, in my computer science degree. Now it is the opposite. Do brute-force search using giant farms of computers, using a huge amount of energy and water, and get results that are not guaranteed to be correct any more. And I was discussing with a colleague this morning that my 2001 laptop ran faster than my current top-range computer for everyday tasks. Of course, it had a much worse CPU and much less ram. And of course the software for things we still do *now* was much faster *then*. I still have that laptop from that time running Ubuntu 4.10 from 2004 in my personal museum of computers. You would be amazed how responsive the system is for everything we do every day with a computer. I recently tested it with my son, because he was curious to see how things were then. So we are using more powerful hardware for getting a poorer experience. The new computers are much better for some things, such as running Agda. But, for everything else I happen to do, they were just as fast, because people programmed them in a more efficient way (they had to - there was no other way).

Replies (25)

@rl_dane@polymaths.social 2026-05-05 22:20

@MartinEscardo@mathstodon.xyz I feel bad for saying it, but I often tell my friends, "Computer Science has reached its rabid-orangutans-smearing-feces-on-the-wall level of devolution." I truly cannot comprehend the level of brainrot and insanity that is SOP these days. Even before all of this "A.I." giga-grifter nonsense. I'm sad for my friends still in the field. And sadder that it has become such a caricature of itself.

Open ##2065648
@MichaelTBacon@social.coop 2026-05-05 21:04

@MartinEscardo I remember when you could still install the old NCSA Mosaic binary on some newer hardware, which I think isn't possible anymore with chip changes. Anyway, a coworker challenged me to try it and I was like sure. I cannot begin to describe how fast it was. Eyeblink fast. Screamingly fast. Comically fast -- I literally guffawed when it loaded one page. Like the whole webpage was a JPG it just loaded from disk. And of course a bunch of stuff didn't work, but most of it was crap I didn't want anyway.

Open ##2075002
@jamesmarshall@sfba.social 2026-05-05 21:04

@MartinEscardo I'm a fan of deterministic computing myself.

Open ##2075006
@luc0x61@mastodon.gamedev.place 2026-05-05 22:26

@MartinEscardo Even back then there were horrible examples. See Windows 2000, which was able to make slow any fast machine of the time. Although Microsoft had a great idea: label it as "Professional". Every "serious" user had to go with a deadly slow machine, but running a professional OS! 😅

Open ##2075017
@synlogic4242@social.vivaldi.net 2026-05-05 22:36

@MartinEscardo I'm writing book on HPC. but I do worry about a possible shrinking market demand by people who might ever care anymore. I am gambling there will always be a slice, and niches, where it will always matter

Open ##2075020
@allypally@mastodon.scot 2026-05-05 22:56

@MartinEscardo Way back in the day, when undergraduate projects were in Fortran IV, the big issue in computer science was how to make calculations more efficient. But they gave up on that and threw memory and power at it. Bloatware running on bloatmachines.

Open ##2075021
@unfmeghan@dmv.community 2026-05-06 00:44

@MartinEscardo I do feel like the entire field of Computer Science has completely abandoned the principle of efficiency.

Open ##2075023
@valpackett@social.treehouse.systems 2026-05-06 02:33

@MartinEscardo outside of the slop thing, that's quite the rose-tinted-glasses take

Open ##2075027
@toroidalcore@masto.hackers.town 2026-05-06 03:25

@MartinEscardo

Open ##2075028
@x_hace@mastodon.green 2026-05-06 05:19

@MartinEscardo and honestly the developers themselves often have the best computers available. Which makes them not having the same experience at all. Back in the days when working on a windows app built in MS Access the fact of not having a filled test database, forgetting an index for an often run query could lead to customer support making the client buy new hardware instead of validating the issue with the software team.

Open ##2075030
@vnikolov@ieji.de 2026-05-06 06:04

Parkinson (Cyril Northcote) noted that work expands to fill all time available for its completion. Similarly with software vs. hardware resources. @MartinEscardo

Open ##2075032
@1div0@mastodon.social 2026-05-06 06:39

@MartinEscardo @Gargron retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// retro:// future:// ad ∞

Open ##2075033
@WvOostveen@mstdn.social 2026-05-06 08:21

@MartinEscardo And when you buy some new tech nowadays, it seems that you don't own it, but you are subordinate to it's commands. (Just returned a new phone after several hours of fight with the beast. )

Open ##2075036
@grishka@friends.grishka.me 2026-05-06 08:36

I once booted MS-DOS on my 2011-ish crappy laptop that I use specifically for testing things. Was amazed by how instant everything is. There's no perceptible input lag at all. Things like dir take zero time as far as I'm concerned. Games launch instantly as well. This made me wonder how fast would Windows 9x be on modern hardware.

