Apologies, no offence to New Yorkers (or the New Yorker) meant.
If you find my handle, "maga", interesting..., I'll have you know that it's been around long before it was appropriated by some movements in US and has nothing to do with them ;)
> It’s unclear whether the suggestions and techniques mentioned in the article came from personal experience or have otherwise been verified or experimented with with a real team and a real project.
Since it's not a New Yorker article, I was hoping to spare the audience a long personal life story and deliver a somewhat succinct list of suggestions that others might find useful.
However, the question is valid, and yes, this is the result of personal experience of following and incorporating AI tools into my own development over the last couple of years, as well as watching my colleagues of various experience levels (in a team of 10 engineers) do the same. These are the practices that we collected, adopted, and trying to codify and develop further.
Code splitting is supported just fine for ESM bundles[1] (and there is little reason to use other formats nowadays), CSS handling works ok as well. I have been using ESBuild since forever and the only instability that I experienced was around sourcemap generation early on.
That's why we introduced the strict mode ("use strict" pragma)[1]. Among other things, it prevents from accidentally declaring a global variable this way, throwing a ReferenceError.
Actually, at this point, with the prolifiration of strict mode and linters, I'd say that these old gotchas mostly belong to quizzes and spec discussions, since personally I have not seen new code written this way even in vanilla JS for years by now.
Great to see dprint on the front page! While here it's speeding up prettier with caching/wasm, it can replace prettier altogether and be even faster (for example, with `dprint-plugin-typescript` for TS/JS dev). So, if you are doing web dev and not yet ready to switch to Deno, you can at least speed up formatting by using `dprint`.
David, I already owe you guys more beer than I can buy for all the great work on the Deno toolchain, but if we could get this issue[1] with HTML template tags going, it would be easier for me to convince the Lit crowd to chip in. Tanks again for the great work!
Leo, your comments save me a lot of typing on threads like this, and since I recently wrote[1] what beeandapenguin wrote above almost to a point (sans security), I feel obliged to expand a bit.
You are right about incompatibility being a major issue; Deno recognizes that as well, hence, they are working on a compatibility mode that allows using Node specific libraries in Deno[2].
> migrating to Deno is, charitably, likely a monumental task that nobody would ever undertake, even considering the upsides.
This is, of course, contingent on the architecture used: for code tightly coupled to frameworks/runtimes it is indeed a monumental task. I have two small to mid size SaaS apps happily running on Node.js, but I'm looking forward to replacing it with Deno solely for the streamlined DX. The apps follow DDD architecture, thus, framework specific stuff is decoupled into a service/adapter and changing it is a day's work. The major technical road-block for now is indeed incompatibility of third-party libraries/SDKs written for Node.js (google sdk, mongdbo driver, etc.).
Yes, TS code is compiled to run in browsers. You can compile and bundle using `deno bundle` if your project is developed in Deno. Since I'm using Deno for library development, I use the `dnt`[1] tool to compile them into ESM modules later consumed by esbuild in my Node-based projects/toolchains.
I have moved all my library development to Deno and described the experience in a blog post[1].
To summarize, the two killer features for me are Web APIs and the built-in toolchain.
I understand the skepticism other Node devs have when faced with Deno. I have been using Node since 2010, and there is hardly anything I cannot do with it (or didn't do at one point), yet Deno makes a lot of these things much simpler. My last straw was trying to run Jest tests on TS code compiled into a Node.js ESM module. I believe I spent more time figuring out workarounds for a myriad of issues in Jest+ Node+TSC chain, than it took me to switch to Deno where all of it came out of the box.
By any chance, are you aware of any good surveys comparing the open models to the cloud providers for other NLP tasks like speech recognition or POS tagging?
Sounds almost too good to be true, but here is hoping for the best.
> Opus audio decoder (enabled)
Considering the track record of codec support, I bet it only works when packaged as CAF, singed by QuickTime, and played during the daytime to a limited audience... I joke, but the amount of gotchas I encounter while working with WebAudio API is mindboggling and reminiscent of good ol' days of IE6. Would you guess that despite all three browsers declaring support for audio recording (MediaSource API) and all three supporting opus on paper, there is no way to produce the same opus file in all three? Safari only supports playing opus in <audio> tag, no recording; Chrome doesn't support ogg, and while both Chrome and FF allow you to record opus in webm container, the resulting file is not seekable... There are of course third-party solutions, but it's just sad to watch in the century of the fruitbat.
> You’re still missing much the point of the article: WASM isn’t for the browser only. WASI, for example, is a full environment you can run stuff in without any such interoperability cost.
Of course, but we are not arguing about the point of the article, I'm contesting the exact claim that "WebAssembly is almost always faster than JavaScript". Admittedly, not because it's such a big mistake on the author's part, but because I'm a bit irate with the usual flow of discussions about WebAssembly here on HN that take this claim for granted and regurgitate it incessantly. This is not the case for the mentioned reasons: data copying overhead and JS itself being compiled by JIT resulting in pretty performant code if well-optimized. Also, the comparison to JS only makes sense when we are talking about running both in the JS engine, not against a dedicated WASM runtime, otherwise it would be tantamount to arguing that Java is not faster than JS.
You do realize that I didn't put up those benchmarks and wrote the article way back in 2017 just so that I can pick on a passage in this article? I was (and still am to a degree) excited about WASM, actually ported a significant chunk of my business logic from JS to C++ only to discover that the whole thing offered only ~20% performance boost, which didn't merit maintaining a whole separate toolchain. I'm not questioning the exciting results that Figma and such achieve, I'm saying that it's not "almost always" a case, and it's disingenuous to present it as such.
And how is the data going to reach the "unshackled" WASM in the browser, for example? I'm not aware of any way for the web page to interact with WASM other than through the JS land with the said overhead. The article explicitly mentions usage inside the web page and presents WASM as a performant alternative.
> The trick then is a sliding scope context window: with a high-level design doc in context, iterate to produce an architecture document.
Absolutely, I will be stealing this!
> It's a very exciting time, it feels like having worked by hand for decades, only to now have access to power tools and heavy machinery.
Very well put, captures my feeling precisely.