Language, communication, are more like the latter, communicarion benefits from everyone using a common language, except when that niche sociolect is domain specific, which an emdash isn't.

But LLM reviewing LLM? I think if the review LLM catches it, then the executing LLM would refuse to run it, and if the prompt fools the executing LLM, it will probably fool the reviewing LLM.

Also it looks very silly? Like I know it sounds like a joke, but optics matter, imagine you are getting paid a salary in tender money to feed your family, would you really want to get caught with this anywhere in the chain at all? Regardless of whether it contributed to the vuln, or just failed to catch it, will you defend your role in a company with this? Unless you are deep into the AI is a god/gold mine, it sounds like buffoonery.

Mathematics is famously rigorously defined, it's roughly analog to AI beating humans at chess. Sure it's impressive, but it's also something you'd expect machines to be good at.

Either way, I'm not reading it, it's a clanker or a clanker collaborationist.

I mean, how would you even write an em dash? There's no button in the keyboard for em dashes, it's not in ascii, it's just not something we write in internet text with, it's a safety watermark put into LLMs by OpenAI to help making LLM generated content identifiable as such.

If for some reason you are an em dash lover that was hurt by the LLM debacle, I'm so sorry for your loss, but look who's on your side, give the em dash a funeral and let it go.

"People are asking the same questions and an answer is generated every time, what if we could like cache the questions and their answers..."

Sounds like someone was using chatgpt to understand how chatgpt works and then asked it to generate a paper based on his proposal to improve it.

Like how the strawberry example was overtrained for, or how the pelican on a bike started being used in official release posts.

I guess it happens when people vibecode and the llm generates code, and it cannot verify that the interface actually works well. Even with agents that use drivers either directly or through playwright/selenium, they might only see a single screenshot, so no video.

Anyways, immediately stopped reading, as the new adage goes, why should I read something you weren't bothered to write?

>The PSP’s web browser is - charitably - pathetic. It is slow, frequently runs out of memory, and can only open 3 tabs at a time.

Alluring, an annoying property of private software development is that making websites and software in general inaccessible to lower end hardware is actually a positive effect, as it filters out 'undesirable' lower-income prospects.

That, along with pressure to produce fast, without much concern for quality (with notable privileged exceptions of luxury software like Apple or 1B+ user software like Google), as well as a disregard for sourcing "I don't care if you do it yourself, or npm install software from effectively unpaid volunteers", ends up in a state of software lacking craftmanship, software that one is not proud of to work in.

Even in the happy case where nothing bad happens, you get a badly integrated product, because you integrate not against the actual vendor, but against a abstraction layer that commoditizes the actual product, effectively forcing you to either use the least common denominator of features, or circumventing the actual aggregation model itself with some kind of 'vendor_specific_parameters' parameter in the aggregator API.

My thesis is drop the vendor neutrality, and build your integration with the vendor directly.

Well, they disclaimed and the user acknowledged

Source: https://www.youtube.com/watch?v=P-QaEB5eXSU Minute 3:00

But you can't do anything about that except asking them to validate their address with an email.

If you can catch 50% of user errors with some complex regex, but the other 50% such errors are uncaught, is that of any benefit during sofware design? No, because you still have to solve that problem, probably with email validation by code. You have reduced your workload by 0%, you just split it into 2 parts (unnecessarily).

Functionally there's no false positives or false negatives

Additionally, if a comment were to hypothetically point out an issue, that is valuable on its own. If someone reacts to a comment that points out an issue this defensively, it's a huge red flag.