plz 1

What is going on here

I have thought for a while that I should write more. Not about anything in particular, I just think that writing will be more important for me as time goes on and practicing occasionally (beyond the technical writing I do at my job) would be a good, minimal sort of investment to make in preparing for that future. So this is that. Also, I love reading Matt Levine’s newsletter and have shamelessly plagiarized his general format to jump-start the writing process. He is probably not the originator of this style, whatever. I’ve tried to make any technical details legible for a non-technical audience.

IBM

There’s this IBM cloud commercial that features Mike Singletary (a former NFL player and coach) hanging out on a football field and extolling the virtues of IBM’s cloud offerings while an anonymous football team runs drills around him:

Why should your business go hybrid? As a coach, my offense is nowhere without a solid line and agile skill players. I need the speed of a running back and the security of a linemen. A hybrid cloud approach with IBM gets your business both so you can manage data and apps, whatever plays come your way. When security meets agility, you’re unstoppable.

So it’s a metaphor. Who is this commercial targeted to, exactly? I mean, I see it all the time on network television so they’re going a little broad here. Maybe that’s a good strategy, I really don’t know anything about marketing. But there are people whose job is to decide what cloud computing provider their company uses. Do those people see this commercial and say ‘wow, compelling pitch Mike, I guess I’ll talk to my CEO on Monday about IBM’s cloud’? Does that commercial speak to C-suite executives on some level that is inaccessible to me? Because that commercial is absolutely bizarre. It seems like if you have the knowledge required to evaluate the merits of on-prem vs. cloud vs. hybrid computing strategies for your company then you will probably laugh at a football-as-computing-strategy pitch.

But IBM spends a billion-plus dollars on marketing annually so I think they know what they are doing here. I think this means that the best way to sell enterprise cloud services is to appeal to a person that has, like, a vague conceptual understanding of exactly what the ‘cloud’ is. IBM has probably done focus groups and A/B tests and paid Gartner five million dollars for a report that says ‘brand association improves the most when you couch your software services in familiar metaphors’. And their ads are sort of all like this. Maybe they found out that the really knowledgeable, domain-expert type customers can’t be sold to via TV commercials so this is the only type of ad that makes sense. Or maybe it’s for general brand-awareness purposes and they’re just getting the word out, I don’t know. There are lots of bad ads for all sorts of things, this is just another one of those.

But the Occam’s Razor conclusion is that there are people who purchase giant cloud services contracts that have no idea what they’re doing. Or, they don’t purchase the contracts but they at least influence and approve those purchases. And maybe there are more than a few of them? Not that they are bad at their jobs — they might be great at business strategy or high-level technical management — but they know so little about IBM’s offerings that a football metaphor is what gets the job done. This Timbaland one even has sort of a FOMO, be-a-cool-kid undertone that does not exactly disabuse me of this idea.

It’s a little weird to imagine, like, a Kroger SVP (having internalized IBM’s brand messages) discussing the benefits of the IBM Hybrid Cloud™ with their peers. They probably don’t really know what makes a specific cloud provider’s technology a better fit for their business than the other guys’ stuff. Apart from speed and security and data and apps, I mean. But I think this sort of thing happens all the time. I get more suspicious any time I hear a Fortune 500 executive say anything about the blockchain (another blockbuster IBM campaign!). Like, you don’t know what the blockchain is! And I am a crypto bear but the underlying concept (cryptographic proof-of-work) is powerful and important and has upsides and downsides and it is painfully clear when someone’s knowledge of that idea stops at the immutable-decentralized-system level of detail. This is clearly a thing.

There’s not really a point to this, other than to quietly posit that some of the emperors are naked and that market behavior and anecdote both kind of support this idea. Also, there is the well-known problem in software companies (and of course all companies are sort of software companies now) where the people who purchase enterprise software are not the same people who use the enterprise software; maybe this IBM ad sheds a different sort of light on how that disconnect manifests. Mostly though I think this is just silly and fun.

Scott Alexander

Scott Alexander had a popular blog called Slate Star Codex. I was not a regular reader, but I’ve stumbled across maybe a dozen of his posts over the last few years. The topics were relevant but wide-ranging and the posts were typically long. My lasting impression of his writing was that it was intellectually honest; he approached topics fairly in an effort to understand them. This piece is a representative example.

Last June, a reporter from the New York Times approached Alexander for a story about his blog. The reporter told Alexander (who wrote under his middle name to avoid attention from his psychiatry patients) that as part of the article his full name would be published, which Alexander objected to. In an attempt to force the NYT to spike the story, Alexander deleted his blog.

Eight months passed, and then yesterday a story was published that (while not exclusively about Slate Star Codex) makes a number of claims and insinuations about Alexander and his blog. Alexander quickly published a response on his Substack (where he has migrated his writing, with compensation).

So that’s the story. I don’t have a lot to say about this except… it’s not a great look for the NYT? You should read the story and the response and make up your own mind. Obviously I am biased but I was not impressed by the reporter’s depiction-by-association of Alexander, and some of the insinuations about him reach a level of stretch where the piece begins to smack of agenda. At any rate, it’s an odd and unfortunate saga between a popular, niche blogger and the paper of record.

