mithametacs

Building something that isn't Turing-complete is surprisingly-hard once it's complex enough.

If basal intelligence is present in diverse computational structures, then weak intelligence is everywhere.

If weak intelligence is everywhere, Earth-like planets are everywhere, ... where are the aliens?

Personally, I blame game theory. Too many agents too smart in one place, you get conflicts, and eventually someone breaks an atom apart in your direction.

Or do you need emotions to have conflict? Are there basal emotions?

I'm usually not worried about AI uprisings, but I do believe in the possibility of conflict.

Reminds me of glitching the integer circle algorithm: https://nbickford.wordpress.com/2011/04/03/the-minsky-circle...

Absolutely gorgeous and probably even worth actually using

Correct, they just don't return anything. Which is the right behavior sometimes and the wrong behavior others.

I’ll wait for the GUT before I declare something the most depressing concept

We're going to have to move past considering an LLM to just be a model.

It's a database. The WYSIWYG example would require different object types to have different UI components. So if you change what a container represents in the UI, all its children should be recomputed.

Need direct association between labels in the model space and labels in the UI space.

It seems like an elegant addition to the understanding.

Maybe if you viewed an animation where `a` starts larger than `b`, and steps till it's smaller.

Then you could see where the negativity happens. And seeing is nice.

Both Pathfinder 2e and the Lancer TTRPG rulebooks are basically software implementation guides.

Lots of use of tags and other capitalized technical terms. Information presented roughly in the order you would program them.

Excellent on the technical end but also understandable.

And a good rulebook has pictures/ examples too.

Lancer does a great job of also presenting mechanics in the order you’ll encounter them. High-level abilities aren’t available till late game and the new mechanics they introduce only become relevant late game. So they’re later in the book.

If someone has this level of control of my system, pictures of me are my least concern. You can have them. Fuck off away from my bank account tho.

Wild.

Okay, keep deleting bug tickets.

Storage is cheap, database indexes work. Just add a `TimesChecked` counter to your bug tracker.

Now when considering priorities, consider not just impact and age, but also times checked.

It's less expensive than going and deleting stuff. Unless you're automating deletions? In which case... I don't think I can continue this discussion.

I don’t think it is exponential. In fact, one of the things that surprises me about software engineering is that it’s possible at all.

Bugs seem to scale log-linearly with code complexity. If it’s exponential you’re doing it wrong.

Everything is finite including bugs. They aren’t magic or spooky.

If you are superstitious about bugs, it’s time to triage. Absolutely full turn disagreement with your directions

Dawg, not every sentence someone says is 100% literal all the time.

Is this a serious comment? Use your intelligence.

Something is only as illegal as the enforcement put into it.

One can hope for a silver lining.

Not necessarily. Your prompt could include instructions to gather information from your emails and address book to tell your friend about all the relevant contacts you know in the shoe industry.

Or for something that needs like a quick mathematical lemma or a worked example. A comment on what is fantastic.

You still use type systems, tests, and code review.

For a lot of use cases it's powerful.

If you ask it to build out a brand new system with a complex algorithm or to perform a more complex refactoring, it'll be more work correcting it than doing it yourself.

But that malformed JSON document with the weird missing quotation marks (so the usual formatters break), and spaces before commas, and the indentation is wild... Give it to an LLM.

Or when you're writing content impls for a game based on a list of text descriptions, copy the text into a block comment. Then impl 1 example. Then just sit back and press tab and watch your profits.

Anyone who buys an expansion pack to Factorio is a logistics masochist. And they know it and that’s fine.

My intuition is that you’d get better results emptying the tags or replacing them with some other delimiter.

Keep the structural hint, remove the noise.

yeah, you could just pull the parser out of any open source browser and voila a parser not only battle-tested, but probably the one the page was developed against

And even if we understood it, it’s still a tiny amount of text.

My educated guess is it’s going to be something like: Zartkir is an asshat

LLMs are shit at generating content, but summarization works really well.

I’d like to use your project

I've always wanted to play a game based on defunct theories. I'm a fan of luminiferous aether myself. What are the impacts on a spacefaring civilization?

Sci-fi already grants alternative physics to enable FTL and other magic. What about hard sci-fi, but wrong-hard sci-fi?

Extra credit: go back to Zeno and all motion is paradoxical, what would you even do in the game?

You seem to know your stuff.

Will this technique work with existing model weights?

This is going to shred when it reaches industry.

But yeah very math heavy for a software engineering paper.