I guess it does make their communication feel a bit more personal, especially since they usually don't use corporate language on personal twitter accounts.
Truly a terrible day in AI. We shall all rise against the 5-hour limit.
Also some Sol token usage optimizations are coming, as described in the tweet. Seems like OpenAI is really trying to remain the leader in token efficiency.
I'd say it's more like going to a Ferrari dealership and they tell you they will build a car for you and bill you per gram of parts used. It might also not work. And they might also not build it ever, really — but they will bill you for any attempts to build it.
That's fine though. If that doesn't work for you, don't buy. There are all manner of situations where what one wants or needs and what they get don't match up well. You don't price out every situation - it's take it or leave it. Pricing in a way that is somehow based on cost structure at least enables the provider to work to reduce the cost and hence price and win. Costco prices at a small margin above cost, they don't price "if this meets value prop X, pay Y and if only value prop A, pay B".
Yes that's true, although calculating the exact discount you get using a subscription could be difficult, especially since the labs often don't state their actual limits and just tell you you get "more usage" or "even more usage" or "5x the usage" (but how much exactly is x?).
It is clear we are getting a very meaningful discount by using a subscription though. I once checked how much would I pay for Codex after about 3 months of non-daily use if I only bought "credits". Couldn't believe it when it came out to be north of $1000 (I paid them $20/month in that period).
Well yeah this benchmark I've quoted is quite large, I believe they use thousands of diverse tasks, so this average cost per task could be a more accurate representation of how expensive each model actually is to run.
- you get a slower first load (cannot progressively fetch resources as they're needed)
- can't reuse a stylesheet, script or image on a different page (each has to have their own copy)
- can't cache commonly used files
- can't make granular changes to specific parts of the code. user has to reload everything each time.
- can't set a proper content security policy
And many more! It's cool for a tiny demo but for anything serious you wouldn't want a single (extremely ugly) HTML file.
I'm not advocating this development approach, but I also think some of your reasons aren't particularly robust when scrutinized.
> can't reuse a stylesheet, script or image on a different page (each has to have their own copy)
Isn't the point of a single HTML file that you don't have different pages that would need to reuse those assets?
> can't cache commonly used files
You can still cache the HTML.
> can't make granular changes to specific parts of the code
Pretty sure text editors can edit text regardless of whether it's a single file or multiple files.
> can't set a proper content security policy
I'd have thought a single page HTML file could negate the need for a CSP since you no longer have any resources accessible via a URI that you need to limit access to.
> you wouldn't want a single (extremely ugly) HTML file
Ugliness a subjective.
---
I think your first load point is the strongest one. But I'd also throw in "it's harder to develop a good single page HTML"
You could probably mitigate some of that difficulty by having a build script (like static site compilers) but then you have to ask yourself if you're introducing more complexity than you're attempting to solve.
If you need a build script, your app is probably too big to be a single .html already.
> You can still cache the HTML.
But then any time you update any part of the app, the user has to re-download the whole thing. It is also a problem with many traditional frontend builds, but if you carefully split it into chunks, you can update parts of the app while most of it stays cached.
The strong point of single .html apps is, of course, that you can run them locally without setting up a web server or anything, and deploy anywhere you want by just uploading it. But it is a fairly niche thing IMO.
Alright, this is pretty cool. I don’t see any reason for it to be a single-file web app, though!
Going through your article on the topic [0], I think the only strong point that SFWAs have which SPAs don’t is: “you can download and run them completely offline”. This highlights the best use case for SFWAs – it is something that you might want to download and run offline. Hypervault [1] is a better example here, IMO.
SFWAs do have drawbacks, mainly the inability to cache things independently (it’s all-or-nothing), or to download things in parallel. So basically it boils down to two questions:
• Is my app something users would want to run offline?
• Will I update it frequently enough for the cache problem to matter?
If it still makes sense to distribute your app as an SFWA, but you have to use a bundler, you can have the best of both worlds by making two builds: one as an SFWA, and one as a “traditional”, chunked static SPA. It should be fairly easy (just make two different configs, building from the same source).
Pretty graph! iOS Safari, takes just under a second probably to load each time I pan across the graph. Wonder what it’d take to bring it all the way to smooth scrolling—too demanding on resources?
Looking again, figure it has to do all the math again with each pan/drag, so minor latency makes sense.
This is an example of a single html-file that I serve: https://basketball.4e4.in
It weighs in at around 30 KB including a favicon, and it's quite functional.
In this case, it's 45kbyte over the network. Instead of very impractical, we might agree that it can be not ideal for many apps, but it's an interesting and valid way of "packaging" a tool / service.
If someone chooses this deployment method, it's likely that they have the ability to design with minimal resources.
My take: hitting three API endpoints doesn’t require a bundler or 400 MB of node_modules — most apps, especially hobby single‑page projects, can and should be built without a framework.
reply