Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
pizza
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
331.
▲
by
pizza
2y ago
Had to whittle the title down a bit; “Cold acclimation with shivering improves metabolic health in adults with overweight or obesity”
332.
▲
Cold acclimation with shivering improves metabolic health in obese adults
(nature.com)
3 points
by
pizza
2y ago
|
1 comments
333.
▲
GPS constellation as a way to image the ionosphere (2011)
(wondering-star.blogspot.com)
29 points
by
pizza
2y ago
|
4 comments
334.
▲
by
pizza
2y ago
But actually whenever this happens you get rich signal about the hidden didactic assumptions of your method of prompting it about things you yourself are unsure you can verify for yourself, and also of how you thought the tool worked. This
335.
▲
by
pizza
2y ago
Yes, anthropic just released model context protocol and mcp is perfect for this kind of thing. I actually wrote an mcp server for claude to call out to openai just yesterday.
336.
▲
by
pizza
2y ago
This seems like the kind of thing that laws and regulators exist for.
337.
▲
What if muscle tension is sometimes signal jamming?
(chrislakin.blog)
5 points
by
pizza
2y ago
|
0 comments
338.
▲
by
pizza
2y ago
You're right - at my coffee shop a cup of coffee is nine
339.
▲
Neural simulation-based inference for parameter estimation in ATLAS
(arxiv.org)
1 points
by
pizza
2y ago
|
0 comments
340.
▲
by
pizza
2y ago
The thing about questions like this is that the complexity of mathematical explanations is highly dependent on what each reader considers "simple." Consider two different approaches to understanding a concept, expressed in informa
341.
▲
by
pizza
2y ago
I'm not sure what's worse: giving parental responsibility to politicians or equating craven paternalism to benevolent parenting.
342.
▲
by
pizza
2y ago
This is awesome. Would it be able to integrate Pocket bookmarks into this somehow? Would love to be able to keep a cache offline of all the things I wanted to read (esp if I can prevent link-rot), and to be able to query them.
343.
▲
llama2psp: llama2.c running on the PSP
(github.com)
2 points
by
pizza
2y ago
|
0 comments
344.
▲
by
pizza
2y ago
I do
345.
▲
by
pizza
2y ago
o1: - better for when the response has to address many subgoals coherently - usually will not undo previous bugfix progress that was made earlier in the conversation, whereas with Claude if you start having extremely long conversations I ha
346.
▲
by
pizza
2y ago
I think OpenAI spec function calls are to this like what raw bytes are to unix file descriptors
347.
▲
by
pizza
2y ago
Yes, you can think of it in terms of (WLOG think of any uniquely-decodable code) prefix-free codes. They're uniquely decodable - for things that are not uniquely decodable, that implies that you could put overlapping codes over that sy
348.
▲
by
pizza
2y ago
"If it were a fact, it wouldn't be called intelligence." I think the other way to read it is that you're fundamentally going to have to choose (by your freedom to control your algorithm's design) which of the maps f
349.
▲
The structure of the token space for large language models
(arxiv.org)
2 points
by
pizza
2y ago
|
0 comments
350.
▲
by
pizza
2y ago
I have a strong hunch that music has a profound emotional effect because we can use the same kind of signal decomposition to encode both emotionality and musicality. Of course, I have no evidence for my claim.
351.
▲
TensorDict
(github.com)
2 points
by
pizza
2y ago
|
0 comments
352.
▲
by
pizza
2y ago
Is there a meta-library someone somewhere has already written for when I just want to write 20% of a compiler, and it more or less takes care for me 80% of the common compiler-building-related things I’m likely to need to do?
353.
▲
How to Reverse Engineer a Proprietary File Format: A Brief Guide
(apriorit.com)
3 points
by
pizza
2y ago
|
0 comments
354.
▲
by
pizza
2y ago
Features from other languages I wish were Python builtins: - Haskell algebraic data types + syntactic sugar - C/Lisp macros
355.
▲
by
pizza
2y ago
I would prefer for my work value to be multiplied by piracy, personally.
356.
▲
Iterative deepening depth-first search
(en.wikipedia.org)
1 points
by
pizza
2y ago
|
0 comments
357.
▲
Copilot Arena
(marketplace.visualstudio.com)
1 points
by
pizza
2y ago
|
0 comments
358.
▲
Accurately computing the log-sum-exp and softmax functions
(academic.oup.com)
1 points
by
pizza
2y ago
|
0 comments
359.
▲
by
pizza
2y ago
But there's really nothing about chess that makes reasoning a prerequisite, a win is a win as long as it's a win. This is kind of a semantics game: it's a question of whether the degree of skill people observe in an LLM playi
360.
▲
by
pizza
2y ago
I think I might have heard so from Nicolas Bourbaki
More ›