vkazanov

28 ms·

421.
by vkazanov 4y ago
A couple of weeks ago I made an absurdly efficient arithmetic expression language compiler in python, complete with a RD/pratt parser, control flow graph IR , optimizer (dead code elimination, algebraic, etc) and a simple x86 codegen.
422.
by vkazanov 4y ago
My personal explanation is how large ML models are stateless while the growing brain is a (very) stateful thing. Stateless imitation will never be able to fully replicate a stateful system. We evolved into having these brains that learn in
423.
by vkazanov 4y ago
Chomsky's ideas around the Universal Grammar are just a theory. Similar to how his formal grammars somewhat represent real languages but never fully, the UG model will never explain it all, or even most of it. Brain biology just doesn&
424.
by vkazanov 4y ago
Yes but what difference does it make? The interface is python. Besides, libraries are much less rigid than language features so maintainers can and do iterate on these much faster. Saying this as a big Lisp/scheme/emacs fan, btw.
425.
by vkazanov 4y ago
Symbolic reasoning and related fields are just areas to were important to be researched back in the day. There are some very interesting results that came of that family of research, especially on the intersection of traditional math and CS
426.
by vkazanov 4y ago
Note then that there's an inherent conflict in programming language design space between introducing too many abstraction building tools and low-level details. It is very hard to make things truly orthogonal. E.g. memory and resource m
427.
by vkazanov 4y ago
And this is what low-level programming is by definition: technical details stand on your way to abstraction building. Memory, error handling, underlying machine details leaking, etc. Zig tries very hard to stay (relatively) simple, make thi
428.
by vkazanov 4y ago
I bought like 4-5 dell laptops with ubuntu preinstalled by now. The experience so far was somewhere between great and ok. The most important thing with these laptops is that I know it's gonna have the right hardware. At some point I st
429.
by vkazanov 4y ago
Well... complaining about people online or in the media making grandiose is like fighting wind. I totally see your point about inherent "derivativeness" of LLMs. This is true. But note how "being alive" or "being in
430.
by vkazanov 4y ago
The breakthrough of ChatGPT is not a brilliant literary work per se. It's how it interprets what people write and provides coherent answers. This was not possible previously. AlphaZero, chess algos do not have to break this barrier, th
431.
by vkazanov 4y ago
Well, I agree with you in that humans inevitable develop moral rules and codify these one way or another. Otherwise a society wouldn't be possible. Religions are one of the traditional ways of motivating these rules... Legal systems ar
432.
by vkazanov 4y ago
And astrology, and magical thinking, and many other things were always there. So what you say is that humanity is not capable of learning moral behaviour without gods? And this is why it needs deities?
433.
by vkazanov 4y ago
For me AlphaZero was boring. :-) Solution space is vast but rules are simple. It was a question of time when somebody could put things together here. There was nothing unknowable about it, unlike how natural languages were always a mystery
434.
by vkazanov 4y ago
I am not excited. I am terrified. Right at the beginning of the current wave (2010-2012) of ML approaches I did some work on ML systems and NLP, and back then I clearly saw how nothing truly outstanding is happening, we were only starting t
435.
by vkazanov 4y ago
Well, I didn't say a brain is just an LLM. It's more complicated than just this. But what LLM showed is that it works using the same simple building blocks + specialised parts of the brain: memory, constant relearning and world in
436.
by vkazanov 4y ago
Why? Do you prefer magical "handwavy" thinking? A lot of things just fall in place if you accept an obvious fact that the brain is just a powerful enough function with certain inputs and outputs. Neurons are not that complicated.
437.
by vkazanov 4y ago
The secret is that there are no grammars in our brains. Rules are statistical, not precise. Rules, idioms are fluid and... statistical. We're a bit more specialised than these new models. But that's it, really.
438.
by vkazanov 4y ago
1400 a day? For an immobilised hospital patient- maybe. It's like not moving at all! How would your body compensate for it? Given lightweight excercise (walking,kettlebells, etc) that's almost impossible!
439.
by vkazanov 4y ago
And yes, abs are not the most important thing. Depends on genetics, max weight, fat distribution. Doesn't really matter In practice. Back when i doing boxing in early 20s I was as low as 69kg, and still didn't see the full 6-pack.
440.
by vkazanov 4y ago
Congrats! That's huge! I know it's a very liberating feeling, both physically and psychologically. My collegue went from 115kg to 85kg some time ago. I know it's a different kind of game. Took a year for him. He was younger,
441.
by vkazanov 4y ago
93kg to 78kg, 180 cm. Did not look fat, but my face got round and belly was forming. Haven't seen my abs for years and years by then. Surprisingly, 80 kg was like no abs at all and 77.8kg was a 4-pack already. I wrote earlier that re
442.
by vkazanov 4y ago
I started walking on my way to work and back (London, UK). It's about 5k + 5k steps. Plus a 4k step walk after lunch. Together with a limited calorie intake (2k cal a day) it made wonders. It was crazy to watch the first 5kg just vapor
443.
by vkazanov 4y ago
Yes! Cannot stress it enough. Move. It doesn't even have to be intense! Running can be traumatic. But calories don't care if it's a walk you're doing or a run. Unless you have some fundamental food-related disorders it&#
444.
by vkazanov 4y ago
Metabolism changes is something like going down from 2100 maintenance calories to 1900 calories. A donut is /at least/ 300+ calories. Age-related metabolism changes are also not that serious. I've seen people losing 40kg (120
445.
by vkazanov 4y ago
I don't understand how counting calories can not work. It's the law of evergy conservation: energy in, energy out. This is not about water at all. Fat is just your main energy storage mechanism. And what's good about fat is t
446.
by vkazanov 4y ago
This approach his its limits. I used to run, climb, do boxing... eat whatever i felt like eating. But traumas do accumulate. 10 km a day becomes impossible. Joints, spine, heart - everything has limits. Counting calories isn't that har
447.
by vkazanov 4y ago
...and there's no official way of finding these limitations. It's amazing for a personal project but not good enough for a serious option.
448.
by vkazanov 4y ago
So is this a clean reimplementation? Given how small this is it probably is. Which means this is not Python but a language that is syntax-compatible ("Python-like"). To prove compatibility something like a (very) extensive set of
449.
by vkazanov 4y ago
Etcd is a key-value storage based on a distributed consensus algorithm. This (sort of) means that you can have a source of cluster truth in the db and make sure it lives through certain distributed system bad scenarios. Sqlite is a (relativ
450.
by vkazanov 4y ago
You're doing something VERY wrong if you don't read whatever is written on your pill's instructions!

More ›