It’s underremarked upon that stablecoins accomplish basically all of the original design goals for Bitcoin qua a transactional mechanism (instant free value transfer anywhere), but that nobody cares about this because nobody ever cared about that part of the pitch.

FAQ: “How do I get better at writing?”
Me: Write a million words.
Follow up: “Hah but seriously.”
Me: Start with 20,000 words. Everyone gets to that 50 times in their first million.
Follow up: “No seriously.”
Me: Nobody expects 10 Quick Tips To Play Violin At Carnegie Hall.

The Internet is so much better at teaching writing than the traditional educational system principally because it removes an artificial and limiting constraint on output imposed by ability of gatekeepers’ limited time to evaluate output.

(Interestingly in fields where we actually care about performance we’re good at encouraging students to practice when no one is watching but for writing we basically assume all the value is created by the teacher-minute of attention. Teacher-minutes are scarce and rationed.)

If you ever want to be deeply, deeply disillusioned about the state of the world, compare what we do where we care about output quality and what we do where we don’t, and which domains sort into which column by revealed preference.

My friends @robwalling and @einarvollset just launched TinySeed, an accelerator for software companies where a successful outcome is a healthy, sustainable business rather than attempting to ride the rocketship trajectory. https://t.co/LSHqZNcVid

I have some thoughts:

As somebody who bootstrapped ~4 companies, I feel like I had to make some clearly suboptimal decisions early in them for lack of what is, in hindsight, not all that much money. But there's a huge gap in the product space for investment options.

It's weird: you can get $25k from Amex trivially, and angels are very willing to write a check for that much, but you have to make representations about your goals/ambitions/market/etc which don't really apply to everyone.

And so you see the traditional angel/VC ecosystem fund companies where honestly the returns are probably not there, and this is knowable pretty early, but the chase of them will wreck what could have been a perfectly happy business.

(To make the math work for traditional VCs the company has to at least have a market-appropriate shot of $100 million a year. There are a lot more $10 million a year companies than $100 million a year companies. That is *not* a bad terminal outcome for founders/employees.)

This lunch is 500 yen ($4.80) at Sukiya, a Japanese fast food restaurant which belongs to a category with about three big competitors.

I love the aesthetics of this category and they’re under remarked upon.

I think people underestimate QSRs in terms of social utility, but Sukiya et al describe themselves as mission-oriented enterprises. I believe this is largely sincere, and goes back to the 60s and 70s, when the clientele was primarily manual laborers who had migrated to work.

Japan was not a rich nation at the time, and day laborers in particular were both unlikely to be able to cook for themselves and unlikely to have much of a food budget, and so the chains sprung up offering an honest-to-goodness cooked meal delivered in under a minute for cheap.

This heritage continued over the years, even after Japan became a much more wealthy nation, and these chains function as social support and dignity for folks in diminished circumstances.

They also are a wee bit of a cartel, and I appreciate the aesthetics of the cartel:

Back when I was first in Japan, in the mid 2000s, there was an increase in the price of beef.

And the heads of the three chains got together, and decided that the price of the basic beef bowl needed to increase, but given the economic circumstances how could they hold the line.

The payments wars in Japan are heating up and one of the battlegrounds is convenience store coffee.

“Coffee? What does that have to do with payments?” I’m glad you asked.

Convenience stores are low net margin businesses, which sell some high gross margin goods/services but a lot of low ones, and have high fixed costs and a low ticket size. The typical transaction is under 500 yen ($5) and many are about $1.

They need repeat custom.

A few years ago, all of the chains had a good idea for increasing frequency of use: make a minor capital investment in automatic coffee machines. Sell access to them for the price of a cup / ice; customers self-serve with the machine.

The price point is $1 to about $2.

Coffee quickly became one of the most frequently repurchased items at convenience stores, in no small part because it’s the one thing they can sell which is phameceutically habit forming but totally unregulated. (Just telling it like it is.)

But the coffee is not very defensible

The problem, such that it is, is that competing chains are everywhere and *all* of them serve Thoroughly Adequate Coffee at similar prices, so you’re back into the brutal economics of “Who is 3 meters closer to 40 customers at 1 office?”

Enter payment apps.

