vidarh

25 ms·

541.
by vidarh 6mo ago
The issue is that many large dams will kill a huge number of people if they fail. The Banqiao dam failure alone is the worst power plant failure in human history by several magnitudes. Not many dams have the potential to kill that many, but
542.
by vidarh 6mo ago
Kimi 2.6 is nowhere near even Sonnet in overall robustness. It can get close when everything goes perfectly. I have about 1KLOC of harness code written by Kimi to work around quirks in Kimi not needed for any other model I've tested, s
543.
by vidarh 6mo ago
Kimi and GLM 5.1 are already capable of handling a good chunk of my tasks. They about to lose the leverage to allow them to drastically increase prices - enough models are 6-12 months away from being good enough large proportions of their c
544.
by vidarh 6mo ago
Competent engineering isn't enough. You also need to never end up being in a war zone, and being able to commit to ongoing maintenance forever , or outlawing all construction far downstream (or finding the even more scarce type of loc
545.
by vidarh 6mo ago
You'd likely do less harm if you just dumped that waste in a heap on a roadside than if you shut down the plants and as a result ended up with more coal plans continuing to run. Where shutting down nuclear would result in wind or solar
546.
by vidarh 6mo ago
The key distinction being "rider-only".
547.
by vidarh 6mo ago
The vast majority of writers do not recoup their investment, not due to piracy but due to a massive glut of works available. I've published a couple of novels. They've sold far better than average, and yet not sold enough to be re
548.
by vidarh 6mo ago
The whole blue 7 thing [1] and variations is very fascinating, but we don't tend to repeatedly pick the same number in the same exact context, though. That's what made this stand out to me - I had a document where Claude had pic
549.
by vidarh 6mo ago
I have, and the dramatic change correlating with the rise of human release of co2 makes it highly improbable that it is natural. But as I said: If it somehow is natural we're far more screwed so we better hope it isn't, because if
550.
by vidarh 6mo ago
Claude, at least 4.5, not checked recently, has/had an obsession with the number 47 (or numbers containing 47). Ask it to pick a random time or number, or write prose containing numbers, and the bias was crazy. Also "something shi
551.
by vidarh 6mo ago
We better hope we're the cause of the warming, because then we conversely have a shot at slowing it or stopping it. If we are incapable of causing a change of this magnitude, then the actions we are taking to slow the change would li
552.
by vidarh 6mo ago
Yes, something prevents llms from being RLed to do this: You can't see through something opaque to determine whether there's something high calorie or low calorie out of sight. The problem itself is unsolvable given the data provi
553.
by vidarh 6mo ago
That's from Mostly Harmless, 1992, so newer than Rick Cook's version. But also sufficiently different that I have no doubt a lot of people have independently coined some variant or other. There's also the decades older (somet
554.
by vidarh 6mo ago
I'm very curious where that saying comes from now. I haven't found anything conclusive, like [1]'s friend I thought it might have been Douglas Adams, but a few references refers to it as "Grave's Law". After a
555.
by vidarh 6mo ago
My fiancees company has no developers, yet everyone has a paid subscription to LLMs. Certainly not $15/hour, and I don't think it's likely they'll ever pay that for everyone , but I don't find it hard to picture th
556.
by vidarh 6mo ago
By your numbers, it'd be $120/day per developer * 5 million = $600m per day , not per year. Of course people don't work every day, but even with European-level holidays that number is off by a factor of 240 or so.
557.
by vidarh 6mo ago
It's fairly easy to quite thoroughly "de-slop" writing: Just feed chunk by chunk to a an agent that you make compare the writing to a good piece of human writing, and adjust the writing to match. It won't address structu
558.
by vidarh 6mo ago
That's kinda the point. It's a far more remote threat.
559.
by vidarh 6mo ago
Not to my knowledge, and that's the point. Though cards could if a provider has poor opsec. But I see no particular reason to worry about that either.
560.
by vidarh 6mo ago
A creation myth of Zen meditation and Shaolin Kung Fu claims that Bodhidharma meditated for 9 years facing the temple wall, and eventually caused the wall to crack.
561.
by vidarh 6mo ago
Common enough even in tech that I've both had contracts try to demand this, and had contracts explicitly rule it out being presented as evidence of how great the company was.
562.
by vidarh 6mo ago
Unless/until there is a risk that the chips themselves are backdoored and trying to exfiltrate data, European companies that host in Europe still solve a big problem for use of certain data in Europe. It's not a purity test. Relyi
563.
by vidarh 6mo ago
I was toying with a DragonRuby game a while back, and did something like that. But DR also comes with recording reproducible playbacks, screenshotting etc. built in, couple hot reloading and easily being able to inject code into the running
564.
by vidarh 6mo ago
They're arguing that we have no evidence that humans have access to our underlying thoughts any more than the models do.
565.
by vidarh 6mo ago
We have no reason to believe it is a special case. The fact that these patients largely functioned normally when you did not create a situation preventing the hemispheres from synchronising suggests otherwise to me. There's no reason t
566.
by vidarh 6mo ago
Sperry's experiments suggests we don't have that awareness, but think we do as our brains will make up an explanation on the spot.
567.
by vidarh 6mo ago
We have pretty much evidence to support that human recollection includes the right data to be able to ascertain why we actually did something.
568.
by vidarh 6mo ago
The thing linking the two is that neither are able to accurately introspect and explain the actual reason why they made a decision. If thinking is the wrong direction to go down, then it is also the wrong direction to go down when talking a
569.
by vidarh 6mo ago
That's not how the term is being used. Six Degrees had all the features we associate with a social network, such as profiles, and the ability to create connections.
570.
by vidarh 6mo ago
> Friendster was the first social network Friendster was not the first social network. sixdegrees.com had it beat by 5 years.

More ›