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

Fast

catherinejue.com

Post reply on HN
Print view View original

171–180 of 450 posts

Re: Fast

#171

Post by sipjca »

ilyakaminsky wrote: post #69

Fast is also cheap. Especially in the world of cloud computing where you pay by the second. The only way I could create a profitable transcription service [1] that undercuts the rest was by optimizing every little thing along the way. For instance, just yesterday I learned that the image size I've put together is 2.5× smaller than the next open source variant. That means faster cold boots, which reduces the cost (and…

ive approached the same thing but slightly differently. i can run it on consumer hardware for vastly cheaper than the cloud and don't have to worry about image sizes at all. (bare metal is 'faster') offering 20,000 minutes of transcription for free up to the rate limit (1 Request Every 5 Seconds)

https://geppetto.app

I contributed "whisperfile" as a result of this work:

* https://github.com/Mozilla-Ocho/llamafile/tree/main/whisper....

* https://github.com/cjpais/whisperfile

if you ever want to chat about making transcription virtually free or so cheap for everyone let me know. I've been working on various projects related to it for a while. including open source/cross-platform superwhisper alternative https://handy.computer

Re: Fast

#172

Post by stevage »

KPGv2 wrote: post #157
Eji1700 wrote: post #114

Earlier quoted context omitted.

Information density and ease of identification is the antithesis of "engagement" which often has some time on site metric they're hunting. If you can find what you want and read it you might not spend 5 extra seconds lost on their page and thus they can pad their stats for advertisers. Bonus points if the stupid page loads in such a way you accidentally click on something and give them a "conversion". Sadly financial…

The one and only thing I'd do is make the font bigger and increase padding. There's overwhelming consensus that you should have (for English) about 50–70 characters per line of text for the best, fastest, most accurate readability. That's why newspapers pair a small font with multiple columns: to limit number of characters per line of text. HN might have over 100 chars per line of text. It could be better. I know I c…

Increased padding comes at the cost of information density.

I think low density UIs are more beginner friendly but power users want high density.

Re: Fast

#173

Post by stevage »

cyanydeez wrote: post #139
hn_throwaway_99 wrote: post #104

Just want to say how much I thank YCom for not f'ing up the HN interface, and keeping it fast. I distinctly remember when Slashdot committed suicide. They had an interface that was very easy for me to scan and find high value comments, and in the name of "modern UI" or some other nonsense needed to keep a few designers employed, completely revamped it so that it had a ton of whitespace and made it basically impossibl…

Ive wanted tp poll HN about how many people actively track usernames. With IRC its basically part of the task, but every forum i read, its rare that i ever consider whose saying what.

Yep. Dang is basically the only one I notice.

Re: Fast

#174

Post by ChrisMarshallNY »

pron wrote: post #152
cycomanic wrote: post #146

Earlier quoted context omitted.

I've already written about this several times here. I think the current trend of LLMs chasing benchmark scores are going in the wrong direction at least as programming tools. In my experience they get it wrong with enough probability, so I always need to check the work. So I end up in a back and forth with the LLM and because of the slow responses it becomes a really painful process and I could often have done the ta…

Programmers (and I'm including myself here) often go to great lengths to not think, to the point of working (with or without a coding assistant) for hours in the hope of avoiding one hour of thinking. What's the saying? "An hour of debugging/programming can save you minutes of thinking," or something like that. In the end, we usually find that we need to do the thinking after all. I think coding assistants would end…

I do both. I like to develop designs in my head, and there’s a lot of trial and error.

I think the results are excellent, but I can hit a lot of dead ends, on the way. I just spent several days, trying out all sorts of approaches to PassKeys/WebAuthn. I finally settled on an approach that I think will work great.

I have found that the old-fashioned “measure twice, cut once” approach is highly destructive. It was how I was trained, so walking away from it was scary.

Re: Fast

#175

Post by globular-toast »

Linus Torvalds said exactly that in a talk about git years ago. It's crazy to think back how people used to use version control before git. Git totally changed how you can work by being fast.

Re: Fast

#176

Post by HarHarVeryFunny »

hn_throwaway_99 wrote: post #104

Just want to say how much I thank YCom for not f'ing up the HN interface, and keeping it fast. I distinctly remember when Slashdot committed suicide. They had an interface that was very easy for me to scan and find high value comments, and in the name of "modern UI" or some other nonsense needed to keep a few designers employed, completely revamped it so that it had a ton of whitespace and made it basically impossibl…

I don't think it was UI that killed Slashdot. The value was always in the comments, and in the very early years often there would be highly technical SMEs commenting on stories.

The site seemed to start to go downhill when it was sold, and got into a death spiral of less informed users, poor moderation, people leaving, etc. It's amazing that it's still around.

Re: Fast

#177

Post by IshKebab »

kccqzy wrote: post #158
Liftyee wrote: post #54

I always have to remind myself of the bank transfer situation in the US whenever I read an article complaining about it. Here in the UK, bank transfers are quick and simple (the money appears to move virtually instantly). Feel free to enlighten me to why they're so slow in the US.

I am convinced that this is in some cases a pro-consumer behavior. A credit card company once pulled money from my bank via ACH due to the automatic payment feature I set up, but that bank account didn't have enough money in it. The bank sent me at least two emails about the situation. I finally noticed that second email and wired myself more money from a different account. The credit card company didn't notice anyth…

In most cases it's definitely better for it to be fast. For example I sold a buggy face to face today and they paid me by bank transfer, and the reason we could do that was that I had a high confidence it would turn up quickly and they weren't trying to scam me. It actually took around 1 second which is really quite fast.

You don't need slow transfers to get an opportunity to satisfy automatic payments. I don't know how it works but in the UK direct debits (an automatic "take money from my account for bills" system) gives the bank a couple of days notice so my banking app warns me if I don't have enough money. Bank transfers are still instant.

Re: Fast

#178

Post by brap »

Yep. I often choose LLM apps not because of how great the model is, but how snappy the UI feels. Similarly I might choose the more lightweight models because they’re faster.

Re: Fast

#180

Post by IshKebab »

SatvikBeri wrote: post #43

I've noticed over and over again at various jobs that people underestimate the benefit of speed, because they imagine doing the same workflow faster rather than doing a different workflow. For example, if you're running experiments in one big batch overnight, making that faster doesn't seem very helpful. But with a big enough improvement, you can now run several batches of experiments during the day, which is much mo…

I think people also vastly underestimate the cost of context switching. They look at a command that takes 30 seconds and say "what's the point of making it take 3 seconds? you only run it 10 times in a day; it's only 5 minutes". But the cost is definitely way more than that.
Post reply on HN