HN user
The title is a reference back to "Is Claude Code a Compiler?" by the same author, linked from the first sentence: https://commaok.xyz/ai/is-claude-a-compiler/
And then claim is attributed right there:
The hands-down highlight was the talk by Erik Schluntz: Vibe coding in prod.
Among other things, he drew an analogy between LLMs and compilers.
You won't make one at home, but cloud chambers[^1] reveal individual alpha particle tracks.
There's one in the Musée des Arts et Métiers in Paris — blew my mind!
[^1]: https://en.wikipedia.org/wiki/Cloud_chamber
Edit: turns out people make these at home all the time. Sick!
Large gems can be broken up and recut for sale. Destroys value (certainly the cultural value) but renders them salable.
His recent "Secret Life of Components" series covers lots of fascinating electromechanical tinkering, including many of the mechanisms in the Novelty Automation arcade machines: https://www.youtube.com/watch?v=6JAgXz6xO0s&list=PLtaR0lZhSy...
Basically linkblogging. I'm not sure this needs to be a separate feed; JSON Feed has a dedicated `external_url` field:
If `url` [optional] links to where you’re talking about a thing, then `external_url` links to the thing you’re talking about.
I'd be shocked if Atom/RSS didn't have equivalents.
This kind of "repost"-ish behavior may just be obscured in the tools people use to construct feeds, so they remain obscure features of the standards. The designers had syndication in mind, very much like what you're describing — ingesting feeds, reprocessing/mixing/extending them, and exposing the result as another feed.
After about a decade of experimentation (NetNewsWire, Feedbin, Miniflux...), I'm self-hosting a feed reader I wrote myself, running on a free hosted LibSQL db.
- NetNewsWire was slick, but wouldn't work on my phone.
- Feedbin was excellent, but eventually I decided to do some subscription cost-cutting.
- Miniflux worked fine, but 1) I found it a pain to set up with remote hosted Postgres and 2) it burned through the Neon free-tier usage limits in a couple days.
So I built one myself and run it on a Raspberry Pi home server.
Made a great little weekend project. The feed standards are known quantities, so a little AI assistance with boilerplate goes a long way.
Deciding you need a new feature and just adding it is refreshing — e.g. I wanted a "read it later" feature like Feedbin's (something missing from Miniflux), and now I have it.
An RTO mandate is also an excellent thing for a CEO to show investors they are doing, if they are not making money and lack better ideas.
I think of Jeffrey Pfeffer's "social contagion" arguments a lot — first with regards to layoffs[^1], but increasingly also to RTO policies and tracked AI use.
It seems very unlikely execs (esp. in small organizations) are taking the time to read and seriously evaluate research about RTO or AI and productivity. (Frankly, it seems much less likely than them doing serious modeling about layoffs.) At some point, the "contagion" becomes a matter of "best practices" — not just a way to show investors what you're doing, but part of the normal behavior shareholders expect.
Bleak if true!
[^1]: https://news.stanford.edu/stories/2022/12/explains-recent-te...
A practical guide for integrating a custom linter into your Go project, a [light] screed against golangci-lint, and a defense of a simpler Go metalinter I’m experimenting with: glint.
It's a good post, super clearly explained. I have to admit it makes me a little nervous about the future of Go packages — of course interfaces aren't intrinsically bad, but omitting them in early versions of the language discouraged reflection and walled out a good deal of this complexity (in exchange for verbosity).
I wind up feeling grateful for simple interfaces, especially the ones disseminated in the standard library.
Strongly statically typed languages let you safely refactor in ways that would otherwise be risky. This is my trick for one very common example: lifting an interface out of a concrete type.
I don't understand how this pertains to my comment. Nobody's arguing whether listeners will, as a matter of preference, avoid Google/Apple/Spotify podcast clients. Nobody makes a claim about how many people know of RSS.
It's up to the podcast producers to decide for/against distributor-exclusivity, determining whether OC has to use YouTube Music.
The vast majority of folks don't know that they have a choice, let alone that they should find/choose an app.
That's... fine. What matters is whether they do have a choice if they go looking for one — again, that's up to podcasters, not up to Google.
This seems like an upstream concern with distribution, no?
