I’ve been thinking lately how lots of tech people dunk on humanities and liberal arts, only to come to forums like this and discuss ethics, reality, and the nature of consciousness (at quite a shallow level)
Alternative hypothesis: these people have a better framework to discuss "ethics, reality and the nature of consciousness" than people who do humanities have.
That explains the Hawking example in this thread, too.
Proposed causal explanation from me: STEM interest and education gives you skill and tools for principled, consistent reasoning, expressing nontrivial concepts (that's what mathematics is largely about), as well as appreciation and deeper understanding for how things in our actual, physical universe work, from biology and physics to computation. Conversely, humanist base with no STEM experience collapses to rhetoric, completely detached from actual reality.
You said it much better than me. Without a grounding in physical sciences, philsophies about topics like consciousness just end up in pointless circles or even absurdities. The atomism theory itself was a circular endless philosophical debate until we got sharper tools and better data for example. So a lot of problems that feel "metaphysical" today may soon fall to the grinding tank of natural science, as they always did over the past several centuries. Another thing is, people who don't know thermodynamics will be prone to thinking things like aphysical items can cause thoughts, without explaining how you are able to reconcile it with energy conservation since information requires energy. Not knowing turing computability makes people posit things like claiming quantum something and thus not a computer.
> Another thing is, people who don't know thermodynamics will be prone to thinking things like aphysical items can cause thoughts, without explaining how you are able to reconcile it with energy conservation since information requires energy.
Yes. This is one of the main key insights that hasn't percolated to the humanities just yet. It's even not fully appreciated in STEM. Computer science is not a pure branch of mathematics anymore, because computation is an actual thermodynamical phenomenon, with identified and measured energy requirements. CS is part of physics now, and not in a trivial sense of "computers are physical devices", but a fundamental one - "computation itself is a physical process".
> The atomism theory itself was a circular endless philosophical debate until we got sharper tools and better data for example.
(EDIT: matched GP's typo fix in quote)
Exactly, and same is true about everything from medicine or sociology - better tools allowed us to see answers to questions that previously were just discussed in abstract by philosophers.
I'm convinced cognitive sciences and sociology are about to see this happening to them, thanks to LLMs. It doesn't matter whether an LLM "truly thinks" or is just a facsimile - it's similar enough behaviorally to be interesting, and unlike a human mind, LLM "intelligence" is a computational artifacts in the engineering sense, reified as a big bag of floats configuring a structurally simple circuit. We can study that directly - slice nad and dice and prod, and what's more, we can study how they're forming - how the training turns the values from completely random to something that makes the circuit understand natural language and perform feats of general-purpose intelligence.
I bet it's gonna solve a lot of philosophical questions too, by allowing us to observe how the thing works in practice.
(Though it might still take decades for those observations to sink in; I mean LLMs are already direct empirical evidence disproving various ideas of Chomsky and Searle, yet I still see people talking about symbolic semantics and Chinese room in earnest every other day on HN...)
Yes atomism felt equally inresolvable back then, but after hard data and better tooling suddenly it is a scientific and physical problem. Consciousness will succumb too most probably. But of course there is the usual difficulty of religious people and others with hidden supernatural beliefs.
moolcool · · focus · HN ↗
sseagull · · focus · HN ↗
“Engineer’s disease” and all that.
TeMPOraL · · focus · HN ↗
That explains the Hawking example in this thread, too.
Proposed causal explanation from me: STEM interest and education gives you skill and tools for principled, consistent reasoning, expressing nontrivial concepts (that's what mathematics is largely about), as well as appreciation and deeper understanding for how things in our actual, physical universe work, from biology and physics to computation. Conversely, humanist base with no STEM experience collapses to rhetoric, completely detached from actual reality.
hardbass · · focus · HN ↗
TeMPOraL · · focus · HN ↗
Yes. This is one of the main key insights that hasn't percolated to the humanities just yet. It's even not fully appreciated in STEM. Computer science is not a pure branch of mathematics anymore, because computation is an actual thermodynamical phenomenon, with identified and measured energy requirements. CS is part of physics now, and not in a trivial sense of "computers are physical devices", but a fundamental one - "computation itself is a physical process".
> The atomism theory itself was a circular endless philosophical debate until we got sharper tools and better data for example.
(EDIT: matched GP's typo fix in quote)
Exactly, and same is true about everything from medicine or sociology - better tools allowed us to see answers to questions that previously were just discussed in abstract by philosophers.
I'm convinced cognitive sciences and sociology are about to see this happening to them, thanks to LLMs. It doesn't matter whether an LLM "truly thinks" or is just a facsimile - it's similar enough behaviorally to be interesting, and unlike a human mind, LLM "intelligence" is a computational artifacts in the engineering sense, reified as a big bag of floats configuring a structurally simple circuit. We can study that directly - slice nad and dice and prod, and what's more, we can study how they're forming - how the training turns the values from completely random to something that makes the circuit understand natural language and perform feats of general-purpose intelligence.
I bet it's gonna solve a lot of philosophical questions too, by allowing us to observe how the thing works in practice.
(Though it might still take decades for those observations to sink in; I mean LLMs are already direct empirical evidence disproving various ideas of Chomsky and Searle, yet I still see people talking about symbolic semantics and Chinese room in earnest every other day on HN...)
hardbass · · focus · HN ↗
>The atomism theory itself was a circular endless philosophical debate until we got sharper tools and better data for example.
You may edit your quote if you want
hardbass · · focus · HN ↗