A thread on HN about bad code in legacy projects both makes me think how little we've learned as a discipline over the years and, honestly, how little credit we give ourselves for some pretty major

Fun going down this list and thinking: "Hmm, plausible at a well-run modern software shop", "Hmm, possible, but requires implausible tradeoffs", "Literally disallowed by languages", and "If you were to attempt doing that our test suite wouldn't let you merge."

I think we as an industry celebrate (not quite the right word) failure too much and don't celebrate success nearly enough. There is no DailyWTF for competent execution, word of which generally stays pretty local to the source while incompetence passes into legend.

Alrighty let me try to thread the needle on being the change I want to see in the world while not giving away anything that will get me in trouble:

Ruby has wonderful developer ergonomics. Typed languages are easier for machines to guarantee the correctness of. We built a type checker for Ruby (and I believe it is slated for OSS release sometime).

c.f.

Java was first released in mid-1995, just 6 months before JS.

By 98, when I went to college, Java was already used for all the first level courses in the CS program.

How did it catch on so quickly (just 3 years) to shift university curriculum, which is usually so slow/behind?

I wonder why Java was so respected and JS was not? Why/how was Java already seen as such a stable and mature language?

Not just Java, other languages (like python) were also released around/near 95, and also seemed to get to "mature" respect status. much quicker.

I bring all this up because my impression of Java in 98 was that it had been around "forever" (as had C++ first developed in 79, but not standardized until 98!), but that's not. true at all. These were all new languages around the same time.

I'm sure the language designs had different merits, and the target applications meant a different perception and reception, but...

it sure seems like the industry (and academics) just decided Java and C++ were the stable mature ones and langs like JS were toys.

BTW, before you assume JS was nascent/immature b/c the web wasn't much of a thing yet, remember that JS was actually released as LiveScript on the server before it was released in a browser. So "server-side JS" was always part of the story, not just Node 15 years later in 2009.

I got a Toshiba T1200, which has a bad screen, and a dead CMOS battery, and I thought the hard drive was broken too...

but I rebooted it and heard a KTHUNK and a really loud spin-up noise and holy shit, the hard drive works!

and it still has files on it. the newest date is January 1994.

It's a 21mb hard drive, with ~37k of bad sectors.

This fucker is never going to spin up again, so I better go find some serial cables and copy shit off now.

The backlight is fucked so I'm shining a flashlight on it.
without it, it looks like this.

There are only 3 real challenges to free speech at the tech level a) the need for CDNs at scale b) distortions of net neutrality c) nation-state firewalls.

All other claims are about social governance on private sites, where applying free speech doctrine is a stretch at best

Baseline free speech has expanded modestly since 1776 or whatever your reference point is.

About 20% beyond core case of government-critical speech in formally public spaces and media I'd say.

You can (for example) have more kinds of conversations in a bar without getting punched or kicked out by other patrons or management. You can teach more things in colleges than 100 years ago.

That's a genuine de facto increase in practical, informal free speech/expression.

But online free "speech" is, as @reneed has pointed out, not an expansion of free speech at all, but an expansion of reach. It is like the secondary freedom of the press to print as many copies as they can afford, not primary freedom to say what they want

If a newspaper wants to print more copies it has buy more printing equipment and front more upfront costs. Reach for a paper publication is not "free", but a function of desired circulation, which in turn is a function of demand prediction etc. But that logic is obscured online.

@zkat__ I'm in the position that I actually find npm / yarn the best ecosystem. Whenever I use something else I always end up stubbing my toe into something thats missing / feels wrong.

Ex. Cargo seems to neither have a concept of devDependencies nor peerDependencies.

I also can't understand why it wouldn't have an "add" command to add a new dependency. And I'm no fan of Toml, json is greate (easy to parse and build tooling around), and the better option in my opinion would be json5.

C / C++ seems to just not have language package managers. The linux / bsd crowd seem to have decided that the system package manager also should be the language package manager. Which might have been fine if every Linux distro used the same system package manager.

Instead we end up with a N x M problem. Where we have a bunch of different operating systems and they all support multiple system package managers. So there's no easy way of distributing, referencing and updating C / C++ packages.

It is also my opinion that the compiler / runtime should be a package dependency. I don't like Rust's split between rustup and cargo (they should have been one tool). Similarly it would be better if you added Node as a dependency to package.json, that way we wouldn't need NVM.