rao-v

5 ms·

31.
by rao-v 1mo ago
Nostalgic for when Dwarf Fortress was 2D and had an ending (if you dug too deep a catastrophic monster emerged as I recall)
32.
by rao-v 2mo ago
A frontier LLM based query rewriter has atleast a masters in economics, and a pretty good understanding of finance informal language. How long ago did you try this? I'd be curious if you find this still to be the case.
33.
by rao-v 2mo ago
Oh I’m sorry to bring such a presentist idea to such a cool project, but if we ever get LLMs inferencing cheaply on consumer hardware and capable of efficient continuous learning, this insane format might be the perfect way to share your un
34.
by rao-v 2mo ago
This is deranged, and perilously close to dumb, which makes it one of the best things I’ve seen on hacker news this year. Absolutely wonderful stuff.
35.
by rao-v 2mo ago
I like the approach here. It’s cleverer than just “dump context to LLM”. I couldn’t quite figure out if it genuinely runs purely locally though. If nothing else - this is how NPCs should work in games moving forward!
36.
by rao-v 2mo ago
Esp32s with eink displays are pretty much this
37.
by rao-v 2mo ago
The article suggests it's unreasonable numbers in the original Uzbekistan data source and that other datapoints may have been worse, the authors just didn't correctly execute their basic checks. "It turns out that Uzbekistan
38.
by rao-v 2mo ago
What is the right recommendation at this point for a Node alternative? Deno?
39.
by rao-v 2mo ago
Even Muse Glimmer (as did GPT-OSS I think) does ~4 sliding window attention layers + 1 full attention layer (like Gemma 4). I’m assuming both labs have good reason to think that gated delta nets are not optimal. Of course it’s possible the
40.
by rao-v 2mo ago
I run llama.cpp and specialized forks on 64GB of HBM and I still cannot figure out where to find the final correct guidance on using the Gemma 4 models. Would appreciate any kind of pointer to the latest!
41.
by rao-v 2mo ago
Read this as "Male humans can't tell the difference." and found it comic
42.
by rao-v 2mo ago
Ha!
43.
by rao-v 2mo ago
I salute your honesty! But [he hastens to add, looking furtively around] I also frown disapprovingly at your choices
44.
by rao-v 2mo ago
Genuine curiosity - is the pop up vaguely factual or sort of randomly generated?
45.
by rao-v 2mo ago
Native dflash support on day 1 helps a lot! High quality speculative decoding speeds up a lot of agentic work.
46.
by rao-v 2mo ago
I have at times mused that writing the internal state of pandoc to disk (yes 13th standard etc) would be the most interoperable file format
47.
by rao-v 3mo ago
Wait it only applies in 14 states? Why do people act like it’s true of the entire country?
48.
by rao-v 3mo ago
I don't quite understand the focus on the token cost of this rewrite. Obviously, we should examine if this rewrite is good, effective, good for the product etc., but the token cost seems ... not important? The marketing value of this t
49.
by rao-v 3mo ago
This is a really neat use of the per-layer embedding trick. It's also worth noting that there viable TTS models that are ~20-30M param, so it might mean you can have a ESP32 with no network access read stuff out to you in near real tim
50.
by rao-v 3mo ago
Food inflation has been moderately high but egg prices were a mix of weird short term events and umm price fixing (look up the recent case). They are back to pre 2022 levels now! https://fred.stlouisfed.org/series/APU00
51.
by rao-v 3mo ago
Good to know!
52.
by rao-v 3mo ago
A genuinely funny and good point!
53.
by rao-v 3mo ago
It seems like such a tempting gap though. The sort of thing you’d think in 2015 would be an obvious capability of 2026 languages!
54.
by rao-v 3mo ago
It’s not unreasonable to suggest that text editors need a nice atomic move (but personally I don’t see the need for it)
55.
by rao-v 3mo ago
It distresses me that we don’t have a language that can do a best effort parallelization of arbitrary loop like code across SIMD, multiple threads, multiple cores and GPU with a small directive. I don’t need it to be optimal, just … handy a
56.
by rao-v 3mo ago
I mean you’d probably just generate random low descriptive length f(x,y,z), check for a const determinant then poke for invertibility. Be fun to ask Fable to write a search program to find more counter examples using only early grad theory
57.
by rao-v 3mo ago
This is so unreasonable! As @__alpoge__ himself notes this is classic crank graveyard territory and yet the counter example is something a grad student in 1997 could have found w a ~3 day computer search. Wild!
58.
by rao-v 3mo ago
The world needs more delightful websites that do one clever thing well like this. Turns out I need to read Percival Everett!
59.
by rao-v 3mo ago
Pretty sure the labs subsidized the well known benchmarks w free tokens for evals
60.
by rao-v 3mo ago
Codex legit does a great job of this with it's auto-review feature. Does a good job of understanding what permissions I'm implicitly giving with the request, and where it needs to actually ask me to grant permission.

More ›