carlossouza

10 ms·

61.
by carlossouza 3y ago
I did that (used other features). This is how new papers are ranked here: https://trendingpapers.com
62.
by carlossouza 3y ago
There's no impact in reading anything . The impact comes with what you do with what you've read.
63.
by carlossouza 3y ago
https://trendingpapers.com/ This tool can help you find what's new & relevant to read. It's updated every day (based on ArXiv). You can filter by category (Computer Vision, Machine Learning, NLP, etc), by rele
64.
by carlossouza 3y ago
These board members failed miserably in their intent. Also, they will find a hard time joining any other board from now on. They should have backed up the claims in the letter. They didn’t. This means they didn’t have how to backup their cl
65.
System 2 Attention (is something you might need too)(arxiv.org)
1 points by carlossouza 3y ago|2 comments
66.
by carlossouza 3y ago
The study introduces System 2 Attention (S2A) as a solution to the problem of incorporating irrelevant information in Transformer-based Large Language Models (LLMs). S2A improves performance on tasks involving opinion or irrelevant informat
67.
by carlossouza 3y ago
“Show me the incentive and I will show you the outcome.” Charlie Munger
68.
by carlossouza 3y ago
Aren't these guys (Business Insider) ashamed about these clickbait tactics?
69.
by carlossouza 3y ago
This is not a story about AI. It's a story about greed, vanity, and envy. Impossible to be more human than that.
70.
by carlossouza 3y ago
Of course he has. Why would he risk political capital by defending sama publicly if he didn’t know for sure he wouldn’t get burned by defending him?
71.
by carlossouza 3y ago
You are not a sample size of 1, I completely agree with you. People hate reading low-quality content created by humans, and as far as I’ve seen, AI-generated articles are even worse. Even if someone manages to use AI-generated content to tr
72.
by carlossouza 3y ago
hahaha that's why I love HN :)
73.
by carlossouza 3y ago
Great essay; impressive content. The fact that it's very unlikely for any of the current models to create something that even remotely resembles this article tells me we are very far away from AGI.
74.
Fast unfolding of communities in large networks: 15 years later(arxiv.org)
2 points by carlossouza 3y ago|0 comments
75.
by carlossouza 3y ago
https://trendingpapers.com/ Pure numbers: the top trending papers surface. They are a function of PageRank (citations and the importance of which papers cite each other), authors' previous body of work, etc... The filt
76.
Language-Driven Scene Synthesis Using Multi-Conditional Diffusion Model(lang-scene-synth.github.io)
3 points by carlossouza 3y ago|0 comments
77.
by carlossouza 3y ago
Thanks! Yeah, I posted as a Show HN, but the post went mostly unnoticed. I will implement the expansion to other areas and post it again once done, as this will be a major change in the codebase
78.
by carlossouza 3y ago
Done!
79.
by carlossouza 3y ago
I’ve built a tool using Arxiv’s API to help explore computer science papers: https://trendingpapers.com/ I hope they keep their great work especially in the API front..
80.
Show HN: Trending Papers, a PageRank-based tool to find papers worth reading(trendingpapers.com)
5 points by carlossouza 3y ago|1 comments
81.
by carlossouza 3y ago
Exactly. Race to the bottom
82.
by carlossouza 3y ago
PyMC or Pyro/NumPyro make the implementation of Bayesian regression dead simple
83.
by carlossouza 3y ago
Exactly. This story is excellent... truly inspiring! It reminds me of PG's mantra: make something people want.
84.
AI-generated books force Amazon to cap e-book publications to 3 per day(arstechnica.com)
1 points by carlossouza 3y ago|0 comments
85.
by carlossouza 3y ago
Their researchers are great, irrespective of how successful their gazillion products might be. Without their papers, there wouldn't be GPTs.
86.
by carlossouza 3y ago
I also stopped using Google, but I have replaced it with Perplexity.ai... Have you ever used Perplexity? How does Kagi compare with it?
87.
by carlossouza 3y ago
Totally agree: it's hilarious because it's true. Good startups target power users. By "good" I mean companies that find true PMF and grow exponentially at near-zero CAC (i.e. the ones that earn - not buy - their growth).
88.
by carlossouza 3y ago
> If your goal is visibility, you are well advised to tailor your messaging to the medium. That means skewing to the visual and away from the linked article. Isn't that obvious?
89.
by carlossouza 3y ago
I'm using pgvector in production, mainly in a table with 500k-1M rows. My main use case is to return search results with pagination : page 1 from 1-50, page 2 from 51-100, page 3 from 101-150, etc. (Think LIMIT and OFFSET). After a lo
90.
by carlossouza 3y ago
Renaissance is the benchmark for AI/ML applied to trading. Naturally, they are completely secret. Euclidean, the fund that manages the personal wealth of some of Renaissance's key personnel, also uses AI/ML to trading... actu

More ›