> do you think there's any appetite in people paying for this type of tool which lets you spin up infra on demand and gives you all the capabilities built so far?
(I'm not the author) The easiest way to charge for this kind of software is to make it SaaS, and I think that's pretty gross, especially for a CLI tool.
> I'm skeptical and I may just release it all as OSS
It doesn't have to be one or the other: you could sell the software under libre[1] terms, for example.
Agreed that SaaS feels ugly. Agreed on selling it as libre. The only thing I could imagine charging for in a subscribed model would be for services that are hosted (e.g. instance/gpu provisioning/monitoring/maintenance in the cloud) while also offering the ability to self-host the same machinery.
I often go one step futher by appending a short random identifier, `{service}.{id}@{domain}`, to make it harder to guess (in case someone learned of my email address policy).
I'm with you. Surprised by the negative reactions here.
A possible piece of the puzzle: I originally read the article on mobile, no issues. Then I opened it on my desktop, and found the design quite jarring. The margins are much too large for my taste, forcing the text into a single narrow column, and the header animations were distracting and disorienting (fortunately the page works perfectly with JavaScript disabled). Perhaps this triggered people?
I really, really, really, really, really, really, really, really, really, really hate the design trend of confining tiny text into a tiny narrow column down the middle of my browser. It's an awful stylistic decision, and this is the petty hill I'm willing to die on. It's so bad that I really can't take a site seriously that does it.
Now, someone's going to come out of the woodwork to remind me, "Well, ackshually, research suggests that it's easier to read text that's constrained by blah blah blah blah" I don't care. It sucks. It's always sucked. It will forever suck. I have a nice 27" monitor, and I want to use the whole thing. I don't want to have to hit ctrl-] ten times just to have text that is readable and spans my monitor.
> If you have a feed reader system there is no need to subscribe [via YouTube] in the first place. You’ve obviated that system.
This approach works, and it's a great way to subscribe to public channels without a YouTube account. The main reason I'm not doing it is that I want to subscribe via YouTube.
> It would be great if it was more discoverable
Oh, hopefully there's a browser extension that detects feeds on a page and lights up and provides a menu. Shame that the YouTube mobile app isn't similarly extensible.
I wrote a program that exposes my YouTube subscriptions as a news feed. It's seriously a relief to finally have this stuff show up in [my feed reader](https://www.newsblur.com).
It's not a "Show HN" because there's no way for others to try it (yet).
Some things I'm curious about:
* Is anyone else here doing something similar?
* Would anyone like to use this service?
* Are there other YouTube-meets-news-feed features that you'd love to see?
I did maintain a list of all my subscriptions in FreshRSS for a while. I started doing this after I noticed Youtube was withholding videos from the subscription page and then showing them too me as something I might like on the home page. This really irritated me it broke my process for finding where I last was.
The problem was while I migrated in a big chunk using some jank javascript this wasn't a process that kept in sync and gradually over time it got out of date and the initial problem was fixed and I removed them all. I haven't yet seen an open source self hosted solution for this for getting the list and providing an Atom feed but its definitely something I want.
I'm encouraged by what you've shared here.
My impression of this topic is that it attracts a lot of cranks: "The secrets of the universe have been revealed, and you are all blind!"
Sounds like you're open to feedback and the possibility of being wrong.
My main concern is that it will be hard for you to find people who could give you excellent feedback on this,
because they probably have epistemic shields up around this topic due to all the surrounding crankery.
I don't know how to solve this, but I wish you luck.
I was pretty confused about why this even exists (it's weird to read a bunch of paragraphs that are semantically valid but don't seem to convey information), and then I read the [Privacy Notice](https://www.mozilla.org/en-US/privacy/firefox). Looks like there's a decent amount of surveillance / ad tech built in to Firefox.
From "How is your data used?":
> Firefox also shows its own search suggestions based on information stored on your local device (including recent search terms, open tabs, and previously visited URLs). These suggestions may include sponsored suggestions from Mozilla’s partners
> Mozilla's partners receive de-identified information about interactions with the suggestions they've served.
> Depending on your location, Mozilla derives the high level category (e.g., travel, shopping) of your search from keywords in that query, in order to understand the types and number of searches being made. We utilize privacy preserving technologies such that Mozilla only learns that someone, somewhere, performed a search relating to a particular category, without knowing who.
> Mozilla may also receive location-related keywords from your search (such as when you search for “Boston”) and share this with our partners to provide recommended and sponsored content. Where this occurs, Mozilla cannot associate the keyword search with an individual user once the search suggestion has been served and partners are never able to associate search suggestions with an individual user.
