golly. it's like, like they are actually discouraging people and making them not want to read or something.... Can't be that though, right?

Yet the inverse of this is that they bugs, just plain don't work right bugs will continue to increase because not well tested contributions will continue to increase many of which will graduate to CVE status. I think they are finding shit that NEVER worked as indented with AI assistants so maybe the spectre of CVE review is necessary on an even broader scope....?

I think so as well, unconstitutional. End users never really owned commercial software because they never really owned or were ever responsible for the data created or its intregity and it was decided it was unconstitutional to make them responsible...... more or less. for the sake of capitalism.

politics. im suspecting this is an angle to actually stop the work from leaving their shop before they figure out how to profit from it? like I said politics, and not a little privilege. every few years there is some moral uproar at google that is shadowed by layoffs but its soon forgotten by the up and roaring.

I think they realized that the language models are an architecture or share one. and as an end user you have little to say about what language model you are using. they never completely stopped using anthropic because it does not matter in reality. like web browsers, what is the actual difference with search results other than the what the providers choose to allow the end user to see?

all of that AND how LLM and AI do not necessarily mean the same thing. So what is actually embedded on the device when nobody seems to care what it is or exactly how it works.

the TRS-100 supports ISDN, problem solved.

hosts files use to be good at catching stuff like this.......

also the unlife state, not necessarily the browser dev's fault is leaving the door broken on the hinges for insecure applications to get in.

Browsers have been slowly sinking in to the service layer as some kind if user agent since the early 2000's or before. Maybe disappearing altogether like how windows device manager became the agent for the user experience. Instead browsers are made to do things like talk to the OS for the user even when they aren't using the system like a transitional state of compromise.

mod parent up!
(me)!

the burocratic fail in many western countries is if you never actually applied for asylum they can decide at the meeting that you have due to bad processing. in other words the oppression you are fleeing is the immigration process at the destination country! they won't tell you or course. so this person has possibly applied for asylum(never did), denied asylum(never applied), and denied immigrant status because the job skills they never listed were deemed undesirable.

Just the same finally people are beginning to realize that every solution does not require a building full of GPU's.

The efficiency of generative AI slop or not is the point. And we've had this discussion about web design in general and how the pigeonhole effect that frameworks have on style choices, style lost. Still .net made microsoft untouchable, to give an example. Back to generative AI, I have a suspicion that the unattractive "slop" is due to constraints in choices put on the agent, like a web framework no doubt.

it made changes like a compiler optimization? no a single person can with no knowledge of a program other than the first 50 lines they parse in the beginning for example, can understand how to make it bug proof. you cannot assume calling functions , functions to be called immediate needs to refactor at all based on current needs..... it looks like the llm is interpreting code intended to work but in the narrow context the user is working in like that 50 or so lines they understand for example....