pseudocomposer

I’d like to echo what others are saying about building stuff you want to exist in the language you want to learn. I learned Flutter/Dart largely by building BeatScratch (https://beatscratch.io), and Rust entirely (and parts of functional React) by building Jonline (https://jonline.io).

But secondly, if you can, get work or your career path to let you build something in a new language. A lot of learning new languages (both programming and human) comes from social reinforcement and riding social waves. Smaller companies are a lot better for this than bigtech.

I learned Ruby at a company I was hired as a Java dev for initially (that had lots of Ruby staff). I learned Go and my first bits of React when a Ruby company I worked for was acquired by a Go company. My current company hired me to work on an Elm frontend without me having written a line of Elm in my life.

I’m my second year into Kagi and loving it. I actually just upgraded to get Kagi Assistant (basically, cloud access to every LLM out there). But the search alone is worth every penny, and it’s built/operated fully in-house as far as I know.

https://kagi.com

Have you considered PRK/LASIK/SMILE?

There's two sides to this: using AI/ML/LLMs to augment your own development ability, and using AI/ML/LLMs within apps you build.

For the former side, Copilot-type implementations are pretty intuitively useful when used this way. I find it most useful as an autocomplete, but the chatbot functionality can also be a nice, slightly-better alternative to "talk to a duck when you're stuck." That said, I'll focus on the latter side (using AI/ML in your actual work) from here.

Generalized AI/ML/LLMs are really just a "black box API" like any of the others in our toolbelt already, be they Postgres, Redis, SSE, CUDA, Rails, hell, even things like the filesystem and C atop assembly. We don't need to know all the inner workings of these things, just enough to see how to use the abstraction. You probably take when to use a lot of these things for granted at this point, but the reason we use any of these things is that they're good for the specific problem at hand. And LLMs are no different!

What's important to recognize is the types of problems that LLMs are good for, and where to integrate them into your apps. And, well, a pretty obvious class of this is parsing plain text into structured data to be used in your app. This is pretty easy to prompt an LLM to do. OpenAI and WebLLM provide a pretty straightforward common set of APIs in their NPM libraries (and other language bindings are pretty similar). It's far from a "standard," but it's definitely worthwhile to familiarize yourself with how both of these work.

For an example, I've made use of both OpenAI and WebLLM in an "Event AI" offshoot to my social media app [1], parsing social media events from plaintext (like email list content, etc.); feel free to test it and view the (AGPL) source for reference as to how I'm using both those APIs to do this.

For projects where you actually have money to spend on your LLM boxes, you'll probably do this work on the BE rather than the FE as demoed there, but the concepts should transfer pretty straightforwardly.

If you're interested in really understanding the inner workings of LLMs, I don't want to discourage you from that! But it does seem like really getting into that will ultimately mean a career change from full-stack software engineering into data science, just because both have such a broad base of underlying skills we need to have. I'm happy to be wrong about this, though!

[1] Source: https://github.com/JonLatane/jonline/blob/main/frontends/tam... | Implementation: https://jonline.io/event_ai

Reputational damage is pretty easy to limit for companies that can control what news a huge majority of people see.

You can have extensible structure and fields with JSON Schema, gRPC, Cap’n Proto, etc. There’s nothing XML-specific about that.

The only thing XML gives you over any of those formats is unstructured mixing of text and data, which is more a foot-gun than anything. Oh, and of course, being significantly more verbose.

I appreciate the feedback! While I haven't had plans to ship binaries to platform-specific package managers, I'll certainly consider it. (I do have plans to publish it to Cargo, but have yet to get to that.) PRs for this would be very welcome, as I have no experience shipping things to apt/yum/etc. :)

In the meantime, it is available pre-packaged as a Docker image for both x86-64 and ARM64: https://hub.docker.com/r/jonlatane/jonline

Dare!!!! Yay!

You might be interested in my AGPL federated social network, Jonline, which I've built specifically for this purpose. Source is at https://github.com/JonLatane/jonline, and I have a few instances up at https://jonline.io ("flagship"/demo server), https://bullcity.social (Durham, NC community), and https://oakcity.social (Raleigh, NC community).

