Pinned toot
Pinned toot

Sometimes, I make really compelling arguments about things that I stop believing five minutes later, because I realise (or am informed) that they're wrong.

Don't take my word for anything. If something I've written has convinced you of something major, it might be worth asking whether I still believe it – but if I don't, don't take *that* as gospel either.

I have a very simplistic view of the world. Sometimes that means I miss obvious things for a long time.

One of these paragraphs is wrong.

Pinned toot
Pinned toot

The Copenhagen Interpretation of Ethics. Almost No One is Evil. Almost Everything is Broken.

Pinned toot

Lately, just fills me with a profound sense of disappointment. It's the best programming language I've ever used, but there are so many things that you *should* be able to do but you just can't.

enum Example {

impl From<u64> for Example {
from = Example::Hi;

That *should* work. In plenty of other places, Example::Hi is treated as a function with the correct prototype. But Rust's syntax doesn't let you express *this*, so you need boilerplate.

Me: Urgh, const generics are still experimental? … *Fine*, I'll use Rust nightly, but only until that's stable.

Me, 5 seconds later: Ooh, shiny!
Two screens full of `#![feature()]`: *silently judges*

There are two valid branches of : combinatorics, and Bayesian.

In my experience, frequentist statistics that is not combinatorics (or accidentally Bayesian) is not correct. I'd love to be proven wrong – frequentist statistics is *so* much easier to calculate – but I won't be.

Internet argument tip #7:

If somebody's questioning your sources or evidence, try to match the types of source or evidence that they use.

The ideal solution is to agree, in advance, what kinds of evidence you'll accept from each other. This usually requires your “opponent” to be invested in the argument, and is best done after a “prove it!” (second best after “biased!”) – though not everyone is open to such negotiation. Experience will teach you who's receptive to that and who isn't.

Internet argument tip #6:

Words can be wrong.

A word is a powerful thing. It decides how to frame an argument. It can conjure into being, or hide from view. It attaches connotations to the neutral.

All its power goes away when you replace it with another. What do you *mean* by “best”? “make a sound”? “behind”?

Your aim is to be right, not to seem right; pick your words carefully, and be quick to discard them (even if they *feel* right).

See also:

Internet argument tip #4:

Some people just aren't worth arguing with. Sometimes it's hard to tell the difference between “doesn't argue the same way as you” and “a rude, stubborn rock”, but if they're making you feel bad, stop engaging and politely leave. Even if they started it. Duty does not call.

Over time you might learn how to be more diplomatic, and, eventually, how to get useful information from such people – but that's not a skill you can learn on the fly. Cut your losses.

Internet argument tip #3:

Beware sets of ideas.

There are coherent, self-reinforcing memeplexes (ideologies) that *seem right*, on the surface, and seem *very* right when you're in them, but are actually wrong.

If you find a proponent of a set of new ideas that seem *very right*, take *just* the first idea they present you with, and tear it to shreds over the course of several days. Believe it's wrong, and try to find out why; repeat until you can no longer justify paranoia.

Show thread

Internet argument tip #2:

Everybody has something to teach you. Not everyone is a teacher. To learn the rarest lessons, you need to stop arguing and start listening.

The best such lessons seem useless at the time, and obvious a week later. (At least, they do to *me* – your response may vary.) These lessons rarely come from your friends. Seek them out.

Internet argument tip #1:

When a correct-seeming gotcha makes an incorrect-seeming point, acknowledge that they're right and ask why they think it's a gotcha. The elaboration will unpack their point and make it clearer.


I've had a lot of internet arguments, and I'm slowly working out tricks for how to divert them away from squabbling, towards productive discussion where people learn things. I'm trying out the idea of documenting those tricks. Should I?

An idol is defined by influence, accomplishments, great works. In effect, a myth. An idol is not a person.

Great works matter, but they are rarely accomplished by emulation. Great thinking matters, but it is not achieved by studying great thoughts.

I hope that some day, my former idols may become my peers – but I have learnt most of what they have to teach, studying their mythos, and they cannot learn in turn from what I have learnt. Not a week goes by when I do not learn from my peers.


Show thread

I no longer have any idols.

Partly, this is due to disillusionment. Mostly, though, it is because I have found those who can be not just contemporaries, but peers,, who I can respect.¹

An idol is somebody one looks up to, and perhaps some day hopes to surpass. One's peers? We're getting there together. Sometimes allies, sometimes comrades, sometimes friends – sometimes rivals.

¹: Respect has multiple meanings; there are few I *dis*respect, but here I mean a specific type of respect.

Why did we make so hard?

Text is, fundamentally, simple. It's a structured assortment of symbols that express meaning. It is *not* a stream of bytes! Attempting to treat it as one causes major problems – in as well as rendering.

So why are we still treating it as a stream of bytes?

@chjara Here's a puzzle, in case you end up separated from anything fun to do:

Prisoners' Dilemma (for THIS-____ player, defect-coöperate > coöperate-coöperate > defect-defect > coöperate-defect), but you get to spot-check other players (i.e. see what they'll do when faced with another player), and you don't know whether you're being spot-checked until afterwards.

If you can present other players with whatever bots you like, what's the best strategy? Okay, but now you've added *that* bot…?


@haskal That being said, CRISPR would probably make the design easier, and it'd allow more variability in the characteristics (because you wouldn't be limited to ways that regular old human cells can be made to differentiate). Sort of like Arbitrary Code Execution v.s. ROM hacks.

Oh, that reminds me: Flappy Bird in Super Mario World.

Show thread


@haskal Might mRNA be a better medium for deploying open source cat ears than CRISPR? From what I hear, mRNA synthesis is as easy as: synthesise some DNA; mix it with the ribonucleotides, your chosen cap analogue and the RNA polymerase; then control the temperature for a few hours.

Admittedly you also need a delivery system, but not altering your genome means better compatibility with other biohacks, *and* it doesn't have the “only one CRISPR deployment each n years” immune problem.

uk pol (lgbt), consultation – actionable? (help wanted) 

“Technical review on increasing accessibility and provision of toilets for men and women”

… Why did they add “for men and women” on the end? Oh, looks like it's just the government trying to ban gender-neutral toilets again – this time under the pretence that it makes the logistics better (it doesn't), and… some argument about urinals?

Looks like they want technical advice via email.

book recommendation 

@chjara Here is chapter 1:

Here is a quote from chapter 1:

> God is awesome in majesty and infinite in glory. He’s not going to have a stupid name like GLBLGLGLBLBLGLFLFLBG.

I think you'll get more of the references than I did. I enjoyed it a lot, but you're more the target audience I think. (Though the target audience is Americans, so you're not the target audience.)

Show thread

book recommendation 

@chjara Have you read Unsong? If not, I think you'll like it. It is very good. Here is the prologue:

It is a work of fiction about Jewish theodicy. Here is the table of contents:

Here is the quote at the top of the table of contents:

> *Jerusalem is builded as a city that is in the public domain.*
> —

Show older

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.