Here, I describe the character of future content which readers can expect from this blog, in order to gently dissipate the second cloud.

What I’ll say

While my original ambition was to say every true thing once, I’ve realized that some readers may prefer that I write more sparingly—about only their scrutability base—instead.

A scrutability base is “a class of truths from which all truths are scrutable, for a given notion of scrutability”. Say the notion of scrutability is “a priori knowable by an ideal reasoner”. (I assume the reader is competent with these terms and satisfied with this definition, as she is no doubt an ideal reasoner. I myself am not certain e.g. what the relevant notion of a class is, but let’s move on!)

So, what goes into our scrutability base? Taking Chalmers and Jackson (2001) for a start, there are four kinds of contingent truths from which ideal reasoners can deduce³ all other contingent truths: P, the microphysical truths; Q, the phenomenal truths; T, a “that’s all” truth; and I, the indexical truths. Call this the PQTI base.⁴ PQTI leaves some blog posts unpublished, but leaves nothing unknown.⁵

While the ideal blog would tackle at least the PQTI base, observe that this even this seems like a lot of work. Below, I sketch what subsets of PQTI I intend to cover.

P: Microphysical coverage

It seems fair to cover some, but not all, of the microphysical truths. I’d like to learn some physics again, and blogging about what I’ve learned would be a fine reward. (For example, I would like to remember how a superconducting qubit works, and what part of the circuit I soldered all day. I’d like to be able to read philosophy of physics papers for various topics in physics.)

While I would never so much as insinuate that the reader could fail to compute macrophysical truths from the relevant microphysical truths, I would nevertheless extraneously cover some macrophysical truths. For example, I might discuss candidates for physical bottlenecks to self-sufficient AI and how soon / at what levels of AI capability they might be removed. (Most metaphysical truths pertaining to AI takeover threat models will probably go on Lesswrong or Substack, but we’ll see.)

Q: Phenomenal coverage

A lot of the phenomenal truths on offer in this blog will pertain to what exactly it is like to be me at time T. While most such truths will be reserved for private musing, some are already haphazardly circulated, and it seems fine for them to go down on this blog and amount to a public record⁶ of my most formative and influential thoughts, feelings, attitudes, (literal) dreams, recollections, aspirations, etc. Perhaps not much else can be scried from this record, but I hope they will at least be entertaining.

I am not confident, to say the least, in my ability to offer the reader other phenomenal truths. The least I can do is try to acquire other points of view which affords access to more phenomenal truths over time (such as birdwatching, I’m told, and perseverant blogging which affords the point of view of a blogger).

In order to know the phenomenal truths I plan to convey, the reader must have, as a prerequisite, the relevant phenomenal concepts (referring to a phenomenal property which is also its own mode of presentation), such as redness or the smell of Yellow Fruits when you squeeze air out of the valve (see below).

Because it is difficult to convey phenomenal truths, I may wax poetic or speak loosely and figuratively and even resort to non-standard punctuation or none at all. I may lavish other forms of media. This deviation from my usual upright writing style must be construed as form forced by function rather than defects in ability or character.

I: Indexical coverage

Indexical truths are truths like “I am 24601” or “Now it’s Wednesday morning, 3 A.M.” which involve indexical expressions whose reference depends on context. My rough understanding is that we only need, for the PQTI base, some small number of self-locating indexical truths.

Unfortunately, my self-locating truths (“I am Alexa”) are just ordinary macrophysical facts (“the author of this text is Alexa”) to you. In order to complete the PQTI base, readers will need to locate themselves.⁷

T: “That’s all” coverage

A “that’s all” truth says “that’s all—there are no further truths”. This truth is thrown in to save us the trouble of enumerating (the infinite and innumerable) negative truths about what doesn’t exist and what isn’t the case (e.g. “there are no pelicans”).

I haven’t looked closely into the motivation for including this in the base; at a glance it seems kind of silly, which makes me reluctant to blog about it.

More importantly, it seems unlikely that I will ever exhaust PQI through writing. So let’s not say “that’s all”, only “until I write again”.

Exceptions and disclaimers

1. I’ll say some things outside the scrutability base, e.g. mathematical and logical truths which are already apparent to the ideal reasoner.
2. I’ll say some things whose status in relation to the base are still yet to be determined by further reflection, e.g. normative facts, grounding principles.
3. I may say some things that are false, unknowingly, since I will say what I believe to be true but some beliefs may be false. (Just in case readers mistook me for infallible.)
4. I may say some things that are false, knowingly, for ironic effect (“O do not love too long”).
5. I may say some things that are not truth-apt (“boo neuralese”, or bad poems).
6. All opinions are not my employer’s.

When I’ll say them

The present aspiration is to publish substantive posts regularly if infrequently (perhaps with each full moon), with occasional short notes.

There is a further question of how long one should wait to say certain facts. Should one say them immediately, from a first-person perspective, while the fact still obtains according to the present time, or with substantial delay, from a third-person perspective, when the fact no longer obtains according to the present time? (If these facts are in fact temporary, but almost all facts benefit from some fermentation and reflection.)

I’d guess the right approach depends on the material (Yellow Fruits go stale, black vinegar and presumably wine improves with time). I’ll think of this blog as closer to coffee for now.⁸

Why I’ll say them

Maybe I should have mentioned this earlier. The goal is, of course, to scry the (your, my) universe and to be scried.

Sadly, even if we have done our best and written down all of PQTI, the universe and I may remain inscrutable. The a priori entailment thesis may not hold⁹; I and the reader may not be able to infer all truths a priori from only the fundamental truths.

Nevertheless, we can talk as if smiles exist.¹⁰

How I’ll say things

Mostly clearly. Sometimes idiosyncratically, hopefully beautifully, hopefully.

I’ll say some in purple and some in pink, some in klein blue.

I’ll say them like Bob Dylan and Jackson Browne.

(Look, I warned you about the phenomenal facts.)

Who I’ll say them to

Following Frank, I will write to you.