fwip

It's not just that we're not innovating on good new products. We're innovating new ways to break the law and exploit human psychology.

Blockchain/NFT hype did a ton of damage to tech's credibility, as well as Uber before it (along with the many Uber for X apps).

Note of course, that the code that comes out of Ghidra is significantly worse than the source code. Aside from the obvious like variable names, more complicated language features, like templates or macros, won't roundtrip.

(Unless I'm misunderstanding your comment, and "horrifying" was not meant to describe the code itself.)

> We are very close though.

Seems like the claim is that they're almost there. Anyways, have a nice day, and Happy Halloween if you celebrate.

I can't think of another thing that is called addiction yet requires this very specific wording. Can you give me another example of something that is required for life, but too much is called an addiction?

"Sentience could be algorithmically obtained" is a much weaker claim than "LLMs are almost sentient."

We've had computers that can do anything computers can do for as long as we've had computers.

"Food addiction" is perhaps not the most helpful way to describe something that every human would die without.

It just means there's no sense trying to read the docs.

One nice benefit of eating the rich, is that we're then free to use the wealth they've hoarded in ways that benefit the public.

Oh, it's AI word-vomit.

Agreed. I generally find that I spend rather more time editing the names of things and the comments in the generated code, than I do the logic.

The 'zinger' / prompt might be injected by the harness, and not under control of the model. I am pretty sure this is the case for gemini.google.com - it's glitched a few times for me in such a way that it's revealed special markup right before or surrounding that leading follow-up question.

Sure, if you want to eat refrigerated burger meal 8 times this week, you can do it for 5 minutes per meal.

They have thought about it, and the goal is for the agent to succeed at doing the scary things. So the knobs that they don't talk about, assume they are tuned to make the scary thing possible.

LLM can recite it because it's in the training data (or Google it). CS Junior could invent it independently.

Some of them are, some aren't.