I've always felt that "proper" and responsible LLM use for search [0] would be two boxes: You can describe what you want in the first, and it'll propose search terms in the second, and then those get executed normally.
Yes, the "average user" [1] might not usually care about the second box... Until they need to because the query/results are wrong.
Showing them in tandem means:
1. Users are at least capable of learning through exposure.
2. Users may realize a key term can be added which the model could never have guessed.
3. Users may recognize a term in there that doesn't make sense, allowing them to detect a translation error.
4. If good search-terms leads to a bad outcome, it is possible for someone to report and diagnose it, rather than a fully black-box mystery.
_____
[0] Not just for websites, but also things like internal business software, or SQL queries.
[1] The average that might not exist. ( https://www.thestar.com/news/insight/when-u-s-air-force-disc... ) There are some features that everybody needs, just at different times.
And regarding “average doesn’t exist” - that’s true. But no company does a/b testing to land on average. I’d assume a good 85%-kinda pass rate for these type of experiments.
95% of users might not want the double-textbox, but 5% might, and some might want to use it someday.
I think, make enough of these decisions, and there's bound to be mounting friction for users with different preferences navigating your app.
Ideally, the app should offer a way for the user to intuitively configure the app to their own liking.
Ideally, yeah maybe, but why bother with extra implementation, support, costs and etc., when people fold and use the new way anyways?
I think the two-tiered input-output approach you propose makes sense. Allowing users to inspect and mutate lower-level languages allows the user to make modifications as needed. I think it's very much in the spirit of free software.
But then again, for text-based search specifically (search engines, notes, etc.), I think there is some value in querying the LLM directly, as it is able to fuzzy search by inspecting its weights. This results in a lesser degree of specificity, which allows for more false positives, but can maybe capture similar words (i.e., synonyms / typos / tenses) or higher-level semantic concepts.
Maybe they're just two different search algorithms, and the user should be able to choose between them.
Thanks for linking the article, I found it very interesting.
For a certain cohort (eg those currently between 35 and 45) who did any sort of grade school computer/library classes in the 90's/early2000's, 'chunking' or keyword searching was one of the main skills taught.
If you are properly lazy, you must expect or at least attempt to get results for the shorter query. Especially on a phone.
Many people who learned to search the Internet in early 2000s learned "boolean search queries" (terms connected with AND or OR). They're talking about the difference between boolean search queries and natural language search queries.
The kids who stopped getting that are probabaly old enough for a bachelors degree now. Today, the schools mostly just hand them an iPad and a Google Workspace sign-in and pretend everything is fine.
All those studies about GUI usability were done in the early 2000s. Before phones. When most people were just learning how to use a computer. The idea of options certainly was confusing, because "clicking on things" was even confusing.
Yet today, we have a populous which, more and more, grew up with computers. And loves to personalize. And wants options. So just keep that in mind, all you frontenders, config options aren't necessarily bad. Ancient studies have little relevance 24 years later.
Sorry to hijack a bit, but felt it was relevant to courses, teaching, and the state of time in 2000 vs now.
https://ometer.com/preferences.html
I'm not sure about you, but I find Gnome very easy to use, although I miss some features. Those features can be added with Gnome extensions that are maintained by third-parties. Meanwhile I find GIMP extremely hard to use, as it packs in too many features to allow their designers to figure out a good UX which doesn't break some component.
I wouldn't have taken keyboarding if I hadn't been forced to at the age of 8. I hated keyboarding classes. I am so glad I took them now. I hated doing boring research projects and looking crap up on Yahoo's Directory. I love that I have a basic understanding of how to rate and investigate a website's trustworthiness. Mfers don't even look at the "About" section anymore.
You were lucky to have that part. Mine, as I remember, was just Microsoft Office throughout.
Answering with AI slop is a choice.
The PageRank algorithm, and importantly, any algorithm which depends on page content and pages linking to each other and referencing each other was DoA, because it only worked in an adversary free context! The very second there was a real incentive to be at the top of Google's search results, their algorithms were going to be attacked and manipulated.
That makes "search" a neverending arms race. The very thing that launched Google as some amazing thing was doomed from the start. They've been laying hacks on top of hacks on top of hacks ever since. It's been shit basically since SEO has existed, which is what the "adversary" is.
Another good point for when google "died" was when they bought doubleclick. They were an ad company from that point on. There was never a good outcome possible after that. If ads were something that benefited you the user, you would have to pay for them.
If you haven’t read it from start to finish you should. If you have, what about it is insufficient?
Instead, the AI answered that it is "actually a famous example from internet culture used to describe how people use search engines. The phrase stems from discussions on tech forums like Hacker News regarding the differences between "keyword searching" and "natural language searching"".
Technically adept people would search using ordered keywords and phrases, which would have excellent results with Google 2006:
Dario Turkey NBA "never coming over"
But regular people agonized over this because to them a search was
What were some of the memes about Dario staying in turkey
Google put immense effort into calibrating search for regular people instead of engineers.