Nothing in this announcement indicates YouTube Music will attempt to control distribution. YouTube might give exclusive content a try (a la Spotify), but — to your point — that won't be a "podcast" in the strict client-agnostic-syndication sense.
As long as podcast creators are publishing mp3s to RSS/Atom feeds, YouTube Music is just another podcast client (like Google Podcasts was) which you're free to avoid.
Okay... here's an attempt at optimism.
In one respect this consolidation makes more sense than the Play Music → YouTube Music switch: many podcasts cross-post episodes as YouTube videos and audio-only feeds. If YouTube collates those feeds and lets users select between modes while maintaining a unified "already watched/listened" list... hey, that reflects how podcasts distribute/monetize in practice.
The YouTube Music app was a dismal replacement for Google Music, and never regained features for managing one's own library of uploads. Of course, that's not a problem for podcasts; episodes are already syndicated.
All that said, this still bodes poorly.
Link isn't loading for me––also on archive.org: https://archive.org/details/TheWozWonderbook
Deletion isn't the only option for escalation; Twitter could assemble a risk factor for a given account and opt to present captchas, for example, rather than automatically deleting. Scale captcha frequency with the risk factor.
Captchas are certainly not foolproof, but they up the difficulty of running a huge number of phony accounts while having a minimal impact on normal Twitter usage.
Also, this reminds me of a Clara Labs (YC S14) project to hide unique IDs in unicode text: https://blog.claralabs.com/invisible-ink-a-simple-but-faulty...
Interesting. Did some quick poking around.
Looks like the system adds a number of non-printing unicode characters into the string. In this thread title, for example, the textmark is stored in the phrase "message has a hidden message inside." (Thanks, textdiff!)
Specifically, it adds non-printing unicode characters repeatedly. I'd guess these form a unique series, which is mapped to a record in the textmark database.
If I wanted to automatically remove the textmark from a text, in order to steal your content, I would write a script to automatically remove any non-printing unicode characters that appear in it. This StackOverflow answer includes a script achieving precisely that task: https://stackoverflow.com/a/11598864/6226586
Yep, we agree entirely! And I think most other technical debaters would, too.
I don't talk to my friends at high speeds, my favorite speakers speak slowly, and I certainly wouldn't talk like a debater when I'm trying to be my most persistent.
The kind of debate McCordick is criticizing is, in my experience, indeed fringe.
There are more mainstream offerings: less technical debate formats such as Public Forum, Parliamentary, and Congress; Mock Trial; Model United Nations; Junior State of America; the Center for Civic Education's We the People competition; popular organized political debates; other oratory activities outside of civics.
To be fair, I've had the good luck to be involved in High Schools with the resources to offer a selection of these programs, which isn't always the case. I wish speech––the kind of slow, articulate expository speech that wins over non-debaters––was a bigger part of core curricula (in civics, but also in sciences).
How do you feel about grindcore?
Certainly it has musical and aesthetic merit, historical importance, and it demands talented composers and performers. Additionally, there is lots of other music for folks who don't enjoy it.
If popular accessibility is all that matters, we could pare down music to the major popular genres. I don't think marches would make the cut.
Fast debate is like grindcore; there's room in civics for specialization, even specialization that isn't aesthetically obvious.
It's hard for most to pick out the logic or musicality from the noise of debate or grindcore, respectively. I would hate to live in a world where grindcore is the only music out there. But I'm glad I have the option of putting on some John Zorn when I'm in the mood.
Competitive cubing is pretty cool, but it's an algorithm/dexterity problem that requires none of the creativity that debate demands.
I disagree with a handful of this article's arguments, and I'm happy to see a number of debaters chiming in on this thread with similar concerns.
[I did parliamentary and policy debate in high school; now I'm a High School parliamentary debate coach]
1. @elefanten correctly points out that arguments are still evaluated at speed. This is my experience; the author characterizes it as "bludgeoning," but good debaters increase their speed to add evidence to a small number of strong arguments. Bludgeoning can happen, but it's uncommon. If McCordick's characterization of Policy were accurate, we could expect the fastest debaters to predictably win all their tournaments. This isn't the case. Speed is a rough correlate to ability inasmuch as it's an early product of commitment, but there are diminishing returns. By a third year in competition a high school debater isn't focusing on speaking faster at all. The real competitive edge is in argumentation.
