HN user
What harness do they use for testing?
Back in 2025 it was common to test models in a different harnesses.
I remember watching a guy on youtube, who was testing every new model in opencode, cline, codex, claude, etc.
Why did it come out of fashion ?
EDIT: ah, yeah. the point was that a harness would often affect results (task completion rate, I think) for more than 10%
note that ARC-AGI-3 has its rules changed.
before today all the contestants were capped at $10k
Model API is not available in your region.
:(
Well, Vietnam is not in the list of restricted territories.
Anyway, what is "your region" ?
Is this where I am now, or is it where I activated my Oculus 2 five years ago ?
This billing cycle I was billed $20 three times.
I contacted my bank and got a reply (from a human) that all three payments are valid.
Emails from Anthropic state that the first two payments failed, but the third went through. Fin says that my question will be elevated to a human being, but so far I was not contacted.
This makes a good benchmark LLMs:
``` look at this paper: https://arxiv.org/pdf/2603.21852
now please produce 2x+y as a composition on EMLs ```
Opus(paid) - claimed that "2" is circular. Once I told it that ChatGPT have already done this, finished successfully.
ChatGPT(free) - did it from the first try.
Grok - produced estimation of the depth of the formula.
Gemini - success
Deepseek - Assumed some pre-existing knowledge on what EML is. Unable to fetch the pdf from the link, unable to consume pdf from "Attach file"
Kimi - produced long output, stopped and asked to upgrade
GLM - looks ok
spatial or not, arc-agi is the only test that correlates to my impression with my coding requests
why don't they publish at ARC-AGI ? too expensive?
I've created Asterisk Codex Skill, but turns out there is ten seconds timeout for scripts
Just in case the author is here: what's the FPS?
Too late. Browser-use local LLMs are already a thing
I think the proper way of estimating the cost is the cost of entire run of a test. Like in aider's leaderboard.
what's point of comparing token prices? especially for thinking models.
Just now I was testing the new Qwen3-thinking model. I've run the same prompt five times. The costs I got, sorted: 0.0143, 0.0288, 0.0321, 0.0389, 0.048 . And this is for single model.
Also, in my experience, sonnet-4 is cheaper than gemini-2.5-pro, despite token costs being higher.
Fizen
I use crypto often. I am Russian, I left Russia when Putin started the war. For me, it is quite hard to open a bank account. So I use crypto. I work remotely for a Singaporian company. Now I'm in Vietnam, I can pay for my groceries with crypto using QR code. I can cash USDT crypto with a rate better than paper bills.
I have two bank accounts in Kazakhstan. Both card credentials were stolen after I used a popular hotel booking website, which, by the words of reddit, shares my card details with hotels. Some money was stolen. Seems like 3D-security only affects my payments, and theifs have a freedom to choose a website without 3D. Now I have to keep that cards always locked. Unlocking them for a short moments, when I need to make a card payment. Like booking an hotel, or buying an airline ticket.
lol. That is too complex for the purpose of long-range movement in VR. Just make a controller with a control patterns of electric unicycle, or maybe "hoverboard".
Is this material safe to handle?
It feels like this thing soon will start appearing all over ebay and aliexpress
and how about playstation 5 ?
and also xbox and that thing from valve?
Another example. Happened to me today. I was unable to pay online for a bus ticket until I set up VPN out of Cambodia to SG. I tried two cards by two different banks on three different ticketing sites. Without VPN all my payment attempts were rejected.
None of above probably. Just buy a laptop which is known not to deviate from reference designs (from intel)
Also, it shows proper dates and times for messages. Not that stupid «1 day ago»
just curious. Can USB3 work without D+ and D- ?
but how will it affect ML applications?
How can we replace nested comments?
you may use multiline string literals
There are two vaccines on the final stages of development in Russia. One from Gamaleya center (Moscow), another from «VECTOR» facility (Novosibirsk). From my (layman's) point of view, Gamalaeya's vaccine is somewhat hurried in development. Also, Gamaleya is known for its lobbying power. In the past, it approved and popularized a number of controversial drugs like Kagocel and a few of «-feron» drugs. VECTOR's vaccine, on other hand, looks more trustworthy.
yes :) RDRAND is implemented differently
Isn't this usecase patented by Apple?
Cool, but we haven't seen compiling benchmarks yet. 12 cores might be bottlenecked by dual-channel memory.
that looks like 2GIS maps
EDIT: check out this map of Dubai: https://2gis.ae/dubai?queryState=center%2F55.274674%2C25.197...
Also check out its search UX. Try searching for "grocery"