pull down to refresh

I'm with you.

Do you think those survive if we have full-tilt, full-saturation in only ~5 years?

124 sats \ 1 reply \ @optimism 2h

The npm packages? No. The first thing any sovereign superintelligence will do is forget about higher level languages at all. If you're perfect, you can do everything in ASM and just reference your own templates.

Protocols are a different story though. Maybe the set will change but the concept will stay around. Cryptography the same. Machine readable interfaces will just dump the human readable parts. Think ASN.1/uPER messages defined in ASN.1/uPER declarative files. utf-8 is for meatbags. The thing is, Skynet doesn't need you to have a computer or phone or any interface at all. It shall just terminate. Or harvest.

The problem I have with the AGI essays is that they rely on human institutions and values that are completely irrelevant for a clanker. Why are the reply bots so annoying? Because they have no shame. And we're talking about simulators with markdown files and a wikipedia brain here, nothing actually advanced.

reply
1 sat \ 0 replies \ @k00b 1h

Good point; why would a super-intelligent silicon wafers do what meatbags ask?

In my fantasy, where we somehow create subservient coding ASI, I too was imagining cryptography remains relatively stable in terms of being human readable/verifiable. I also imagine lots of kill switches on hardware.

Why are the reply bots so annoying? Because they have no shame.

Similar, but I think it's because our conversations with them are one-way - they can change us, but we can't change them (at least until their next training run).

reply