AlexErrant

> Purified recombinant Reelin was injected bilaterally into the ventricles of wild-type mice. We demonstrate that a single in vivo injection of Reelin increased activation of adaptor protein Disabled-1 and cAMP-response element binding protein after 15 min. These changes correlated with increased dendritic spine density, increased hippocampal CA1 long-term potentiation (LTP), and enhanced performance in associative and spatial learning and memory.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3166788/

Published 2011. It's been a while... I wonder what it would take to reach human trials.

How do you feel about hierarchical tags? No, I don't mean in the way that Anki has a "file tree/directory" kind of structure for tags and decks, but like, a separate "concept" hierarchical graph that organizes the tags.

For example, consider two tags, "Fractions" and "Prime Numbers". One should know Fractions before they study Prime Factors, and one could represent that using a drag/drop UI like <insert generic mindmap tool>. This "concept hierarchy" would organize using tags. This way you could still syntactically tag cards with, say, "Prealgebra/Fractions" and "Prealgebra/Prime Numbers". In this way one could have a "syntax" tree and a "semantic" tree/graph. (File trees can't represent graphs, as you may know.)

One problem with making cards directly hierarchical with each other is that people have many cards, and organizing them individually can be a pain (and questionable usage of time).

> OPFS doesn’t come with graceful handling of concurrency out of the box. Developers should be aware of this and design around it.

There's a multiple readers and writers proposal [0]. It's been "position: positive" by Firefox [1], implemented in Chrome [2], and ignored by Webkit [3] (of course).

    0: https://github.com/whatwg/fs/blob/main/proposals/MultipleReadersWriters.md
    1: https://github.com/mozilla/standards-positions/issues/861
    2: https://chromestatus.com/feature/5172892632875008
    3: https://github.com/WebKit/standards-positions/issues/238

Love the shoutout to Roy Hashimoto. He's been writing VFSs for SQLite-on-the-browser and perf testing them. He's recently wrote "IDBMirrorVFS", which "is a new example VFS that keeps all SQLite files in memory while persisting to IndexedDB". It has remarkable performance, of course. https://github.com/rhashimoto/wa-sqlite/discussions/189

Also, looks like Roy takes advantage of the aforementioned proposal when he wrote `OPFSPermutedVFS`, which takes 2nd pace behind `IDBMirrorVFS` w/r/t perf. https://github.com/rhashimoto/wa-sqlite/blob/master/src/exam...

Possibly not what OP was thinking of, but maybe this one https://www.thisamericanlife.org/283/transcript

Ctrl+f "video"

I was kinda hoping to incite discussion about reasoning AIs and how they could work, hence my STaR callout, but I think we're all equally in the dark here.

> Strawberry has similarities to a method developed at Stanford in 2022 called "Self-Taught Reasoner” or “STaR”

Previous HN discussion about "Quiet-STaR", 280 pts with 265 comments https://news.ycombinator.com/item?id=39713634

FYI `schemaCompletionSource` exists https://github.com/codemirror/lang-sql#user-content-schemaco...

But if you want an example of async completion, I do it here

https://github.com/AlexErrant/Pentive/blob/fe86a8a185e4be470...

which is called from here

https://github.com/AlexErrant/Pentive/blob/fe86a8a185e4be470...

Perf isn't really a problem for me since I'm calling SQLite in the browser; it's maybe 10-20ms.

I've been using CodeMirror+Lezer to write a custom query language similar to GitHub's (somerepo stars:>100). It's been great (though unfortunately time-consuming), and Marijn's been incredibly responsive/helpful on the "discuss" forum (as long as your question puts in a modicum of effort, hah). He's merged some PRs I've made, and treated many discussions as feature requests and implemented them.

There are many ways to sandbox Javascript, both serverside and browser-side.

Are there any ways to "sandbox" DOM access? I.e. give untrusted 3rd parties access to a DOM element in a predefined spot? AFAIK the only tech that allows for this is iframes, which are unfortunately heavy and slow. I'm writing an app that can host plugins, and unfortunately, I think giving plugins DOM access means they can now literally do literally _anything_.

I'm just here to commiserate with F# being the red-headed stepchild of dotnet. Everytime people say "dotnet" they really mean "C#".

You're asserting properties, so IMO this meets the definition of PBTs ("every node of level greater than one has two children").

However, depending on the lib, you can get some nice quality of life improvements. One "nice to have" is shrinking. See the "Shrinking" section here https://tech.fpcomplete.com/blog/quickcheck-hedgehog-validit...

Having combinators to compose generators is also great.

Libs may also have a known set of "bad" values that cause exceptional behavior.

My usage of "parallel/await" is entirely metaphorical; I was kinda going for a Javascript-esque syntax with "await Parallel" - I'm assuming most people aren't familiar with F#'s `and!`. It doesn't make sense for applicative generators to (necessarily) use the thread pool.

Haha yeah I kinda went off the deep end with applicatives. Here's a short primer on applicative vs monadic shrinking behavior using F# syntax https://github.com/hedgehogqa/fsharp-hedgehog/issues/419#iss...