The ex-FAANG meme

FAANG is an acronym that refers to Facebook, Apple, Amazon, Netflix, and Google, though in practice it is really a more abstract way to refer to ‘the big important tech companies’. Usually people also view Microsoft as being part of this group (and sometimes they exclude Netflix). There are other combination-specific acronyms. Lots of people are obsessed with these companies because they are very influential or because they are desirable places to work or because they are perceived as being allies or foes of one group or another. Unsurprisingly, an easy way to get clicks on your article is to have ‘FAANG’ (or its affiliates) in the title.

There is another easy but less obvious way to get clicks on your article which is to write about someone who used to work at one of these companies and refer to them titularly as an ‘ex-FAANG engineer’ (this works with other roles as well, but engineer seems to have the most cachet). This happens quite a bit on Hacker News. It is a renewable resource for clickbait. The general fascination with FAANG companies seems to extend to its individual employees and especially to former employees.

That doesn’t really make sense? Of course there are notable individuals who work at these companies, but FAANG companies collectively employs hundreds of thousands of engineers. It is odd that even average employees can capture a piece of the collective FAANG clout for themselves. One explanation for this is that these companies are viewed as being very exclusive, so if you’ve worked at one (or more) of them people will just assume you’re the best of the best. This can’t really be the case (again, hundreds of thousands of people). A more realistic explanation — at least in the case of ex-employees — is that lots of people want to work at a FAANG company so if you do work for one of them and willingly walk away you must be crazy/a genius/the next big thing.

What can you do with this? Not much, really. Some people capitalize on the attention (see above, clickbait). That doesn’t seem to dampen it and may even be a feedback loop. A common startup recruiter tactic is to mention the former employers of founders and team members as second-hand credentialing. If you are running a startup, I guess you have to weigh the tradeoff of potentially overpaying for these engineers (they are disproportionately popular) against the reputation they could confer on you. The reputation probably depreciates as the group of people gets larger and larger, though. The meme-ness of all of this indicates these strategies are already a little played out.

Replacing JavaScript

JavaScript is important because it is the only programming language that (popular) web browsers can understand. If you want to write code that runs in a browser, it has to either be JavaScript or a language that transpiles to JavaScript[0]. This has made a lot of people very angry and been widely regarded as a bad move. Because of its exclusive (so far) ability to run in the browser, JavaScript is the most popular language in the world. It is popular in spite of most of the things about it! JavaScript has improved considerably over time but it still lacks many of the features of other popular languages.

It would be nice if you could just wholesale replace JavaScript with something else. How would you do that? Well, first you’d have to create a better language that could transpile to JavaScript so that people could use the language for all their JavaScript tasks. You would also have to build a whole suite of development tools and libraries to make this an attractive option (even if your language was much better, the JavaScript ecosystem is compelling for developers). Over a period of years, your language and technologies could be adopted by the industry at-large.

Let’s say you did a really good job with all of this and your language became very popular, more popular than JavaScript! You would then have to persuade one of the major browser vendors (Apple, Mozilla, Google, and Microsoft) to include a virtual machine[1] for your language in their product. If your language was performant, this could result in a significant improvement in users’ experiences (remember, lots of web code is now written in your language) so they would have some incentive to do this. If the results were positive, other browser vendors would likely follow suit.

And that’s it! Everyone now writes code in your new language and JavaScript is a quirky chapter in the history of the web. Maybe it eventually gets sunset, I don’t know. That part would be difficult and messy. Actually, all of this would be difficult and messy. Almost impossible, really. The only language that got even part of the way down this path is Dart.

Dart 1.0 was released in 2013 and there was a version of Chrome with an embedded Dart virtual machine[2]. The language syntax was similar to JavaScript, but Dart was packed with features that JavaScript lacked and (especially at the time) looked very strong by comparison. But that’s where the parallels between Dart and my hypothetical plan end. Dart never gained significant traction in the broader developer community, despite being popular inside of Google. Many years passed and in 2020 Google announced that it would shift its external Dart focus from AngularDart toward Flutter, which has its own web support.

Disclosure (ha!): I work at Google which created Dart and I use Dart professionally, but I wasn’t at Google when all of this happened and I’m not privy to any non-public information about Google’s intentions or strategies at the time. I just think it’s interesting to picture a world where some other language replaces JavaScript and Dart is the only one that ever took a meaningful step in that direction.

Bookmarks

Shitbowl: “The algorithmically powered in-home physical caching platform”. Ubuntu reps creepily contact Azure users. British couple travels across the U.S. by train. Calvin and Hobbes search engine.

[0] This is increasingly not the case, but WebAssembly has a ways to go before it can be a plausible alternative for most of the things that JavaScript is used for. As I understand it, the largest obstacles right now are memory footprint and web ecosystems for non-JS languages. In any case, this is a backwards-looking story.

[1] A piece of software that lets the computer run code in your language directly, rather than transpiling to JavaScript.

[2] Dartium, a riff on the name of the open-source Chromium engine that powers Chrome.