bomewish

10 ms·

61.
by bomewish 2y ago
Signs of a sick society. Real sad to see this stuff.
62.
by bomewish 2y ago
This exists in CeLLM and various other implementations. Well done on a smooth version. The sad thing is that when one clicks in an LLM populated cell, one cannot see/edit the content as if it were normal text. One only gets the formula
63.
by bomewish 2y ago
Certainly one could build that on top of quickwit — which also doesn’t allow crud — but it’s not trivial. You need to insert a new record with changes, then delete the record you want to update. The docs instruct that the latter action is e
64.
by bomewish 2y ago
The big issue with tantivy I've found is that it only deals with immutable data. So it can't be used for anything you want to do CRUD on. This rules out a LOT of use cases. It's a real shame imo.
65.
by bomewish 2y ago
Archive.is seems even sketchier doesn’t it? Who is paying for all that? Total mystery. I wouldn’t be surprised if it just vanished one day. I feel like the solution is some version of local personal archive + zenodo plus those others.
66.
by bomewish 2y ago
What on earth is wrong with that company? This is just so incredibly brain dead.
67.
by bomewish 2y ago
Tailscale affected?
68.
by bomewish 2y ago
I also have a custom rga-fzf function, I think adapted from that: rga-fzf () { local query=$1 local extension=$2 if [ -z "$extension" ] then RG_PREFIX="rga --files-with-matches --no-ignore" else RG_PREFIX="
69.
by bomewish 2y ago
Makes me think the govt should maybe straight up just punish obvious loophole exploitation. Clearly they were acting against the spirit of the regs for personal benefit. It seems obviously corrupt. Just punish them to deter this kind of non
70.
by bomewish 2y ago
Maybe they got some free credits? Site does not give off the right vibes either tho.
71.
by bomewish 2y ago
Maybe we could use GCP infra and the trick will work better ?
72.
by bomewish 2y ago
I just did some tests and the perf is actually worse than that — but it’s actually still perfectly fine for our use cases. - 15m row db will return a search for an extremely common term in ~3s (9m rows have that term); - When adding terms,
73.
by bomewish 2y ago
We are running FTS5 with a bunch of sqlite databases for an internal search tool, one with 15m records (we sharded that one, one shard per core), others with 200k-2m records. Speed is very fast -- the biggest dbs are hundreds of ms response
74.
by bomewish 2y ago
No one else would find these kinds of questions simply intolerable? I'm reading these thinking "seriously???"
75.
by bomewish 2y ago
You may be thinking of McFly —- it’s very good.
76.
by bomewish 2y ago
ACL issue?
77.
by bomewish 2y ago
If this is implemented properly it’ll be a game changer for collaboration on papers. Means one can write a paper with colleagues in markdown and then easily knit with pandoc/quarto. Cheaper than overleaf etc.
78.
by bomewish 2y ago
Unfortunately not! I do very similar workflows honestly but have basically defaulted to the norm or compromises like: - write own cli in python that does the sql wrapping and basic transforms and cleaning - read in the csv with R to make
79.
by bomewish 2y ago
It is complicated... but honestly I have found claude 3.5 to just 'fix it'. So you hardly have to spend much time spelunking. You just give it all your dependencies and tell it what you want. It'll whip up a working flake in
80.
by bomewish 2y ago
Thanks. I looked through the notes you posted but... honestly, it seems a bit disorganised and rather complicated. Not sure the juice is worth the squeeze for max 10% (we are getting 2-3s query time on 15m records after sharding and that se
81.
by bomewish 2y ago
Have you tested PGO for FTS in SQLite? It wasn’t clear from all those links. Would be very interested in that.
82.
by bomewish 2y ago
Are you just describing something close to Nix?? In any case, Nix solves a lot of these problems.
83.
by bomewish 2y ago
Quite interesting — but how is it fundamentally more productive than being in VS code in R or python? You don’t get any of the benefits of an IDE here. I often find myself doing very similar workflows but default to either VS Code or the sh
84.
by bomewish 2y ago
+1. More transparency around usage limit would be very helpful. Message based seems arbitrary if different messages are different lengths. Maybe token based with a bar that tracks how much you’ve used. It would go down with use and up with
85.
by bomewish 2y ago
But if all that never results in actual consequences for those responsible, what good was it? Seems like theatre to make us feel good, then.
86.
by bomewish 2y ago
What’s the threat model? If an attacker has access to your shell history then don’t they then already have access to basically everything else?
87.
by bomewish 2y ago
So I understand right, there are two solutions that would have handled this before it even got to prod or at least found it in prod fast. 1. Bunch of tests that simulate exactly the scenario of signups. Hundreds of them actually inserting d
88.
by bomewish 2y ago
Fantastic! I resorted to coloring in the light parts with a permanent marker once another library written in python stopped working to turn the damn things off. Will give this a go. Note the link to the examples folder is broken. There appe
89.
by bomewish 2y ago
OP: how many records are you storing and what’s the total size in gb? I can’t find this info on the site and am eager to get a sense of tantivy perf from a real life case.
90.
by bomewish 2y ago
Eagerly awaiting the day someone can figure out a tantivy extension to SQLite. That would be the best of all worlds…

More ›