Open ##2075037
@bayindirh@mastodon.sdf.org 2026-05-06 08:48

@MartinEscardo I completely agree, and I still strive to write lightweight, fast and responsive code, and take pride at that. For anyone debating me on this, I point them to demoscene. Namely "elevated", ".the.product", and "debris".

Open ##2075038
@antsu@p.antsu.net 2026-05-06 08:54

@MartinEscardo There is a quote attributed to someone named Henry Petrosky that I think about a lot: "The most amazing achievement of the computer software industry is its continuing cancellation of the steady and staggering gains made by the computer hardware industry."

Open ##2075039
@GoteerTxuria@derg.social 2026-05-06 11:22

@MartinEscardo@mathstodon.xyz This is something that bothers me in general because it means hardware that is perfectly fine gets left behind by virtue of modern software or even just websites being too much, despite not providing much more to justify the extra required power. It especially bothers me with phones. What do you mean the app launcher is running slow on this old phone? Do the background services really need to use up all the resources when they do almost the same stuff as back when the phone was new?

Open ##2078625
@phf@dmv.community 2026-05-06 11:25

@MartinEscardo Welcome to 1995. https://dn710706.ca.archive.org/0/items/pdfy-PeRDID4QHBNfcH7s/LeanSoftware_text.pdf

Open ##2078626
@FrankEndrullat@mastodon.social 2026-05-06 12:32

@MartinEscardo @Gargron Also today I noticed yet again that most webpages seem to be optimised for phone screens, meaning that UI elements are comically big when visited with a desktop browser. This I find particularly annoying as there was a time when webpages used frameworks that would display them correctly and efficiently based on the browser and device you used to visit them. Desktop, phone, tablet was all taken care of and elements displayed appropriately. Even UI frameworks are lazy now

Open ##2092779
@pointlessone@status.pointless.one 2026-05-06 17:39

@MartinEscardo@mathstodon.xyz It's not a new thing. A good decade before your laptop Wirth plead for leaner software. Three decades later we can't buy normal laptops because there's not enough RAM on the planet.

Open ##2097708
@publicvoit@graz.social 2026-05-06 21:24

@MartinEscardo Mostly because almost anything nowadays goes over the #Internet which is still adding substantial delay because of extremely inefficient #programming, sometimes causing hundreds of requests for one single user interaction. See also human scale explanations like https://pesin.space/posts/2020-09-22-latencies/ Back in the old days, we had software that was much more optimized on algorithm level, less layers, everything local since network connection was a rare thing to have. 👴 #HCI #CHI #computers #efficiency

Open ##2111845
@odr_k4tana@infosec.exchange 2026-05-06 21:30

@MartinEscardo the thing is that the computers back then were not what they are now. Even 25 years ago, they were used for very specific things and had a very limited range of tasks you could use them for. Now the same machine is expected to work for CAD, games, office, video, audio, network etc. We moved to generalist machines. It's only natural that you get efficiency losses that way.

Open ##2111846
@grishka@friends.grishka.me 2026-05-06 23:01

💡𝚂𝗆𝖺𝗋𝗍𝗆𝖺𝗇, a 2011 CPU is fast enough that this difference shouldn't matter. I'm sure its memory bus bandwidth is enough to transfer an entire VGA framebuffer every frame, many times over. GPUs are really good at rendering images and videos very fast, and all modern web browsers use hardware-accelerated rendering as much as possible. So it's not that either. I'd blame modern OSes and their schedulers.

Open ##2123851
@tschundler@leds.social 2026-05-07 00:25

@MartinEscardo I noticed today that that PayPal app on my phone is 650MB. It's hard for me to imagine how it can be a full CD worth of content. (And how much bandwidth does it waste on every update - I don't think updates are patch files like they used to be)

Open ##2123852
@t0maz@floss.social 2026-05-07 05:07

@MartinEscardo@mathstodon.xyz I also remember times when computers where booting 5 minutes or when you need to remember to run command to park disk drive headers before powering off system to prevent their permanent damage. Hell, I do even remember when loading a game from tape was taking around half an hour :o But I do agree that due to hardware constraints software was often much better designed and there was no bloatware everywhere (Win11&mobiles) or unnecessary complications like systemd etc

Open ##3180598