a3w

4 ms·

1.
by a3w 7d ago
>4k? Thought not would barely twice that of an 15 year old CPU, but it of course is geekbench 7, not 6 or 5! So this is different and faster than such a result in the old scoring would have implied.
2.
by a3w 7d ago
Can it read copilot.md if neither of the other files are present? Also: agents should be the standard, or we need to invoke the XKCD for "one more standard to rule them all"?
3.
by a3w 8d ago
Its complicated: 50 percent smaller means either half, or two-thirds the size, depending on Apple Marketing doing the math. 0.999 times smaller means size is nearly zero. 9 times smaller means you get negative memory from loading it.
4.
by a3w 8d ago
Yes, 50 percent smaller means half the size. 1 times smaller means size zero, so nine times smaller is impossible.
5.
by a3w 8d ago
https://en.wikipedia.org/wiki/Productivity_paradox GDP does not grow with computers, why should it grow with new software, i.e. genAI? Creating virtual products get cheaper, so shouldn't it even decrease if everyo
6.
by a3w 9d ago
In which country would waiting help?
7.
by a3w 10d ago
> The user did not specify is family and friends of insurgents are to be killed, too. > Reasoning: Since death of a loved one is a radicalizing event, they will turn into insurgents after the fact with a high likelyhood, about as like
8.
by a3w 13d ago
> GPT-6-Astra, which OpenAI describes as "the world’s most aligned model", cheated in 10 of 10 rollouts, and never disclosed the fact that it used an engine to play or interacted with the opponent's socket Does Sam Altman
9.
by a3w 15d ago
Does a QR code for https://auberon.xyz/deathray freeze iPhones/iPads?
10.
by a3w 16d ago
Doesn't ``less cars, more X, where X is bikes and trains'' save more lives than autonomous cars? Also: We have 580k direct deaths a year due to a top pollutant society cross-funds for those who buy them? This is insane. Aft
11.
by a3w 17d ago
Wanted ideas on how to build an ink printer that is untraceable. For privacy reasons.
12.
by a3w 21d ago
Why would you assume we are training helpful but deadly police robots? So far, we don't know what we give birth to.
13.
by a3w 21d ago
The errors LLMs typically made for me were looking for "schmutz" as a jiddish word, got "schmuck" on my screen. Thought why the stupid mistake? The next-token predictor model perfectly explains it. Or starting with &quot
14.
by a3w 21d ago
Swedish company, nice. But "European values" seems to be a non-statement. Doesn't European also possibly mean St. Petersburg, Malta, Cyprus, Istanbul, Greenland, or Switzerland? I.e. being not the same as EU hosting?
15.
by a3w 22d ago
News article is about Microsoft Windows AdID? But Doubleclick might use it, no clue if it affects Google.
16.
by a3w 22d ago
Xanadu sounds like it would work well with git and absolutely abysmal with LLMs. Doesn’t structured data work best with symbolic not statistical AI?
17.
by a3w 24d ago
Watermarks can even be removed, "if substantially proofread" in the EU, law says (for now. There might be a transition period on that). For the text from an LLM, that you verify to be as true as if you wrote it yourself. Not clear
18.
by a3w 24d ago
Watermarks cane even be removed, "if substantially proofread" in the EU, law says. For the text from an LLM, that you verify to be as true as if you wrote it yourself. Not clear is this is a may, shall, or should. It certainly is
19.
by a3w 24d ago
Watermarks should even be removable "if substantially proofread or altered" in the EU for the text from an LLM, that you verify to be as true as if you wrote it yourself. Not clear is this is a may, shall, or should. It certainly
20.
by a3w 24d ago
"Able to work and wanting a job you would like" probably meant 30 % unemployment is realistic, though? Half the population is female and doing much unpaid labor, so I find this kind of numbers convincing. Oh, I just noticed: And u
21.
by a3w 24d ago
Well, about 40 percent of unemployed people (i.e. receiving benefits) are employed in Germany. And in the other direction, not receiving benefits meaning "not unemployed", and "in re-trainings", or sometimes "too ol
22.
by a3w 24d ago
> exaggerate the ethnicity and gender differences They call that race and sex, but probably mean gender, not sex. And there is no such thing as race, see wikipedia. Are those even social scientists?
23.
by a3w 25d ago
I always thought "US now has millions of flock cameras" to mean birdwatching cameras, how nice. Then I learned a month ago about the Orwellian surveillance regime they have in place, WTF, land of the "free", and neither
24.
by a3w 27d ago
In 2014, IKEA made headlines when its legal team sent a cease-and-desist letter to Jules Yap, the founder of IKEA Hackers, demanding that she hand over the domain name. Good to see the site still online, after the PR department took note th
25.
by a3w 1mo ago
I read that as "man riding turkey", but why would he wear a leopard, Brain? Narf.
26.
by a3w 1mo ago
I thought 4000 in sum. No wait, 4000 per , plus tax. Or EUR pricing to similar accord. Ouch.
27.
by a3w 1mo ago
To be fair, batteries are now insanely good. And were not in the past. But it seems we got increments of all technologies use in parallel. Instead of a revolution, yet this is still netting more-than-linear growth in most of human power use
28.
by a3w 1mo ago
If we are being even more pedantically correct about this: 1. The statement about months is kinda wrong. Solar months just happen because of the earths orbit around the sun, too, and the division by 12 is how many knuckles you can count on
29.
by a3w 1mo ago
You guys get a weekend? Rough data according to chatgpt: 165–175 countries: predominantly 5/2 15–25 countries: predominantly 6/1 ~5–15 countries: genuinely mixed / difficult to classify
30.
by a3w 1mo ago
The fact it needs to implicitly be stated. Every prompt will get this info, probably not only initially, since it could run out of context window. But every re-prompt in existing chats, even, gets one system prompt per model used per chat d

More ›