ianbicking

7 ms·

31.
by ianbicking 1y ago
Respect (or lack thereof) goes both ways: both the writer and reader. I have frequently felt disrespected by producing documentation, planning/etc that isn't read. In the end I mostly rely on oral transmission of knowledge because
32.
by ianbicking 1y ago
My experience writing in a professional setting is that people mostly don't read what I write, and the more effort I put into being thorough the less likely that it will be read.
33.
Hidden Door at Launch: Design Review of an LLM-Driven Story Game(ianbicking.org)
1 points by ianbicking 1y ago|0 comments
34.
by ianbicking 1y ago
Probably ignoring things like robots.txt, I'm guessing? But I'd be curious what exactly the list of things is, and if it's growing. Would it go as far as ChatGPT filling in CAPTCHAs? autocomplete="off" is an instanc
35.
by ianbicking 1y ago
The surgeon general warning came along with a large number of other measures to reduce smoking. If that warning had an effect, I would guess that effect was to prime the public for the other measures and generally to change consensus. BUT,
36.
by ianbicking 1y ago
I'm doing something similar, so some thoughts: 1. I really like the "commitment" concept. That solves a real conversational problem where the AI can be too easy to redirect, moving on too fluidly from previous conversational
37.
by ianbicking 1y ago
Thought: if one of these automation tools wants to do some deep research task, is it legit if it just goes to chatgpt.com or notebooklm.google.com? Obviously Anthropic or OpenAI doesn't need to do this, but there are a dozen other brow
38.
by ianbicking 1y ago
The use case is not that obvious to me, but I _think_ it's trying to give the LLM more context to write code by dipping into these code objects. Perhaps telling the LLM what modules are already imported or otherwise available? Creating
39.
by ianbicking 1y ago
It feels like there’s some kind of Party Train opportunity here, similar to a party bus.
40.
by ianbicking 1y ago
I'm around so much wildfire smoke lately that my sky expectations have changed... I wonder what it would take to account for weather?
41.
by ianbicking 1y ago
I went to some lengths to ground the action in my game: https://ianbicking.org/blog/2025/07/intra-llm-text-adventure That said, I think there's a lot of prompting techniques that can help here. (Actually
42.
by ianbicking 1y ago
"Why would AI be any different?" It might not go any differently, but I think all us folks here have an opportunity to make it different. Any product needs to pursue enough engagement that the user actually gets value. I have a bu
43.
by ianbicking 1y ago
I wish the doctest approach won over the BDD approach. I still find myself recreating the doctest approach in miniature... Specifically I try to make a nice string representation of state, and then tests match off that state. So for instanc
44.
by ianbicking 1y ago
The piece (and lots of commentary) keeps talking about "AI Companies" as though it's this fixed set of companies that are destined to all be the same. But anyone can start an AI company... the models are all available, and th
45.
by ianbicking 1y ago
Yes, that's the one, thank you!
46.
by ianbicking 1y ago
I remember reading an article on one of the classic rationalist blogs (but they write SO MUCH I can't possibly find it) describing something like "rational epistemic skepticism" – or maybe a better term I can't recall ei
47.
by ianbicking 1y ago
Using modern tools I would naturally be inclined to: 1. Have the LLM see the image and produce an text version using a kind of semantic markup (even hallucinated markup) 2. Use that text for most of the RAG 3. If the focus (of analysis or c
48.
by ianbicking 1y ago
I'm confused, where's the content? For instance I looked in Meta-Cognitive Techniques and it mentions "Articles in This Series" but the list doesn't have any links...? https://promptz2h.com/chapter_0
49.
by ianbicking 1y ago
For something a little different than a coding task, I tried using it in my game: https://www.playintra.win/ (in settings you can select Mercury, the game uses OpenRouter) At first it seemed pretty competent and of course v
50.
by ianbicking 1y ago
FWIW, I remember regular models doing this not that long ago, sometimes getting stuck in something like an infinite loop where they keep producing output that is only a slight variation on previous output.
51.
by ianbicking 1y ago
Some things you could do: 1. Look up compiler research in relevant areas 2. Investigate different parsing or compilation strategies 3. Describe enough of the language to produce or expand test cases 4. Use the AI to create tools to visualiz
52.
by ianbicking 1y ago
There's a hundred ways to use AI for any given work. For example if you are doing interesting work and aren't using AI-assisted research tools (e.g., OpenAI Deep Research) then you are missing out on making the work that more inte
53.
by ianbicking 1y ago
Some of the issues with parsers and knowing what to type could be helped quite a bit if the text adventure interfaces had a bit more affordances. Making important objects bold is an easy one (unless you feel that determining the importance
54.
by ianbicking 1y ago
Yes, the author is deciding How People Really Are by looking online, when the distortions of online discourse are also what she is complaining about. On the other hand, most interactions I have with people are devoid of this kind of involve
55.
Intra: Design notes on an LLM-driven text adventure(ianbicking.org)
2 points by ianbicking 1y ago|0 comments
56.
by ianbicking 1y ago
For text adventures an important kind of reasoning is Inferring Authorial Intent. Or maybe Seeing Chekhov's Gun. Or Learning The Metagame. The game is deliberately solvable, and elements are introduced to that end. Inferring that is im
57.
by ianbicking 1y ago
I've been building a memory system myself, so I have some thoughts... Why use a knowledge graph/triples? I have not been able to come up with any use for the predicate or reason to make these associations. Simple flat statements s
58.
by ianbicking 1y ago
What he describes feels so familiar to me... the ideas and projects I've cared about most have usually landed flat. And because the ideas matter to me I try over and over, hoping that there's something I can change or explain or i
59.
by ianbicking 1y ago
Yeah but you didn't actually answer the question...?
60.
by ianbicking 1y ago
Where does stuff like the prompts go? If you put them in the frontend then you have a bit of a security, monitoring/etc concern. If you don't put them in the frontend... then you have a backend. (But maybe a simpler backend for de

More ›