montebicyclelo

11 ms·

91.
by montebicyclelo 2y ago
The core concept of agent graphs seems quite similar to an LLM agent framework I created last year, [1]. Imo it is a more promising way to make agents, than the prior approach of just giving an LLM N tools to use at once. [1] https:/&
92.
by montebicyclelo 2y ago
Their 2d music programming language, Orca is very cool/fun. Some examples: https://youtube.com/watch?v=MNF8SF69QvM&list=PLb1uDATFJPcEEG... https://youtube.com/watch?v=DBI8eBMGyYs https:/&#x2F
93.
Training a Simple Transformer Neural Net on Conway's Game of Life(sidsite.com)
1 points by montebicyclelo 2y ago|0 comments
94.
Training a Simple Transformer Neural Net on Conway's Game of Life(sidsite.com)
2 points by montebicyclelo 2y ago|0 comments
95.
by montebicyclelo 2y ago
People throw that proof around all the time; but all it does is show that a neural net is equivalent to a lookup table; and a lookup table with enough memory can approximate any function. It's miles away from explaining how real world,
96.
by montebicyclelo 2y ago
The author has experience of teaching box plots in various organisations. The author has found that compared to other types of plots, people struggle to learn how to intepret box plots. The author proposes some alternatives that they believ
97.
by montebicyclelo 2y ago
Related side note: I've always had a suspicion that one of the ideas I was working on using Google Colab was viewed by an employee and leaked, because someone wrote a blog post with the exact same idea (very niche) before I got round t
98.
by montebicyclelo 2y ago
Yep interesting that massively useful and impactful work like this doesn't seem to get as much interest on HN compared to gimicky new architectures that have only been shown work well only on tiny datasets like MNIST and CIFAR, rather
99.
by montebicyclelo 2y ago
IAgreeThatSnakeCaseIsEasierToRead,ThoughIfWeWereToUseCamelCaseMoreWeWouldBeSavingAGoodAmountOfSpace,WhichMightBeUsefulForNarrowBlocksOfText.I'mSuprisedItHasn'tCaughtOnInProse.Actually,LookingAtThis,MaybeIt'sAGoodThingItHasn&#
100.
by montebicyclelo 2y ago
This seems likely. I wonder if explaining it through the packaging, e.g. coloured stripe marked "stir space" and an explanation on the back saying don't worry you are only paying per gram of peanut butter...
101.
by montebicyclelo 2y ago
The introduction / training could be made clearer with some GIFs / animations showing what to press and the effects.
102.
by montebicyclelo 2y ago
y = x @ W + b Is how you'd write it in NumPy and most (/all?) deep learning libraries, with x.shape=[batch, ndim] I'd personally prefer that to be the convention, for consistency.
103.
by montebicyclelo 2y ago
From the YC guidelines: > One good trick for describing a project concisely is to explain it as a variant of something the audience already knows. It’s like Wikipedia, but within an organization. It’s like an answering service, but for e
104.
by montebicyclelo 2y ago
The success we're seeing with neural networks is tightly coupled with the ability to scale - the algorithm itself works at scale (more layers), but it also scales well with hardware, (neural nets mostly consist of matrix multiplication
105.
by montebicyclelo 2y ago
It does seem to be the norm for interesting things to be presented in a boring way, rather than the exception. I'd guess this is partly because it's difficult to explain concepts clearly in the first place, let alone make the expl
106.
by montebicyclelo 3y ago
It's fairly easy to use, not that compute intensive (e.g. can run on even a small-ish CPU VM), the embeddings tend to perform well and you can avoid sending your data to a third party. Also, there are models fine tuned for particular d
107.
by montebicyclelo 3y ago
Cosine-similarity is a useful metric. The cases where it is useful are models that have been trained specifically to produce a meaningful cosine distance, (e.g. OpenAI's CLIP [1], Sentence Tranformers [2]) - but these are the types of
108.
by montebicyclelo 3y ago
Hmm, typically the models where people use cosine similarity on embeddings have been deliberately trained such that the cosine similarity is meaningful. It looks like this paper is looking at examples where the models have not been delibera
109.
by montebicyclelo 3y ago
Nice, here's an attempt to compute the HN-index for all users. (Differs from blog post; not sure if because the data from the two sources is different, or there's a mistake somewhere): WITH Ranking AS ( SELECT
110.
by montebicyclelo 3y ago
It appears to be working for me: SELECT count(), min(time), max(time) FROM hackernews_history 39838828, 1970-01-01 00:00:00, 2024-03-11 13:04:47
111.
by montebicyclelo 3y ago
The open Google Bigquery table of Hacker News might possibly be a nice way to get this data. (Although last time I checked it was a few years out of date, so maybe not... Hoping that someone reads this and starts updating it again...)
112.
by montebicyclelo 3y ago
Lineage: - Peter Norvig (original (?)) - http://norvig.com/21-days.html#answers - Jeff Dean (slides) - https://www.cs.cornell.edu/projects/ladis2009/talks/dean-key... - Colin Scott - https:&
113.
by montebicyclelo 3y ago
Thanks, I do agree with this, and I am rooting for them. Would love to find out at some point that it's turned around.
114.
by montebicyclelo 3y ago
Guessing there are a few of us in a position where we are frustrated with our past experiences with ROCm software, (e.g. being awful to install, in the past there were long guides with loads of steps to follow, and not the greatest clarity
115.
Show HN: Wikipedia pages as poems, written by LLM(enwikipoems.com)
1 points by montebicyclelo 3y ago|0 comments
116.
by montebicyclelo 3y ago
This addresses a real pain point, of Numba not being so easy to profile, (especially compared to Python and line-profiler), so will definitely be trying this out!
117.
by montebicyclelo 3y ago
There's an example here of using Numba to speed up data science code, (disclaimer, I'm the author): https://sidsite.com/posts/python-corrset-optimization/ > for most "normal" uses, you'
118.
by montebicyclelo 3y ago
Consumer laptops, such as apple silicon macbooks with just 32gb ram, can run fairly large models, such as Mixtral-8x7b, an open source model that's comparable to gpt-3.5, well enough for interactive chat, (and seems likely models of th
119.
Extracting Copyrighted Text from GPT(sidsite.com)
3 points by montebicyclelo 3y ago|1 comments
120.
by montebicyclelo 3y ago
Yes, I saw that and still consider the methodology questionable. A fairer approach might be: time from the cli, including compilation, for compiled languages. (Or warmup the jit compiled code.)

More ›