Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
binarymax
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
361.
▲
by
binarymax
2y ago
I predict in 2 years there will be some aspect-ratio fill-in using genai. The thought of which terrifies me but I can definitely see it happening.
362.
▲
by
binarymax
2y ago
I was thinking the same thing, and then I also thought - if the long term plans are colonization, is sterilization and non-contamination viable, or even desirable? While the missions remain scientific I absolutely agree contamination must b
363.
▲
by
binarymax
2y ago
Maybe that’s part of it but there are so many moving parts in a near-realtime CMS and world class publishing platform, so 600 people doesn’t surprise me when you consider the scope and scale of the NYT
364.
▲
by
binarymax
2y ago
It’s impossible to answer that question without knowing what content/query domain you are embedding. Checkout MTEB leaderboard, dig into the retrieval benchmark, and look for analogous datasets.
365.
▲
by
binarymax
2y ago
It’s frustrating how hard it is now to buy pure cotton or <gasp> wool, from a store. Even if it’s 3% synthetic it’s still not what I’m looking for.
366.
▲
by
binarymax
2y ago
It’s gotta be just barely holding on in some legacy environment right now. I’m surprised they don’t just start over using new tech. Maybe that’s the end goal with on-device inference to sunset cloud Siri.
367.
▲
by
binarymax
2y ago
Sigh. We’re here as people asking other people who hopefully have expertise. If we wanted an AI summary we could get it from perplexity ourselves. I say this as someone who actually builds and sells AI research software - but there’s a ti
368.
▲
by
binarymax
2y ago
Yes, but it was obvious and I didn’t use equipment. It was in Jan 2020 at an Airbnb “hotel” in Nashville. There were two smoke detectors in the room. One of them was not like the other. I wrapped it in toilet paper. I left it wrapped s
369.
▲
by
binarymax
2y ago
I’ve been saying computering since the 90s but it never took off. Need to get a petition for OED going.
370.
▲
by
binarymax
2y ago
Computering
371.
▲
by
binarymax
2y ago
Don’t worry, a judge in Texas in the pocket of some big company will shoot this down, just like the attempt to abolish non-competes
372.
▲
by
binarymax
2y ago
I certainly understand that. But again I don't know what to make of the rankings. Yes, I know this is a niche community and yes I know there are more games than these classics...but comparison on Crokinole being #47 among thousands o
373.
▲
by
binarymax
2y ago
Eh, chess is ranked #453, go #219 and backgammon #1545. The highest ranking game is "Brass: Birmingham" which I have never heard of - so I don't know what to make of these rankings.
374.
▲
by
binarymax
2y ago
Maybe that's fine for Microsoft/Github. Where are the bootstrapped startups supposed to go?
375.
▲
by
binarymax
2y ago
Try buying a .com domain these days and let me know how that goes.
376.
▲
Ask HN: Why is the .IO gTLD uncertainty not getting press here?
10 points
by
binarymax
2y ago
|
24 comments
377.
▲
by
binarymax
2y ago
Pluntonium? In semiconductor manufacturing? Tell me more please.
378.
▲
by
binarymax
2y ago
Great project! Do you support geocoding with a nominatim api?
379.
▲
by
binarymax
2y ago
Function calling uses JSON mode. While it has been mostly correct, I do get an incorrectly formatted response sometimes (maybe 1 in 10k requests?). So it sounds like this fixes that bug.
380.
▲
by
binarymax
2y ago
I’m not entirely sure but I suspect through their ML hosting/inference/training products and services they have working with enterprises. But I’m still baffled how they cover their AWS storage and egress fees :) I remember seeing
381.
▲
by
binarymax
2y ago
I see where they’re coming from but they need to sort out the license confusion. Stack Exhange data really is the worlds best open Q&A dataset. Far cleaner and more reliable than anything else. But LLM trainers are going to use it no m
382.
▲
by
binarymax
2y ago
This is the text I received in an email just now, which is for people on the 'Starter Plan' and easier to understand than the linked article. 1. Today we’re deprecating the Billing Starter plan (your current plan) and moving all c
383.
▲
by
binarymax
2y ago
The same problems exist in lots of cities. Enforcement and recidivism prevention is not easy. You need competent staff and big budgets at all levels. For instance - what if 20 of those 200 are minors. Family court might just release the
384.
▲
by
binarymax
2y ago
Crime and homelessness aren’t areas that can be addressed quickly - unless the method is morally questionable
385.
▲
by
binarymax
2y ago
I don’t have an account and can’t read the article - but this is obvious to anyone with decent ML experience. Models are good with data they have seen. New unseen data is hit or miss. Most laypeople using AI these days could benefit grea
386.
▲
by
binarymax
2y ago
This won’t change Congress, which is more partisan than ever. It just means what we have in code now is what we are stuck with for a really really long time.
387.
▲
by
binarymax
2y ago
All the boxes must be checked. To whoever is unchecking them, please reconsider and come over to the checked side.
388.
▲
by
binarymax
2y ago
They got the Golden Arches, mine is the Golden Arcs. You see, they got the Big Mac, I got the Big Mick. We both got two all-beef patties, special sauce, lettuce, cheese, pickles and onions, but they use a sesame seed bun. My buns have no se
389.
▲
by
binarymax
2y ago
I’m not at my machine to whip up an example but this is an ideal use case for pandas. You can read or stream the excel file and split it in probably <20 lines of python.
390.
▲
by
binarymax
2y ago
"Two and a half million bucks. Spent it all on candy." A true artist. Here's my favorite work of his: https://www.youtube.com/watch?v=Psk3rmjonQA
More ›