Mastodawn

Tommi Mäklin Jul 15, 2023

This vid about optimizing Mario 64 source code just for the fun of it is coincidentally also a pretty neat example of why/when function inlining and loop unrolling are useful https://youtube.com/watch?v=t_rzYnXEQlE

FIXING the ENTIRE SM64 Source Code (INSANE N64 performance)

Could this bring Luigi back?Patreon: https://www.patreon.com/Kazestuff🎥 https://www.twitch.tv/kazesm64🐦 https://twitter.com/KazeEmanuarMERCH: https://kazem...

YouTube

Tommi Mäklin Jul 13, 2023

Vidar Hokstad / Galaxy Bound Jul 13, 2023

Beating neural approaches to sentence classification for out-of-domain datasets with ... nearest neighbour over *gzip* (finding the ratio between the compressed size of the documents combined vs. smallest compressed size of the inputs being compared).

This is both hilariously simple, and somewhat damning of how the other methods do in this context.

https://aclanthology.org/2023.findings-acl.426/

#computerscience #programming #machinelearning

“Low-Resource” Text Classification: A Parameter-Free Classification Method with Compressors

Zhiying Jiang, Matthew Yang, Mikhail Tsirlin, Raphael Tang, Yiqin Dai, Jimmy Lin. Findings of the Association for Computational Linguistics: ACL 2023. 2023.

ACL Anthology

Tommi Mäklin Jul 13, 2023

Luca Visentin Jul 11, 2023

Hey #bioinformatics , #OpenScience and #computerscience hive minds of mastodon, I have need of your help.

See https://genomic.social/@MrHedmad/110606060559853537

I need a way to make our lab data FAIR without having to explain or use anything very complex. RO-crate does not cut it, for our personal use-case. So I came up with https://github.com/MrHedmad/data-myr

Can anyone take a look and tell me if:
- It is dumb / useless / makes no sense
- There is something identical to it

and any other feedback. Thank you! Love you! xoxoxoxo

Luca Visentin (@[email protected])

A 🧵 [1/n] So, I was thinking about FAIR data. I wanted to start cataloguing our lab's data and make it FAIR, even just locally (with the idea that in the future we can start uploading somewhere). So I wanted to find some software that can bind metadata and data in some way. I found a series of LIMS software that seems to do what I want (e.g. makes protocols, data types, authors, instrument pages and links them all together) [...] #FAIR #OpenScience #help #researchsoftware

genomic.social

Tommi Mäklin Jul 10, 2023

Wrote a blogthing about how chatgpt is pretty bad at being a self-replicating machine https://maklin.fi/post/computer-science/2023-07-08_llms-and-self-replicating-machines/

Tommi Mäklin

Are large-language models a type of self-replicating machine? Probably, but very lousy ones.

Tommi Mäklin Jul 9, 2023

Alberto Perez-Posada Jul 9, 2023

People on the #FOSS community: the world of Open Source Software is vast an well documented, with numerous past and present examples of communities commited to promote and educate its philosophy and principles. But what about #hardware ? Is there anything similar about the hardware? I’d like to start reading more on sustainable hardware (i.e. devices that can be repaired, that have long term support, software designed to keep supporting them, sustainable materials, etc). Thanks!

Tommi Mäklin Jul 9, 2023

Philipp Bayer Jul 9, 2023

Most baffling thing about 'geek culture' is how ultra-conservative it is. People who grew up shouting "hack the planet" grew up to be averse against the smallest changes, from the small scale ("well it works for *me* so why should we change this 15 step setup") to the large ("i refuse to think about the social impacts of my work")

Tommi Mäklin Jul 8, 2023

David Lohner Jul 8, 2023

This. Is. Hilarious.

The PhD-Simulator

https://research.wmz.ninja/projects/phd/index.html

#PhDs #phdlife #IchBinHanna #IchBinReyhan #WissZeitVG #AcademicChatter

PhD Simulator

A completely random event based text game. Simply make your choice at the beginning of each month and see if you can graduate in time. All outcomes are determined by the random number generator and do not take them seriously. Sometimes the RNG can be brutal :)

Mianzhi Wang

Tommi Mäklin Jul 8, 2023

Bede Constantinides Jul 7, 2023

Preprint: How best to remove human reads from microbial FASTQs? Our tool Hostile removed >99.6% of human reads while retaining >99.997% and 100% of simulated reads in bacterial and mycobacterial metagenomes
https://www.biorxiv.org/content/10.1101/2023.07.04.547735v1

Tommi Mäklin Jul 3, 2023

Qagggy

Jun 22, 2023

2008 v. 2023: I don’t think people realize the extent to which parked cars degrade public space.

Tommi Mäklin Jul 3, 2023

James Pistole Jul 1, 2023

If the server costs are too much, have they considered #Twitter by mail? One of these bad boys can fit 1440000 / 280 = 5142 tweets! Just mail one out once a week! That’s 734 tweets a day! Problem solved!