js8

29 ms·

541.
by js8 1y ago
I would not be so sure. There are many fields where reductionism was applied in practice and it yielded useful results, thanks to computers. Examples that come to mind: statistical modelling (reduction to nonparametric models), protein fold
542.
by js8 1y ago
> The people involved all seem very... Full of themselves ? Kinda like Mensa?
543.
by js8 1y ago
I know a nice proof of volume of tetrahedron being 1/3 of the corresponding paralellepiped. You split it into smaller tetrahedra by midpoints and count them. Also there is a nice visual proof that in an equilateral triangle, for every
544.
by js8 1y ago
Sorry, but that's reductionism. We don't know how human brain works, and that you won't get there by studying quantum electrodynamics. LLMs are insanely complex systems and their emergent behavior is not explained by the algo
545.
by js8 1y ago
Playing a devil's advocate - if you have to scrap by with little memory and few services, you focus on your mission, not distractions like code elegance, library choices and build systems.
546.
by js8 1y ago
But look the GDP has increased!
547.
by js8 1y ago
> All of modern mathematics is conditional. I am not sure this is true for "all" mathematics. You're still using some metalogical axioms that are always true. In particular, laws of untyped lambda calculus (or any other mo
548.
by js8 1y ago
> They're intended to stabilize volatile situations and prevent escalation. Wouldn't it be better, in many of these cases, to actually talk it out? When I was young and there was a conflict of this sort, we were taught to apolo
549.
by js8 1y ago
Hi, I like your work. Do you think there could be a universal combinator (like iota) and some binary encoding that would produce even smaller interpreters than BLC? I read your paper and you show that a simple encoding of S and K combinator
550.
by js8 1y ago
I see LLMs (AI if you will) as both regression and progression (and that's why I think people are so divided on it). It's progression for obvious reasons - it automates natural language processing (human language), which even 5 ye
551.
by js8 1y ago
Maybe you will, but maybe not. Hence the title - 2.5 attempts sounds about right.
552.
by js8 1y ago
Somewhat agree, but it depends. I agree that core of the operating system (kernel) should be responsive, but there are many tasks in the operating system that don't require immediate response and can be run in batch. Usually these are
553.
by js8 1y ago
I get what you're saying but I disagree. You're talking about people who ask others to go to war, not people who actually decide to be violent. Of course some people get bored being well-off by indulging in violence but most peopl
554.
by js8 1y ago
"It is what made us "civilized"." Not true, it was technology what civilized us (and jury is still out), not growth. Yes, if people are not hungry or sick, spend time doing meaningful work, can talk to others and general
555.
by js8 1y ago
I agree with what you wrote, except possibly: "We're moving towards better representation of common design elements." I would say big component (maybe 90%) of the new abstractions created is fashion. So the representation is
556.
by js8 1y ago
I don't see the problem. So medical and nursing care is expensive. We can change the society so that relatively more people work in those fields. Instead, we are cutting on those services. It's just a thinly disguised attack on th
557.
by js8 1y ago
"Mandatory pensions are not a vacation sponsored by the government" Who are you talking for? I think plenty of people would prefer it to be that. We should respect our elderly, as a society.
558.
by js8 1y ago
I agree with you. I think the fundamental problem is we don't have a good unified theory of fuzzy reasoning. We have a lot of different formal approaches but they all have flaws. Now LLMs made a big breakthrough that they showed we can
559.
by js8 1y ago
I don't see how the two things are related. Whether acquisition of human language is nature or nurture - it is still learning of some sort. Yes, maybe we can reproduce that learning process in LLMs, but that doesn't mean the LLMs
560.
by js8 1y ago
No, it's skepticism. I think proponents of that hypothesis should come up with some testable prediction, how is that supposed "loss of quality" affecting the distribution of possible LLM outputs. Then we can have debate on it
561.
by js8 1y ago
> Maybe someone would like to contest that. But that should be done directly. Instead of making pedestrian statements I am not sure what you want from me here. Yes, I do contest that "that a lot of LLM brilliance comes from datamini
562.
by js8 1y ago
Unfortunately many languages have poor support for named type definitions and higher order types, unlike e.g. Haskell. That would definitely help to avoid these problems.
563.
by js8 1y ago
What about a wearable split 34-key keyboard? That could work, for outdoor coding. Or voice interface, there was a video of a guy coding in Emacs using that.
564.
by js8 1y ago
It's not the same argument because I don't make any assumptions about how LLMs work. All I am saying that people have been able to keep reality in check in the presence of cultural artifacts, and will continue to do so even if suc
565.
by js8 1y ago
Another example is free market ideology. This was a question I posed to libertarians - how can you claim that free market enhances human freedom, when it always tells you what to do in the name of efficiency?
566.
by js8 1y ago
Of course AI has a way to judge truth -it's what we tell it to. We say to it, forests are real, but dragons are not. If it didn't discern it, it would lose competitivness with other AIs, the same way delusional humans are shunned
567.
by js8 1y ago
Has the quality of art gone down since art was invented? Or has the quality of the written text gone down since writing was invented? I think the answer is clear no. Humans have been trained on "human-generated data" (cultural art
568.
by js8 1y ago
That's exactly how I meant it. Another alternative to Java at the time was Smalltalk, with different syntax. Choosing C++ syntax for Java was clearly a marketing move (as it was for C++) - we shouldn't underestimate network effect
569.
by js8 1y ago
I believe there were alternatives at the time, but expensive. Java was free and also syntactically compatible with C++. I think that rather these properties (than being somewhat unique) are the main causes of its success.
570.
by js8 1y ago
I think any HDL that is more inspired by functional languages (with better composability) is good. But yeah there is lot of inertia in using existing tools.

More ›