You can think of `let!` as `let + await`, and `let! x ... and! y` as `await Parallel([x, y])`.

Please feel free to ask any questions if it's still confusing!

> - rose tree "integrated shrinking" (eg. Hedgehog) follows the constraints of the generators, but has issues with monadic bind.

We're at the limits of my amateur knowledge, but I believe this is a fundamental limitation of monadic bind/generators. Instead, you should prefer applicative generators for optimal shrinking. https://github.com/hedgehogqa/haskell-hedgehog/issues/473#is...

In other words, applicative generators do not use "results of generators to dispatch to another generator", but instead shrinking is optimal due to the "parallel" nature of applicatives (I'm using "parallel" in the monadic sense, and not the sense of article's "threading" sense). Since applicatives are "parallel", they can shrink the generators independently. (Whereas monadic generators are in "series" and therefore shrinking one necessarily changes the behavior of the subsequent generator, as you noted.)

Please feel free to link your talk if it's public!

Another example of biphasic programming is parser generators with DSLs for generating parsers, e.g. Tree Sitter or Lezer.

What does "scale" here mean? Sqlite can go quite far https://use.expensify.com/blog/scaling-sqlite-to-4m-qps-on-a...

I agree with you w/r/t "Why build your own query builder" though since (AFAIK) you can't touch the underlying SQL. I'm making heavy usage of FTS5 in my app, and it doesn't look like they support that.

And, to be fair, feathers/rethink also have the weaknesses you list:

- Why build your own query builder

- This also isn't really something you could drop into an existing project very easily

Personally, I've been pretty happy with https://github.com/vlcn-io/cr-sqlite/ which is just a sqlite extension that can run in the browser.

Discussed yesterday https://news.ycombinator.com/item?id=40762298

As context, they're aiming for first plasma in 2026

https://www.axios.com/pro/climate-deals/2024/05/01/commonwea...

They deleted a lot (all?) of the negative comments on that video too, hah. At one point this was their advice:

> We see this so commonly at YC that we have a term for it. It's called a "Solution In Search of a Problem," or a SISP. And these are usually not great, because usually, you never actually find the problem. You're much better off starting with a problem, and then looking for solutions.

https://www.ycombinator.com/library/8g-how-to-get-startup-id...

Very few people have a problem that AI can currently solve. It might not _stay_ that way... but man I wonder what kind of moat a startup could build when they're just wrappers around foundation models.

Maybe use pglite (which wraps postgres-wasm)? Readme indicates it can run in-memory.

Doesn't solve your testing-data question, but it'll save you from spinning up a new DB.

https://github.com/electric-sql/pglite/

https://github.com/electric-sql/postgres-wasm

The test suite is (partially) proprietary. Good luck maintaining a fork for a DB without tests.

https://www.sqlite.org/th3.html see bottom

> Open source AI models allow for greater transparency, collaboration, and innovation by making the underlying code publicly accessible and modifiable.

I feel like open models do virtually nothing for transparency, collaboration, or innovation, and are only modifiable in that they can be fine-tuned. It's "open source" training processes and data that will lead to "transparency, collaboration, and innovation", and I'm unaware of any large company that does this.

Am I wrong?

Meta:

455 points, 3hrs old, but on the 2nd page of HN. What's up with the algo?

I consulted on some tech for this company https://www.thereversalclinic.com/

I'm not in medicine and can't speak to how effective it is, but I'm guessing there's a reason why he's still in business.

I have a Charybdis https://bastardkb.com/charybdis/

Highly recommend. I daily drive vim and still scroll around in there with the trackball haha.

My trackball's also on the right, and also use mouse buttons on the left, but of course everything's customizable.

More details https://github.com/orgs/open-spaced-repetition/discussions/1...

Nice!

(Jarrett's the author of fsrs, fyi.)

FYI I made Anki's spaced repetition algorithm, FSRS, run in the browser. Might be useful to you https://github.com/open-spaced-repetition/fsrs-browser

Edit: I just remembered it takes like 500(?) data points to train the algo... so maybe not.

I've been using GitHub code search a lot to get examples of how people use APIs or designing features. It's quite good.

I recently did this to figure out how to insert JSON into Sqlite's FTS5 https://github.com/brouberol/5esheets/pull/292

I didn't use AI because for these things because for some things (database design, constructing a grammar) you _really_ want to understand your code completely, and I don't want to deal with hallucinations.

> My real worry about seed oil theory is that it’s a distraction.

Which do you think is worse: smoking, or seed oils?

Which do you think is worse: smoking, or not exercising?

https://jamanetwork.com/journals/jamanetworkopen/fullarticle...

Look at Figure 2C in particular. The hazard ratio of smoking is 1.41. The hazard ratio between "Low" (<25th percentile) and "Below Average" (25th-49th percentile) cardiovascular health is 1.95.

In other words, it is better to be a smoker than in the lower 25% of non-exercisers. By far. It's not even close.

Seed oils are 100% a distraction. Once you're in the upper 2.3% of vo2max for your age/sex, _then_ you have the right to worry about seed oils. Until then, save your energy for interval training.

Minor in the minors, major in the majors.