Security through obscurity is mitigation basically. You reduce risk/impact, not eliminate it. There are problems - such as denial of wallet attacks - where you can only mitigate and can't eliminate the problem completely
All of logic and math is a convincence tool. There are no, circles, quantities. Reality just is. We created these tools because they're a convinent way to cope with complexity of reality. There are no "objects" in a sense that chair is just atoms arranged chair-like. And atoms are just smaller particles arranged atom-like and yet physics operate in these objects treating them as something that exist.
So, now we have created these mental tools called mathematics that are heavily constrained. Then we create models that are approximately map 1:1 to some patterns that exist in reality (IE patterns that are roughly local so that we can call them objects). Due to the fact that our mental tools have heavy constrains and that we iteratively adjust these models to fit reality at focal points, we can approximately predict reality, because we already mapped the constrains into the model. But we shouldn't mistake model for the reality. Map is not territory.
Yep. Humans (and other animals) have an inbuilt ability to count small numbers of objects, so whole numbers seem more natural to us, but it's just a bias.
In fact, automated regression tests done by ai with visual capabilities may have bigger impact than formal verification has. You can have an army of testers now, painfully going through every corner of your software
Will only work somewhat when customers expect features to work in a standard way. When customer spec things to work in non-standard approaches you'll just end up with a bunch of false positives.
This. When the bugs come streaming in you better have some other AI ready to triage them and more AI to work them, because no human will be able to keep up with it all.
Bug reporting is already about signal vs noise. Imagine how it will be when we hand the megaphone to bots.
A hybrid will likely emerge. I work on a chat application and it's pretty normal that LLM can print custom ui as part of the chat. Things like sliders, dials, selects, calendars are just better as a GUI in certain situations.
I've once saw a demo of an AI photo editing app that displays sliders next to light sources on a photo and you are able to dim/brighten the individual light sources intensity this way. This feels to me like a next level of the user interface.
1. There's a "normal" interface or query-language for searching.
2. The LLM suggests a query, based on what you said you wanted in English, possibly in conjunction with results of a prior submit.
3. The true query is not hidden from the user, but is made available so that humans can notice errors, fix deficiencies, and naturally--if they use it enough--learn how it works so that the LLM is no longer required.
Yessss! This is what I want. If there is a natural set of filters that can be applied, let me speak it in natural language, then the LLM can translate that as good as possible and then I can review it. E.g. searching photos between X and Y date, containing human Z, at location W. These are all filters that can be presented as separate UI elements so I can confirm the LLM interpreted correctly and I can adjust the dates or what have you without having to repeat the whole sentence again.
Also, any additional LLM magic would be a separate layer with its own context, safely abstracted beneath the filter/search language. Not a post-processing step by some kind of LLM-shell.
For example, "Find me all pictures since Tuesday with pets" might become:
Then the implementation of "fuzzy-content" would generate a text-description of the photo and some other LLM-thingy does the hidden document-building like:
Description: "black dog catching a frisbee"
Does that "with pets"?
Answer Yes or No.
Yes.
My first impulse is to say that some languages have better SNR on the internet. (less garbage autogenerated or SEO content compared to useful information)
