Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
topaz0
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
151.
▲
by
topaz0
9mo ago
Back of the envelope, they'll need to use this on the order of a billion times to break even, under the (laughable) assumption that running claude code uses comparable compute as the computer he's running his code on. So more like
152.
▲
by
topaz0
9mo ago
You could think of it as wanting your desire to hear the song again build up to a sufficient level to make it worth a relisten, sort of how a bus driver might want potential passengers to accumulate at a bus stop before picking them up, and
153.
▲
by
topaz0
9mo ago
Universality broadly construed is well understood since the 70s. Particular universality classes are newer and will likely continue to be discovered, but they all come to be in a qualitatively similar way.
154.
▲
by
topaz0
9mo ago
This isn't crank stuff, and operates on different kinds of problems/scales than "grand unified theory" type cranks. This is about emergent statistical order in complex interacting systems of sufficient size, not about th
155.
▲
by
topaz0
9mo ago
My sad prediction is that LLMs and humans will both get worse. Humans might get worse faster though.
156.
▲
by
topaz0
9mo ago
They also enslave human workers to do all the manual labor.
157.
▲
by
topaz0
9mo ago
They're both trust, but one is trusting a person or group of people with some intentionality.
158.
▲
by
topaz0
9mo ago
Generously, it could be that some relevant phenomenon does exist, but the studies are testing a hypothesis that is too strong or misaligned with the real phenomenon. Like, say, maybe the studies are attempting to show that people divide alo
159.
▲
by
topaz0
9mo ago
> Of course with Nvidia being a near monopoly here, they [...] will pump out cards/servers with shitty reliability rates simply because people keep buying them and they don't suffer any economic loss or have to sit in front of
160.
▲
by
topaz0
9mo ago
I'm in your camp, that both of these are appropriate for some situations. In particular I like starting with a variation on "you're absolutely right" when it appears my interlocutor has identified the wrong disagreement,
161.
▲
by
topaz0
9mo ago
It is what many saw, just not the people who take the US foreign policy establishment at its word.
162.
▲
by
topaz0
9mo ago
"...achieve prime quality texture ... add two, three, or five ..."
163.
▲
by
topaz0
9mo ago
with the occasional "much worse than useless" thrown in as a bonus
164.
▲
by
topaz0
9mo ago
Because the infrastructure required to document the deaths systematically was bombed to hell.
165.
▲
by
topaz0
9mo ago
This is what people should want from an intellectual slave, sure, and I don't think it's going to happen for llms.
166.
▲
by
topaz0
10mo ago
Wages and profits are conceptually different sorts of things, even if it's sometimes hard to draw a bright line in specific cases.
167.
▲
by
topaz0
10mo ago
Who do I have to talk to to get somebody to pay me thousands of dollars to beat a game from the 90s?
168.
▲
by
topaz0
10mo ago
Just me, or are all of these one sentence approving comments (at top level) posted by bots?
169.
▲
by
topaz0
10mo ago
And slop is one of the worst ways
170.
▲
by
topaz0
10mo ago
The ones that are moving the wrong item to the wrong place are fixable, but there will always be a problem when scrolling is quantized if there is a possibility that any item in the list is shorter than the scroll quantum.
171.
▲
by
topaz0
10mo ago
Does not work with an indexed scrollwheel -- one click of the scroll wheel moves like 4 lines in your list, which seems to break the assumptions of your code. I get very strange behaviors, e.g. "pick" Five, scroll up and down, see
172.
▲
by
topaz0
10mo ago
I got about a page in before finding out this is drivel. The final straw was "AI companies showed up and solved it in an afternoon". No faster way to show you don't know what you're talking about.
173.
▲
by
topaz0
10mo ago
And that projection is mostly based on wishes and dreams.
174.
▲
by
topaz0
10mo ago
As a start I'd say you should articulate your thoughts or ideas, rather than your self.
175.
▲
by
topaz0
10mo ago
I wonder how much of that RAM is sitting in GPUs in warehouses waiting for datacenters to be built or powered?
176.
▲
by
topaz0
11mo ago
Are they building nuclear reactors to power those?
177.
▲
by
topaz0
11mo ago
It's the reverse printing press, drowning all purposeful human communication in noise.
178.
▲
by
topaz0
11mo ago
Could be both.
179.
▲
by
topaz0
11mo ago
This is a big oversimplification. First, the 20% (of all datacenter usage being LLMs) is based on 2024 estimates, while meanwhile all of the LLM players are putting billions of dollars into building more and larger datacenters. This number
180.
▲
by
topaz0
11mo ago
I think the reason people talk past each other on this is that some of them are using LLMs for every little question they have, and others are using them only for questions that they can't trivially answer some other way. Sure, if all
More ›