It's a pure Rust server so it's extremely lightweight to self-host compared to things like Mastodon, plus it supports events (with iCal/Google Calendar export). I'm also happy to host an instance on my DigitalOcean Kubernetes cluster for anyone interested and integrate it into my CI/CD pipeline, which is an option on my GitHub sponsors page (and I'm working on a load balancer which should eventually bring the cost of this down as well).

Can it integrate with mail, messages, calendar, reminders, notes, journal, etc. the same way Siri can? Otherwise it seems like a competitor to Alexa or Google Assistant at best.

So that’s why some devs can somehow still manage to make flexbox layouts difficult :D

Would this not be easily solved with a one-time 50-90% wealth tax when they exfiltrate their money from the place where others earned it for them, rather than the ~1% annual wealth tax they want to leave over?

> It feels like the current system is the Imperial system and somewhere ought to be the SI metric system.

Any reason for that beyond “feeling?” And are you very deeply familiar with how the current system resolves differences between the single chromatic and 12 diatonic scales?

Full disclosure here: I’m a pianist who can generally sight read nearly anything (that’s well-edited) put in front of me, to the point of having music directed opera and theater for a living for a while, rather than using my CS and math degrees immediately after earning them from UNC.

I’ve also created an app, BeatScratch (https://beatscratch.io) that, among other things, attempts to resolve some of the challenges of editing (which I’d argue, again, really involves understanding how musical notation maps chromatic and diatonic scales between each other - ie “picking the best sequence of flats and sharps automatically to maximize readability”).

To me, a lot of these “musical notation is outdated” arguments seem the same as “there will soon be no need for programmers; AI can write all our code,” or those arguments insisting some “visual programming” paradigm should replace languages like Rust and Python.

That is to say… they’re arguments made by people who don’t really understand why the existing systems work the way they do and have done so successfully for decades (or centuries).

My AGPL app Jonline does what your customer wants. Every conversation has “discussion” (most recent at top) or “chat” mode, and you can star any reply at any depth to have it accessible from a menu. (TODO: push notifications, but when they’re done, starred posts/replies will also support them.)

Here’s a sample convo where you can try out the functionality: https://jonline.io/post/4zHQSj

I’m actually a UNC grad so it breaks my heart too :-(

It’s really because (1) it costs $14/mo to add an instance currently (though my aforementioned cost saving measure can bring it down to $1/mo), and (2) I’m not sure whether to register https://tarheel.social, https://chapelboro.social, or what really. Each has specific connotations and it’s a bit weird with a college town (while Raleigh and Durham are more diverse cities).

If you’d be interested in partnering and standing an instance up, it’s AGPLed for exactly this reason :-) Feel free to DM me on Insta (@jons_difficult_truths), contact me via my GitHub above, or just reply here with your preferred medium if you’d be interested in talking about it over coffee/drinks!

I’ve actually built and AGPLed my side gig, which correlates pretty closely to what you described here:

https://github.com/JonLatane/jonline

Personal/Demo server (links to/includes data from the other two servers automatically): https://jonline.io

Servers for Durham, NC and Raleigh, NC (link to each other only): https://bullcity.social and https://oakcity.social

But um… how make money with it? I’m paying $60/mo for these 3 servers (and can bring it down to $40), and really stuck on how to monetize and get people using it.

How is money spent on childcare different from disposable income? Differentiating them artificially would fail to account for people who want children, but don’t, because they don’t feel they have the income. Which is a lot of the middle class at this point.

Who says IP was ever right in the first place? A bunch of wealthy, powerful people?

I’d work on my app BeatScratch full time. (https://beatscratch.io)

I think it should be “authentication” and “permissions.”

There’s pretty much no word in the English language to describe login/account creation/etc than “authentication.” The word “login” is a poor substitute. There are no good synonyms for “authentication” that encompass all its applications in computer systems.

Meanwhile, there are already lots of synonyms used for “permissions.” Given the abundance of these, and the lack of synonyms for “authentication,” choosing “authorization” to describe permissions is, frankly, an asinine decision. It adds unnecessary cognitive overhead for everyone.

(That’s not to say there’s no place for, say, Unauthorized responses, etc. Just that we should be calling the topic “permissions” or really anything other than “authorization.”)

