vjerancrnjak

7 ms·

31.
by vjerancrnjak 9mo ago
I do not need JOINs. I do not need single row lookups or updates. I need a compute engine and efficient storage. I need fast consumers, I need good materialized views. I am not treating anything like OLTP databases, my opinion on OLTP is ev
32.
by vjerancrnjak 9mo ago
Thanks for having this discussion with me. I believe I don't want a time series database. I want to be able to invent new queries and throw them at a schema, or create materialized views to have better queries etc. I just don't fi
33.
by vjerancrnjak 9mo ago
This internal compulsion is just learned behavior. The society conditions you to work instead of play. Nothing wrong with that, I have that compulsion as well. Having a compulsion to play, purely for the sake of playing is a much healthier
34.
by vjerancrnjak 9mo ago
It’s funny how there is continuous reinvention of parsing approaches. Why isn’t there already some parser generator with vector instructions, pgo, low stack usage. Just endless rewrites of recursive descent with caching optimizations sprink
35.
by vjerancrnjak 9mo ago
These things cost money, Redshift handling live ingestion from Kinesis is tricky. There is no need for Athena, Redshift ingestion is a simple query that reads from S3. I dont want to copy 10TB of data just to have it in 1 file. And yes, def
36.
by vjerancrnjak 9mo ago
Redshift does not fit into aws ecosystem. If you use kinesis, you get up to 500 paritions with a bunch of tiny files, now I have to build a pipeline after kinesis that puts all of it into 1 s3 file, only to then import it into redshift whic
37.
by vjerancrnjak 9mo ago
What can I say, I expected more than what they actually offer. A Redshift job can fail because S3 tells it to slow down. How can I make this HA performance product slower given its whole moat is an S3 based input output interface. As a comp
38.
by vjerancrnjak 9mo ago
Both products are nothing but reliable. Redshift can’t even go around partitioning limits, or S3 limits. But what’s funny is that Claude Code is from US company so can’t be used in a boycott scenario
39.
by vjerancrnjak 9mo ago
I remember being quite surprised that my implementation which uses manual stack updates is much slower than what compiler had with recursion. Turns out, I was pushing and popping from stack on every conceptual "recursive call", bu
40.
by vjerancrnjak 9mo ago
Bruteforce thinking works in this case, given that there's only ~12*2^12 total states and transition matrix is very sparse, 1/11 is quick to calculate. But not all of these states are valid, visited set is just defined by 2 marker
41.
by vjerancrnjak 9mo ago
https://www.scylladb.com/2019/12/12/how-scylla-scaled-to-one... I like this one where they put a dataset on 80 machines only then for someone to put the same dataset on 1 Intel NUC and outperform in query tim
42.
by vjerancrnjak 9mo ago
There's a nice song by Daleka Obala - Morska Vila. That's a first ring.
43.
by vjerancrnjak 9mo ago
If you remove diacritics its completely valid BCS and same meaning.
44.
by vjerancrnjak 9mo ago
Where's the great poet or singer songwriter to capture the essence of our humanity today? All the myths have been captured and repackaged already... that lucky old sun.
45.
by vjerancrnjak 9mo ago
They market it as a general purpose store. Successfully, even though hc cs wizards wouldn’t touch it ever, c suite likes it Best example is IoT marketing, as if it can handle the load without bazillion shards, and since when does a text eng
46.
by vjerancrnjak 9mo ago
Well, you would no longer have thousands of open transactions maintaining the locks. Of course, if q on top of psql is reasonably implemented (lease), what they say makes no sense.
47.
by vjerancrnjak 9mo ago
google.com/maps to capture your location across the whole domain easier than before
48.
by vjerancrnjak 9mo ago
They probably did not batch. It’s realistic they will have issues if code is written to handle 1 job at a time and needs to make several roundtrips to the same db inside the same locking transaction. Leases exist for a reason.
49.
by vjerancrnjak 9mo ago
Not sure how that helps. They mention SKIP LOCKED but then show a job with 15 minute duration. How will you hold an open transaction for 15 minutes without seriously compromising the performance of the database? Allowing people to do this e
50.
by vjerancrnjak 9mo ago
RAG, llm pipeline industry just continues in the same fashion of throwing even more glue, insanely slow, expensive, but works due to somehow companies having money to waste, perpetually. Not that much different from the whole Apache stack o
51.
by vjerancrnjak 9mo ago
Feeling is also a thought. All perceived sensory information is thought.
52.
by vjerancrnjak 9mo ago
I miss what.cd, bandcamp almost a replacement
53.
by vjerancrnjak 9mo ago
No, LLMs like NextJS style injection as much as anyone. There is nothing magical about topological sort and calling constructors in the right order, which is all DI is. I dislike it a lot, it is exactly like any other construct that allows
54.
by vjerancrnjak 9mo ago
I dont understand that. I always assumed meditation is not a practice or an antipractice. There is no goal to reach. Nothing to gain.
55.
by vjerancrnjak 9mo ago
I am doubtful of the scientific validity and am definitively not dismissive of cultures that use meditations for inquiry into Self. Strength training is a good example because it is an immensley stressful activity with adaptations that some
56.
by vjerancrnjak 9mo ago
I find it interesting how meditation eventually becomes an anxiety reduction method, or general emotion management. What should it be if there is no burden of stress or negative impression of any emotion? Why rid of stress? It comes and goe
57.
by vjerancrnjak 9mo ago
I like how classes escape closures but then dependency injection frameworks have to work around that because now it is hard to think of construction as passing arguments to functions in the correct order. I am so glad LLMs eliminate all of
58.
by vjerancrnjak 9mo ago
Haskell is usually used as a DSL. I believe Haxl is used at Meta and TidalCycles is a good example of another DSL built in Haskell. Although I agree the usual lens, optics, machines, pipes or other higher kinded libs are completely unnecess
59.
by vjerancrnjak 9mo ago
It is interesting that despite this we still have programming languages and libraries that cannot exploit pipelining to actually demonstrate IO is the bottleneck and not CPU
60.
by vjerancrnjak 9mo ago
Reading the article I did not ask at all the question for why they're doing this. I see no reason why that is important.

More ›