Reminds me of Mike Jittlov's vehicle in Wizard of Speed and Time (1988)

I assume this is a double-bluff and the blog post WAS written by an AI o_O ?

I don't get it (and I'd call this cumulative not incremental)

Why not at least wait until the key is complete - what's the use in a partial key?

Doesn't it do exactly that?

> As a consequence of 1 and 5, we only add a property to an object once we have the entire key and enough of the value to know that value's type.

Their example in the README is extremely misleading then. It indicates your stream output is

name: A name: Al name: Ale name: Alex

Which would suggest you are getting unfinished strings out in the stream.

How is it misleading? It shows that it gives back unfinished values but finished keys.

Cumulative is a good term too. I come from the browser world where it's typically called incremental parsing, e.g. when web browsers parse and render HTML as it streams in over the wire. I was doing the same thing with JSON from LLMs.

If you're building a UI that renders output from a streaming LLM you might get back something which looks like this:

  {"role": "assistant", "text": "Here's that Python code you aske

Incomplete parsing with incomplete strings is still useful in order to render that to your end user while it's still streaming in.

In this example the value is incomplete, not the key.

incomplete strings could be fun in certain cases

{"cleanup_cmd":"rm -rf /home/foo/.tmp" }

If any part of that value actually made it, unchecked, to execution, then you have bigger problems than partial JSON keys/values.

Incremental JSON parsing is key for LLM apps, but safe progressive UIs also need to track incompleteness and per-chunk diffs. LangDiff [1] would help with that.

[1]: https://github.com/globalaiplatform/langdiff/tree/main/ts

Why not just chunk the json packets instead?

Only because you have access to the incomplete value doesn't mean you should treat it like the complete one...

Yeah, another fun one is string enums. Could tread "DeleteIfEmpty" as "Delete".

I imagine if you reason about incomplete strings as a sort of “unparsed data” where you might store or transport or render it raw (like a string version of printing response.data instead of response.json()), but not act on it (compare, concat, etc), it’s a reasonably safe model?

I’m imagining it in my mental model as being typed “unknown”. Anything that prevents accidental use as if it were a whole string… I imagine a more complex type with an “isComplete” flag of sorts would be more powerful but a bit of a blunderbuss.

Currently working on http://untamed.earth

Scratching an itch, the intention is that its a map, centered on the users that shows all (configurable) things of interest near by. Think of Atlas-Obscura but much more local - eg AO doesn't list every prehistoric burial mound on the planet, but I want to know where they are ;)

Perhaps we could ever so slightly dissuade malware actors by assigning less desirable names.

E.g. instead of 'BlackLotus' => 'LameEffort-12'

'BlackLotus' => 'BlackerLotus', way less desirable.

Or BlackOrifice as a pun on the old rootkit.

Cybersecurity has become big business since then though, and as such a much more carefully PRed rubber tile community, sadly.

Especially because Black Lotus Labs exist and researches malware.

https://www.lumen.com/en-us/security/black-lotus-labs.html

I put an episode up for each star-sign this weekend.

Most of them have had a 40-50 listens but the one linked here has 400+ listens!

The date was my mistake - the 'Aquariums' was for LOLs. (see also 'Librarians' and similar)

It's prerecorded snippets that came out of my mouth ;)

Sure.

For dates etc - you got it. I think from memory it would be 'Wednesday' + 'the 18th' + 'of' + 'may...' + '20' + '22'

For the narrative speech it would be more words in a file. There are plenty of files (EDIT: just checked 350ish files that cover all the variations of script that can be generated at the moment)

In general the TTS - part of the project is the 'art of the almost possible' (if TTS engines sounded really good - I'd have just used one of the shelf)

How did you come up with the initial list of phrases? Did you do some kind of analysis of other horoscopes?

Listened to a few podcasts, read a few - then tried to come up with some combinations that (I hoped) were funny :)

Here's all of the current 'starts' for the main prediction:

(Note they're all pretty non commital - so anything could come next)

A bite from a wild animal

A financial matter coming to a head

Completion of a long delayed task

A seemingly generous gesture

A sudden realisation

An agreement with a headstrong peer

An unavoidable slowdown

Being pulled between two emotional options

A sudden eruption of feelings

Investigating a proverbial - light in the woods

Involvement with a purely privte project

Making peace with the past

The chance of a big win

Todays socialising

Time to start lobbying for a 'generated'=true/false flag on RSS feeds?

Also, I promise to only churn out inane content for LOLs.

Yeah my bad - trust the human in the loop to put date in wrong.

As to the point, its programming practice, perhaps a stepping stone to more elaborate content-generation systems, and jolly good fun too.