Unions have historically solved this problem. Non-union accreditation programs don’t seem to have the same effect, though in some industries they do have a positive effect on salary.

Or the city could just invoke eminent domain, buy it at a very low price if the owner can’t demonstrate it having a high market price, restore the building and democratically choose an equitable process to rent it or sell it, hopefully favoring existing citizens.

But of course, we’d have to be a lot more aggressive about invoking eminent domain for this to be effective. To that end, it would make sense to also establish policies ensuring it’s easy to invoke it fairly and transparently - ie, easily allowing other parties to bid against the city for higher prices in cases where eminent domain is to be used this way (stipulating that those parties also take measures to restore the building, etc).

Between OnlyFans and a cultural shift in masculinity that makes paying for sexual attention more of an "ick" for many, this doesn't seem like a very good economic indicator, especially today.

Do you and your friends in London all drive to work? Or do you walk/metro?

The lifestyle shift the grandparent poster was referring to is typical of suburban and rural areas, but not really urban areas. I suspect people in more rural parts of the UK are very different from your London cohort.

Seems like they need to organize/unionize then, no? Their failure to do so isn’t the “shelf stackers’” fault. If anything, this gives them leverage to get paid more. But they have to use it, because capitalists will never pay someone fairly if they don’t fight for it.

I’d recommend Flutter Desktop if you/your team are already comfortable with React development. Even if you don’t know Dart, it’s easy to pick up, has mature tooling, and Flutter’s UI API is very similar to React and will be comfortable for a lot of devs very quickly.

It also leaves open the possibility of eventually supporting your app on web (with, say, a backend or even WASM doing the FFMPEG stuff). Few desktop UI lots offer this.

What was pro-China about the comment you responded to exactly? It was anti-multinational-corporation, for sure. But it literally put China in the same category as those corporations.

I’m not that your post is suspicious; I’m saying outright that you are either a CIA/corporate spook or being manipulated by them. Any criticism of the companies that have dismantled American unions, healthcare, education, childcare, elder care, and more, for the sake of shareholder profits, seems to be unjustifiably labeled “pro-China” or “terrorist” or some other label by some anonymous person who hasn’t really contributed anything to the community otherwise.

Until you start comparing the bandwidth/speed of that RAM and SSD, at least…

I see a lot of “7M” chords in generated outputs, which isn’t a type of chord I’m familiar with. Is this meant to be “M7” (major 7)?

Here's one way to put it: Suppose people in the Northern US had, since the end of WWII, been using a mix of economic and military tactics to force Canadians near the border out of their homes, declaring the new territory part of a New England-centric ethnostate where only people of New England descent were welcome, to the point where the only remaining Canadian areas were completely surrounded by American settlements with militarized borders. Suppose between last January-October of last year alone, Canada lost 100-500 or so children directly due to military action and thousands more civilians due to inadequate healthcare and nutrition resulting from the military blockades.

And then, suppose that a militant group of, at most, 100-1000 of those Canadians, out of 5M or so total Canadians, did a parallel to Oct. 7.

By the time we'd killed 20,000 unrelated civilians, I'd be pretty upset. I would be upset at seeing rhetoric on TV saying that Canadians were "animals" or "grass that needed to be mowed." As a Southern American, I'd be upset that my tax dollars were funding this.

It's also worth noting that this really parallels 9/11 quite well. I was slower to be upset at our tactics in Afghanistan/Iraq then, but having seen it all now, Israel's attack is clearly the same fruitless-for-workers, military-industrial-complex-feeding meat grinder as it's evidently always been.

And to be more "objective" about that comparison: 9/11 killed more Americans than Oct. 7 killed Israelis. As of 2023, Afghan civilian deaths from our war on Al Quaeda (the entire period from 2001-2023) number [1] "more than 70,000." In half a year alone, 30-35,000 Palestinians have been killed.

But I'd say that measuring human lives against one another is kinda gross, focused on hindsight, and ineffective. To do the best thing we can in the present, we need to make the choices that maximally preserve human life. And selling/giving them weapons just ain't it.

[1] https://watson.brown.edu/costsofwar/costs/human/civilians/af...