AnodicElegy

4 ms·

1.
by AnodicElegy 6d ago
The most irritating way in which W is odd is that it requires three syllables to name, while all the other English letters are monosyllabic. This leads to many unfortunate situations where it takes substantially longer to say the acronym of
2.
Stepfun Step 5 Preview (LLM): On AA Pareto frontier(artificialanalysis.ai)
15 points by AnodicElegy 7d ago|2 comments
3.
by AnodicElegy 7d ago
https://artificialanalysis.ai/?models=granite-4-2-3b%2Cclaud...
4.
by AnodicElegy 8d ago
The game is broken for me. The ball just sits in the middle.
5.
Joining the dots between big AI(ft.com)
1 points by AnodicElegy 8d ago|0 comments
6.
by AnodicElegy 9d ago
The #1 thing these frontier model companies can do to help alignment is to provide the user with the chain-of-thought traces, as the open models do. But let's be real, their commercial considerations are a much higher priority than ali
7.
by AnodicElegy 10d ago
"The same kind of argument was used when China was admitted to the World Trade Organization. And indeed, lots of new jobs were created, just not in the Western world." China's entry into the WTO is really not a good evidentia
8.
by AnodicElegy 12d ago
In many areas of criminal law, willful ignorance is not an excuse. One could potentially argue that the sandbox setup at OpenAI that led to the hacking incident at Hugging Face could fall into this category. Beyond that, there's always
9.
AI changing the calculus for Canadian students interested in comp sci(theglobeandmail.com)
3 points by AnodicElegy 14d ago|1 comments
10.
How do we know if writing is AI?(paulgp.substack.com)
2 points by AnodicElegy 15d ago|1 comments
11.
by AnodicElegy 19d ago
"OpenAI’s primary bet here has been chain-of-thought monitoring. It is based on an appealingly scalable idea: a lot of the model’s capability comes from a verbalized reasoning process (chain-of-thought). If we scale optimization on the
12.
by AnodicElegy 20d ago
If someone uses an LLM to write and is able to tailor their writing such that it isn't obviously written by an LLM, then I'm fine with it! But two of my otherwise-favourite news sources -- the Hacker News front page and FT Alphavi
13.
by AnodicElegy 21d ago
This update really gives OpenAI a boost. Not saying there's anything inaccurate or untoward about that, but the timing is unfortunate. It would have looked better had it been done prior to the Fable 5.1 and GPT 6 releases. I guess AA w
14.
by AnodicElegy 21d ago
It's very different than it was earlier today, so pretty sure it's v4.2.
15.
by AnodicElegy 22d ago
Everything in moderation! With the big predators you get mercury, with the bottom-feeders you get persistent lipophilic compounds like PCBs and PFAS. I could just fish for perch, but what's the fun in that? (I'm mostly kidding, I
16.
by AnodicElegy 22d ago
Any angler will know that, in some lakes, it seems like every fish is full of worms. I've always dealt with that by frying the crap out of them -- no medium-rare bass for me, sorry. Alternatively, I'm sure they're fine after
17.
by AnodicElegy 22d ago
If you scroll down in the Artificial Analysis page you linked, you'll see all the individual benchmarks.
18.
by AnodicElegy 22d ago
Confirmed: the compound proposed by the author has been made and studied since at least 1992 ( https://pubs.acs.org/jmcmar/article-abstract/35/22/4135/7112... ). It had activity interesting enough to
19.
by AnodicElegy 23d ago
This is pure speculation, and he really doesn't know what he's talking about: "The main problem is that DXO also contains an amine, and that amine can react during the process. So the first step would be to temporarily protec
20.
by AnodicElegy 23d ago
It was in the link in the parent of the thread to which I replied. But you can find it on Artificial Analysis's website now.
21.
by AnodicElegy 23d ago
This stood out: "Artificial Analysis Intelligence Index v4.1.1 61.2" So on the Metacritic of LLM benchmarks, it's.. basically where everyone else is (except for Fable 5.1, which is a bit ahead).
22.
by AnodicElegy 25d ago
Zitron has staked his bear position and isn't budging, so regardless if he's been wrong and wrong again, he'll be remembered for calling the bubble if/when it pops, if only because so few in the media have done so withou
23.
by AnodicElegy 25d ago
Fable 5.1 is actually more expensive than 5.0 when run on the Artificial Analysis suite: https://artificialanalysis.ai/#intelligence-efficiency-tabs
24.
by AnodicElegy 25d ago
The data and graphs are great, but it would have been a much higher quality report if the text and titles were written by a human.
25.
by AnodicElegy 25d ago
Why call out "drink less" but not "smoke less"? Both are recreational drugs, both are socially stimulative, but only the former has the propensity to lead to serious harm in the short term.
26.
by AnodicElegy 29d ago
I switched two of my PCs over to Linux Mint last week. The automatic driver support is shockingly good compared to when I last installed Linux on a PC (almost 20 years ago). It even found a driver for the random no-name USB wireless receive
27.
Harvard and the Cloudflare Governance Fight(ft.com)
2 points by AnodicElegy 29d ago|1 comments
28.
by AnodicElegy 29d ago
Accessible with free FT account
29.
by AnodicElegy 1mo ago
Artificial Analysis benchmark is out: https://news.ycombinator.com/item?id=49450353
30.
by AnodicElegy 1mo ago
Impressive. It kicked everything between itself and Sol xhigh out of the Pareto frontier. Can't wait to try it out.

More ›