Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
frotaur
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
31.
▲
by
frotaur
5mo ago
That would be great, but I always wonder: how do you prevent people from photographing an AI generating image through the 'legitimizing' camera?
32.
▲
by
frotaur
5mo ago
Not much info on the actual robot... For instance, I wonder how it has enough battery to follow a whale for 'months'? Which seems really unrealistic, as sperm whales can dive more than a kilometre, can't imagine an autonomous
33.
▲
by
frotaur
5mo ago
Yes! There is an excellent video on the subject, though it is in french ( https://youtu.be/lo7iJnq_U9M?si=FFz6iHI_W4lz9V8D ) He did extensive polling with different framings to see how these affect the outcome.
34.
▲
by
frotaur
6mo ago
I've been using the /ralph-loop plugin for claude code, works well to keep the model hammering at the task.
35.
▲
by
frotaur
6mo ago
Why couldn't the calendar app expose in an API the read_calendar and update_calendar functionalities, and have a skill 'use_calendar' that describes how to use the above? Then, the minimal skill descriptions are always in the
36.
▲
by
frotaur
6mo ago
Are you certain? My understanding was that this is automatically injected in the context, and in my experience that's how it worked. I never see 'ReadFile(claude.md)', and yet claude is aware of some conventions I put in ther
37.
▲
by
frotaur
7mo ago
This is my understanding as well, I thought tools where allowed.
38.
▲
by
frotaur
7mo ago
Afaik the experts are not usually very interpretable, and generally would be surprised if at least one does not change every token. I don't know what happens in practice, but I know at least during training, nothing is done to minimize
39.
▲
by
frotaur
7mo ago
Nitpick, but the volume increases cubically (it scales with volume), not exponentially.
40.
▲
by
frotaur
7mo ago
Personally defined <dtf> as 'don't touch files' in the general claude.md, with the explanation that when this is present in the query, it means to not edit anything, just answer questions. Worked pretty well up until no
41.
▲
by
frotaur
7mo ago
One important operation I've noticed in the examples that do end up with abiogenesis is having a 'copy' operation. In the bf version they use in the paper, one head can copy the byte under it at the location of the other head
42.
▲
by
frotaur
7mo ago
Seems Claude is also writing the comments for you?
43.
▲
by
frotaur
7mo ago
In a world where there isn't work schedule and in general the whole of society's schedule which works around the arbitrary time, I agree with you.
44.
▲
by
frotaur
8mo ago
I'd venture this article is written by AI with the density of 'it isn't X, it's Y'
45.
▲
by
frotaur
8mo ago
and : kessler syndrome
46.
▲
by
frotaur
8mo ago
I think the website was just down when you tried. Skills should work with most models, they are just textual instructions.
47.
▲
by
frotaur
9mo ago
I don't understand, you CAN use claude code through the API.
48.
▲
by
frotaur
9mo ago
The problem is not knowing whether someone is typing, as far as I understand. But that you may extract some information about what keys are being typed, based on the small differences in timings between them.
49.
▲
by
frotaur
9mo ago
Would love to contribute! Have some fun projects + lectures on there. www.vassi.life
50.
▲
by
frotaur
9mo ago
I'm sorry, but 'not being surprised if LLMs can rederive relativity and QM from the facts available in 1900' is a pretty scalding take. This would absolutely be very good evidence that models can actually come up with novel,
51.
▲
by
frotaur
9mo ago
I don't know much about Lean, but I attended an introductory talk at some point and I was particularly bothered by these partial function definitions. The example was sqrt, which would give 0 on the negatives. Now, of course, if you&#x
52.
▲
by
frotaur
10mo ago
This is the plane picture meme
53.
▲
by
frotaur
10mo ago
It astonishes me how people can make categorical judgements on things as hard to define as 'understanding'. I would say that, except for the observable and testable performance, what else can you say about understanding? It is a f
54.
▲
by
frotaur
11mo ago
It's crazy that the 'piss filter' of openAI image generation hasn't been fixed yet. I wonder if it's on purpose for some reason ?
55.
▲
by
frotaur
11mo ago
I mean no disrespect, but are you a trained physicist, or at least familiar with the 'mainstream material'? Because there is no shortage of 'crackpots' that have 'obvious' solutions to unsolved physics problems
56.
▲
by
frotaur
11mo ago
Thanks for the context. Nuts that this is not seen as a strictly better outcome vs a bailout.
57.
▲
by
frotaur
11mo ago
That's crazy to me...
58.
▲
by
frotaur
11mo ago
But not really right? It could happen in the market, company A chooses to bailout company B by buying it and investing money to keep it afloat. Except company A in this case is the government. No? Why is it that when it is the government do
59.
▲
by
frotaur
11mo ago
Genuine question : When companies are bailed out by the taxpayer, why can't we then give ownership of the company to the taxpayer? Effectively 'buying' the company to save it, instead of just gifting it money for it to surviv
60.
▲
by
frotaur
11mo ago
'We all benefit either way'? I am not sure that making labour obsolete, and putting the replacement in the hands of a handful of investor will result in everybody benefiting.
More ›