> You may be able to opt into an enhanced search experience, which will result in Mozilla processing additional personal data, including your technical data, location and search data. Some of that information may be shared with our partners on a de-identified and/or aggregated basis.
> We use technical data, language preference, and location to serve content and advertising on the Firefox New Tab page in the correct format (i.e. for mobile vs desktop), language, and relevant location. Mozilla collects technical and interaction data, such as the position, size, views and clicks on New Tab content or ads, to understand how people are interacting with our content and to personalize future content, including sponsored content. This data may be shared with our advertising partners on a de-identified or aggregated basis.
> In some instances, when ads are enabled on New Tab, additional browsing data may also be processed locally on your device to measure the effectiveness of those ads; such data will only be shared with Mozilla and/or our advertising partners via our privacy-preserving technologies on an aggregated and/or de-identified basis.
> When you allow us to do so, Firefox sends Mozilla data about the website domain or specific advertising campaign (if any) that referred you to our download page to help us understand and improve our marketing efforts. Firefox also shares information with our marketing partners to measure and improve these campaigns; what information is specifically shared varies (depending on how you discovered Firefox and your operating system) but generally includes how you were referred to our download page and whether you actively use Firefox. Where Firefox is pre-installed on your device, technical and interaction data (your device type and whether Firefox is used) will be sent to our marketing partners, and shared with Mozilla.
One way I treat LLMs is as a "semantic search engine". I find that LLMs get
too many things wrong when I'm being specific, but they're pretty good at
pointing me in a general direction.
For example, I started learning about OS development and wanted to use Rust. I
used ChatGPT to generate a basic Rust UEFI project with some simple
bootloading code. It was broken, but it now gave me a foothold and I was able
to use other resources (e.g. OSDev wiki) to learn how to fix the broken bits.
* Avoiding reading the entire manual
It feels like a lot of software documentation isn't actually written for real
readers; instead being a somewhat arbitrary listing of a program's features.
When programs have this style of documentation, the worst case for figuring
out how to do a simple thing is reading the entire manual. (There are better
ways to write documentation, see e.g. <https://diataxis.fr/>)
One example is [gnuplot](http://www.gnuplot.info/). I wanted to learn how to
plot from the command line. I could have pieced together how to do it by
zipping around the
[gnuplot manual](http://www.gnuplot.info/docs_5.4/Gnuplot_5_4.pdf) and building
something up piecewise, but it was faster to instruct Claude directly. Once
Claude showed me how to do a particular thing (e.g. draw a scatter plot with
dots intstead of crosses) I then used the manual to find other similar
options.
* Learning a large codebase / API
Similar to the previous point. If I ask Claude to write a simple program using
a complex publicly-available API, it will probably write a broken program, but
it won't be *completely* bogus because it will be in the right "genre". It
will probably use some real modules, datatypes and functions in a realistic
way. These are often good leads for which code/documentation I should read.
I used this approach to write some programs that use the
[GHC API](https://hackage.haskell.org/package/ghc). There are hundreds of
modules, and when I asked Claude how to do something with the GHC API it wrote
relevant (if incorrect) code, which helped me teach myself.
* Cross-language poetry translation
My partner is Chinese and sometimes we talk about Chinese poetry. I'm not very
fluent in Chinese so it's hard for me the grasp the beauty in these poems.
Unfortunately literal English translations aren't very good. We've had some
success with asking LLMs to translate Chinese poems in the style of various
famous English poets. The translation is generally semantically correct, while
having a more pleasing use of the English language than a direct translation.
> I just asked ChatGPT what it knows about me, and apparently I am known for my work in the field of AI. [..] people taking this stuff at face value for anything important is terrifying.
Yeah, I'd be more comfortable if it could say "I don't know" instead of generating bullshit.
Simon Willison's article Think of language models like ChatGPT as a “calculator for words”[1] helped me think about this topic better:
> If you ask an LLM a question, it will answer it—no matter what the question! Using them as an alternative to a search engine such as Google is one of the most obvious applications—and for a lot of queries this works just fine. It’s also going to quickly get you into trouble.
All of this is why I only every use LLMs for things that I can verify.
(I'm not the author) The easiest way to charge for this kind of software is to make it SaaS, and I think that's pretty gross, especially for a CLI tool.
> I'm skeptical and I may just release it all as OSS
It doesn't have to be one or the other: you could sell the software under libre[1] terms, for example.
[1]: https://en.wikipedia.org/wiki/Free_software