2. McCordick's comparison between "corrosion" in debate and "corrosion" in American national politics is extremely hasty. National politics have not become more inaccessible because they're technocratic and opaque. The 2016 presidential debates were not run at 300+ wpm, nor did they involve detailed policy advocacies. Debate is characterized by technical elitism, national politics by right- and left-populism. How is debate jargon remotely like "the kind of language currently poisoning our public sphere?"
3. McCordick speculates that debate produces students who don't care about truth and public speech, but there's no real evidence of this. Even if there's validity to his stylistic criticisms, it's unlikely debaters are on the whole less civically engaged or politically literate or eloquent than their peers. My experience, at least, is that they're dramatically ahead.
4. He ignores all the skills besides rhetoric. Research and acadmic/technical literacy are critical to prepared debate, as they are to college academics. Debate teaches a number of mental models that I've personally found extremely useful––e.g. for grokking multiple opportunity costs in decisionmaking. It teaches work ethic and self-motivation, both skills admired in athletics. Also topic knowledge at a very accelerated level for high school students.
5. McCordick gives the impression that debate is homogeneously inaccessible, which is inaccurate for a few reasons. A) High-level competition makes up the vast minority of High School debate; the vast majority of debaters are active on nationally uncompetitive local circuits where traditionalist judge expectations keep debate slow and untechnical. B) Other commenters correctly point out the rise of new performative styles that preference debate on relatable terms, which have been highly competitive for years now. C) There are plenty of debate formats that aren't hypertechnical––McCordick even competed in two of them: Public Forum and Worlds. Parliamentary debate often plays a similar role where it's offered.
6. I agree with @ashark that debate's idiosyncracies are a product of Goodhart's Law. Competition is an essential incentive for the workload, in the same way that international science fairs incentivize research (the only other high school competition in which I've seen a similar work ethic). @ashark suggests in a different comment that the only way to counteract specialization is to allow subjectivity in judging. This would just substitute judge-appeal as the metric for competitive success, which is hardly a better proxy for truth-seeking and communication skills given judges' pre-existing political biases on a topic.
7. I understand frustrations with debate's apparent inaccessibility to laypeople. It's a frustration I experienced myself when I first started. Why, though, is there so much frustration over specialization in debate? Why not similar frustration over other specialized activities? Debate definitely doesn't bring politics to the polis, but it needn't. Not all politics is about eloquence; it's also about cost-benefit analysis, philosophical dilemmas, wonky details. It isn't necessary that everyone is a technical expert, and it's naive to believe that policy doesn't demand expertise. I find grindcore offputting, but that doesn't mean grindcore is "corroding" or "perverted."
Interested in developing such expertise? Competitive debate is one of many ways in which to develop it. It certainly isn't perfect, but it teaches skills that don't exist anywhere else in high school curricula.
Does it erode general advocacy skills and interest in truth? Certainly not for debaters who are good at it. If anything, it builds them. Skills are not mutually exclusive.
On a petty note, the article credits Jack with winning "the high school debating national championship"––America Magazine should be more precise. Regis won first in World Schools Debate, not one of the major formats.
McCordick argues that debate is becoming an inaccessible rhetorical monoculture; you point out a trend towards rhetorical diversification, affect theory, and personalization.
This is arguably competitive debate's biggest culture clash (though debaters from neither camp could be successful without proficiency in the other)––even if you believe them to be corrosive, it's misguided to construe them as parts of the same trend.
Additionally, it's interesting that you select 2014 CEDA finals as an example, which was not only criticized but also met with fairly vicious racism––why not the 2015 NDT, which exemplifies the kind of debate McCordick criticizes? Or the 2013 NDT, which shows the two styles in interaction?
This sort metrics-buying isn't unique to books:
The RIAA changed its rules so Jay-Z's 'Magna Carta Holy Grail' would go platinum instantly after Samsung bought one million copies to give away to its users: http://pitchfork.com/news/51392-riaa-changes-gold-and-platin...
A similar partnership with Sprint meant '4:44' also went platinum almost immediately, even though it was only distributed on TIDAL: https://www.forbes.com/sites/garysuarez/2017/07/05/jay-z-444...