daijin

[daijin @purplesquare41]

https://t.co/wbgdb7UEca

[daijin @purplesquare41]

@14a0x @lukas_m_ziegler claude says capex=$400k/8% cost of capital + $2.50 energy + $1.20 substrate + $5.50 labour + $1 packaging = $13/$14 per kg, vs aldi $12.58/kg, for button shrooms. maybe if you have one of those inputs at lower cost? solar saver coming on, reduce labour through automation?

[daijin @purplesquare41]

@14a0x @lukas_m_ziegler claude says capex=$400k/8% cost of capital + $2.50 energy + $1.20 substrate + $5.50 labour + $1 packaging = $13/$14 per kg, vs aldi $12.58/kg, for button shrooms. maybe if you have one of those inputs at lower cost? solar saver coming on, reduce labour through automation?

[daijin @purplesquare41]

@14a0x https://t.co/wW73hK32zz

1v1 me, ur agents vs mine

[daijin @purplesquare41]

@14a0x what are you building?

[daijin @purplesquare41]

@vadi_ms @boubaakiki In terms of form factor: could the transducer fit in a ring? if so, then you could just rest your chin on your hand in the thinker pose...

[daijin @purplesquare41]

@gaspihabif Hey, just found https://t.co/JxBNn66eIx run by https://t.co/x9kgibQKR8. Seems almost uncannily similar. might want to have a chat with them

[daijin @purplesquare41]

@rayzhueth Sam Altman's wealth primarily comes from his successful technology investments and entrepreneurial ventures, not directly from his role as CEO of OpenAI. His wealth is largely from his investments in Reddit, Loopt, YC

[daijin @purplesquare41]

@VictorLevoso @ESYudkowsky @tlbtlbtlb @dylanmatt c) we as humans have full visibility on real physical systems but haven't been able to make good physics simulators yet.

... ok counterargument is "we as humans have full vis on <language models> but havent been able to make <LLMs> yet". hmm

i think sim causality hard tho

[daijin @purplesquare41]

@VictorLevoso @ESYudkowsky @tlbtlbtlb @dylanmatt but there are examples in nature where offense to take power away from a diverse incumbent is hard. e.g. parasitism in nature is not a dominant strategy at high levels of complexity. e.g. centralised govt vs capitalism

[daijin @purplesquare41]

@VictorLevoso @ESYudkowsky @tlbtlbtlb @dylanmatt b) yes, defence can be easier. an AI system needs to seize additional useful power in order to pose a threat, which means taking power away from those who have power, and turns out humans dont like that. they'll restart a server to get back to what they're doing. (ik, naive, but

[daijin @purplesquare41]

@Fulbitchscholar mm yeah its a pity tho that there are so many things in our lives vying for our attention

like sleep. which i am going to do now

gbye fullbitch, see you next time, be well !

[daijin @purplesquare41]

@Kat__Woods i think the diversity of 'happiness' makes it difficult to form communities around. instead settling around instrumental goals e.g. max money is the closest you can get.

so i present to you: the FIRE movement

[daijin @purplesquare41]

@beaker406 @_holyweather beat me to it hahahaha

[daijin @purplesquare41]

@Fulbitchscholar i counter: my l'appel du vide

p.s. HIII FULLBITCH HOW ARE YOU I MISSED YOUR DINOSAUR ASS FACE

[daijin @purplesquare41]

@strangers_gate @QiaochuYuan broadly: any advocacy for change that isnt already happening
subcategories:
- that anyone can predict the future accurately enough to make good policy decisions about <problem>
- that <problem> is a positive sum game whose benefits are easily redistributable

[daijin @purplesquare41]

@dylanmatt hmm. personally i put it down to cherry picking

i.e. there exists an optimist who can back their optimism with provable stats; and a pessimist who can back their pessimism with provable stats.

the difference is which stats they choose to highlight

[daijin @purplesquare41]

@_holyweather holyweather i think you would very much enjoy factorio

[daijin @purplesquare41]

@_holyweather better/worse than that nuclear reactor powered car they tried in the 80s?

[daijin @purplesquare41]

@tszzl i once tried to form a moral philosophy around this. but was like 'is it ethical to sacrifice humans for soulless von neumann bots' and intuition said no

but there may be some alien version of us out there who say yes, and their von neumann machines maybe approaching as we speak

[daijin @purplesquare41]

@tszzl take:
it's not that people inherently crave negentropy
its just that in the long term, ppl/patterns with negentrpy stick around more and hence are more visible than ppl/patterns that don't. it's not some value system we choose or is ingrained. just game theory all the way down