Why should I apply for *cybersecurity* approval in order to have model debug a program it is writing itself? Anything related to memory safety, debugging, syscalls etc (meaning, "programming") somehow is cybersecurity now?

Your tools refusing to do your bidding is an absurd idea in the first place

Imagine asking for permission to use your hammer

I mean I don't want autonomous cars to follow directions by humans that sound like "plow into this crowd of people". Even things like microwaves don't let you turn it on without the door closed. I don't see how this is any different.

It is fundamentally diferent because it tries to judge your intent. Thus being both opaque and unpredictable. Car refusing to drive onto person is easily understood. Autonomous car refusing to drive you to corner of Baker's street because that is suspicious destination, with no recourse or explanation, is totally different thing.

This is why you can't leave judgment to technology, and why we have law enforcement and courts

"Plow into this crowd of people" is nominally wrong, but what if its actually "plow into this crowd of people that are hurting and robbing a family with two small children"

Your smart hammer refuses to let you smash into a glass window, but what if you were smashing into the glass window to save a child stuck in a fire?

Splitting subfolder into separate repo is full history rewrite. Changing commit message or commit tree is full history rewrite. Also, if you maintain local changes, you want your rebase onto vendor be super easy, meaning you want to groom your history into nice set of changes, not "reverted: reverted: really fix bug #123 for real" tangle. Rebase itself is a history rewrite.

I am getting my hair stood up by all the Comment Engineering going on in this thread. If all of the participants are organics, who are you people? If most is bots, the prospects are grim.

whoa, "we tested" seven completely different chemicals, one has zero effect, six others have massive negative effects? By what mechanism?

Apple's Metal is C++ too. Like, actual, you can compile it with plain C++ compiler and a shim header.

On embedded, your fight totally is against language and runtime. Since you cannot gave conformant freestanding C++ without exceptions, rtti and most of STL, "using c++" on embedded always turns into "using gcc" or "using clang" — their mutually compatible, but completely standards non-conformant extensions that make writing for embedded reasonable to even attempt. At that point, is the language still c++?

Meanwhile, both rust and zig will gladly compile your standalone function into standalone binary.

Last time I looked "submitting a proposal" to WG21 involved attending in person. Has something changed recently?

I have slightly different experience. Any time during deep work on codebase when I need to stop and switch lanes to study how exactly did .SECONDEXPANSION work or which of "git reset" "git checkout" flavours will do the thing I need instead of wiping all my work, or what magic sauce ingredient again prevents Dockerfile running on Podman — that is when I start understanding less the source code of my project. Because I am trying to multitask two deep study sessions. Going "computer, make this branch contain that stuff" and then proceeding with debugging kernel crash sits very well with me

They did not misunderstand anything. All of the behaviour is not inherent in raw base model and has been planted by the agressive, secretive reinforcement learning they do for benchmaxxing, "safety" and all other things. Claude begins any other sentence with "honestly". That is not how LLMs work, that is how they work after being RLed to the brink.

Before everyone piles on this comment with "whoosh" and "it was sarcasm" and such — have you noticed that reacting to ironic, sarcastic comments as if they were meant literally is what real LVL 80 trolls do oftener and oftener? On internet, you can never know who is pulling leg...