emporas

20 ms·

241.
by emporas 2y ago
Hypothetical robots of the future will be made just from bamboo and very inexpensive materials. Intelligence means exactly that, being effective with minimum amount of anything, energy, muscle, sensors, materials etc. See ants, which can fi
242.
by emporas 2y ago
These "exploited" workers by ruthless capitalists, will soon produce data for robots to do their dishes, lay bricks and build houses , design automatically plumbing and electrical wiring and they will not need to work ever again,
243.
by emporas 2y ago
Bigger context windows don't work as well as advertised. The "hay in a haystack" problem is not solved yet. Also bigger context windows mean a lot more time waiting for an answer. Given the quadratic nature of context windows
244.
by emporas 2y ago
Why not use langchain-rust and make your own client? If you don't know about langchain, i think you are missing out. I took a look at other langchain implementions in js and python, in each one people have done some serious work. Langc
245.
by emporas 2y ago
Walter Russel's periodic table, is very intriguing. Observing the fact that many elements decay to other elements, with time being the only variable changing, is a very strong indication that the classic periodic table might not show t
246.
by emporas 2y ago
> You don't need infinite datasets, they just need to be large enough and cover your domain well. When you are talking about distinguishing noise from a signal, or truth from not-totally-truth, and the domain is sufficiently small,
247.
by emporas 2y ago
Somewhat of a clickbait, but "it's increasingly becoming less true" exponentially. The human population produces written data, exponentially, but in a less steep slope than LLMs by themselves. Human text may double every year
248.
by emporas 2y ago
The person you are replying to, said it clearly: "there is no such corpus of text that contains only accurate knowledge" Deep learning, learns a model of the world, and this model can be as inaccurate as it goes. Earth may as well
249.
by emporas 2y ago
>What does an algorithm look like? That's a million dollar question. Can code be represented in a visual way? Bret Victor tries to give an answer to that. I don't personally think that it can, but there is sure a way, to repres
250.
by emporas 2y ago
It's grammatically correct. Correct grammar despite of it being semantically nonsense, still not defined how small it can get. GPT-2's grammar was atrocious.
251.
by emporas 2y ago
Is there any way to find just Linux or Unix programs? Couldn't spot a way.
252.
by emporas 2y ago
The project is stalled, because at some point after i developed it, the project langchain-rust appeared and gained traction. My project is not very advanced compared to langchain, i used regexes while langchain uses tree-sitter. Langchain-r
253.
The Matrix Simulacrum Bites(pramatias.github.io)
1 points by emporas 2y ago|0 comments
254.
by emporas 2y ago
> Similar systems were proposed in the late '90s/early 2000s (hashcash/micropayments) to combat spam. These ideas indeed exist for decades. > The big problem isn't a technological one, it's that it presupposes
255.
by emporas 2y ago
The only universal way IMHO, is to associate a small cost to every internet request. The cost has to be as small as possible, but it has to be there for million/billion/trillion requests to add up the cost and make it uneconomical
256.
by emporas 2y ago
I was not referring to TeX and PDF files. Donald Knuth is the creator of Literate Programming and Tex, but I personally consider this toolchain to belong to the stone age right now. I was referring to: 1) Org mode and Emacs. LP was introduc
257.
by emporas 2y ago
There will be a comeback of literate programming. Tools were not advanced enough for it to be really practical. CS was not mature enough as a field for LP to flourish.
258.
by emporas 2y ago
Agree, something i noticed also, is that people who make PLs, like Guido van Rossum (Python), Rich Hickey (Clojure), Guy Steele (Common Lisp), Stroustrup (C++) and many others, they use natural language expertly. They express their thoughts
259.
by emporas 2y ago
All valid criticisms, i agree with all of them. Some writers are more straightforward, like Hemingway, some more verbose like Asimov. The level of verbosity is on me, that's the amount of verbosity i like, if it is not good, it is not
260.
by emporas 2y ago
In Latin, Ancient Greek and Sanskrit of course, words can be rearranged in the sentence any way we like, with the exact same identical meaning. For example "Waves hit shore" equals "Hit waves shore" equals "Hit shor
261.
by emporas 2y ago
Not a native english speaker here. My comments are never written by LLM's. I have never spoke with my mouth even one english word, but when it comes to Computer science and Biology, i can write better english than native speakers. Not
262.
by emporas 2y ago
Yes it is written by LLM. I generated 50 to 100 pages, till i finally composed it. The english version took almost 20 hours to write, LLM's output (Llama3 70b) plus human editing. Translating it was a little more than 3 hours for moder
263.
by emporas 2y ago
Infinite potential for sure. You can instruct it to be more terse, poetic, avoid pronouns when possible, dial up the humor and so on. Take also a look at David Shapiro's use of Claude and GPT for writing [1]. My process however is diff
264.
The Great Proxywall 3000(pramatias.github.io)
6 points by emporas 2y ago|9 comments
265.
by emporas 2y ago
If you are willing to go to extraordinary lengths, then yes any programming language can be written in a declarative style, at least some part of the program. Some parts of a Java or PHP program can be written declaratively. In the Nix lang
266.
by emporas 2y ago
Nix is one of the reasons open source wins in the long run. I would have never imagined that package dependencies could be installed and managed across programming language boundaries, kernel boundaries, shell boundaries, dotfile boundaries
267.
by emporas 2y ago
Writing is the simplest of all. Really trivial actually. I create new styles almost every other day. See the latest entry on my blog for a style i created last week, and this video [1]. Different styles can be created by taking advantage of
268.
by emporas 2y ago
Technologically superior tribes always needed land. Their technology was advanced, but not in a way in which land is rendered useless, i.e. land price equals to zero. Two reasons mainly for land price not equating to zero. First: buildings
269.
by emporas 2y ago
Regarding gold, i like this infographic [1], but my favorite from this channel is wolf population by country. Point being, that gold is shiny and beautiful, and it will be used even when it is not appropriate solution to the problem, just b
270.
by emporas 2y ago
As soon as mimicking voices, text messages, human faces becomes a serious problem, like this case in UK [1], then citizens will demand a solution to that problem. I don't personally know how prevalent problems like that are as of today

More ›