kesor

6 ms·

31.
by kesor 5mo ago
It is just a Linux device. Other people will install NixOS on it anyway, and use specializations if the whole idea of swapping device roles in-and-out is viable. I don't really understand why would the team that already got a full plat
32.
by kesor 5mo ago
Instead of re-inventing Linux distributions for FlipperOS on top of Debian. They should just choose to base it on NixOS which already has these "profiles" as a built-in feature called "Specializations" https:/&#x2F
33.
by kesor 5mo ago
People have been sent to the gulag by the Canadian government as well, not too long ago. Governments are disgusting.
34.
by kesor 6mo ago
So you have some folders with markdown files ... which are insanely hard to query without a tool ... impossible to traverse via their relationships ... and you call that a graph database? WHAT?! Clicked the link expecting to see some tool o
35.
by kesor 7mo ago
160,000 if you take the revisions into account
36.
by kesor 10mo ago
I would argue that Desktop Linux finally took off because of Steam Proton, and because of Windows 10/11 and macOS starting version fartascular or whatever their versions are named.
37.
by kesor 10mo ago
OpenSCAD can load STLs and cut holes in them.
38.
by kesor 10mo ago
Diffusion models for images are already pretty much binary code generators. And we don't need to treat each bit individually, even in binary code there are whole segments that can be tokenized into a single token. Regarding training, w
39.
by kesor 10mo ago
Why would you think its more complex? There are less permutations of generating transistor on/off states than there are all the different programming languages in use that result in the exact same bits. Who said that creating bits effi
40.
by kesor 10mo ago
You are missing the whole point of the article.
41.
by kesor 10mo ago
You, like many others, seem to imply that humans write "good" code, but they absolutely do not--unless they are running some million dollar team with checks and cross checks and years of evolving the code over failures. I've
42.
by kesor 10mo ago
Why you hate on prose? This article has been a joy to read, unlike a lot of the other slop on the internet.
43.
by kesor 10mo ago
Binary bits are also a language. A structured language that transistor-based computers execute into some result we humans find valuable. Why wouldn't a model be able to write these binary instructions directly? Why do we need all these
44.
by kesor 10mo ago
Probably because then it wouldn't be software anymore. That is because you will be required to do a physical process (print an integrated circuit) in order to use the functionality you created. It can definitely be done, but it takes i
45.
by kesor 10mo ago
I don't think there is a need for an output language here at all, the LLM can read and write bits into executables directly to flip transistors on and off. The real question is how the input language (i.e. prompts) look like. There is
46.
by kesor 10mo ago
Don't train them to emulate humans. Train them to emulate compilers instead.
47.
by kesor 10mo ago
No need for using assembler, LLMs should be writing machine bits directly into executables. It is also fairly simple to teach them how to do it. Provided we have some executable that was written in some human-readable computer language. We
48.
Michael Burry Foundations: My 1999 (and part of 2000)(michaeljburry.substack.com)
2 points by kesor 11mo ago|0 comments
49.
DoiT Alliance with Ingram Micro, Reshapes the Value of FinOps Worldwide(finance.yahoo.com)
1 points by kesor 11mo ago|0 comments
50.
DoiT and Ingram Micro Team on FinOps for AWS Partners(channelinsider.com)
1 points by kesor 11mo ago|0 comments
51.
by kesor 1y ago
So ... dating apps just need to remove text messages and replace them with voice messages. Problem solved. Thanks!
52.
by kesor 1y ago
This is just like vibe coding. In vibe coding, you snapshot the results of the LLM's implementation into files that you reuse later. This project could use something like that. Perhaps ask the LLM to implement a way to store/cache
53.
Reinforcement Fine Tuning a Pangu Model(youtube.com)
2 points by kesor 1y ago|0 comments
54.
Reasoning with Sampling: Your Base Model Is Smarter Than You Think(arxiv.org)
2 points by kesor 1y ago|0 comments
55.
by kesor 1y ago
CueCat it was an affordable barcode scanner that anyone could have connected to their computer, and it scanned barcodes. It took almost two decades before we could finally do it again with our mobile phones.
56.
by kesor 1y ago
Skype ; Because my R.I.P. grandma was using it to talk to her relatives overseas just like she would use a phone, but it didn't cost an arm and a leg (unlike phone calls).
57.
by kesor 1y ago
ICQ ; It was the first instant messenger, the technology could have adopted voice (and not get disrupted by Skype) and mobile (and not get disrupted by whatsapp) and group chat (and not get disrupted by slack/discord). But they didn&#x
58.
by kesor 1y ago
Geocities ; It was a "put your html here" Free web hosting back when people barely knew what html was. Today you have to be a rocket scientist to find a way to host a free static "simple" page online.
59.
by kesor 1y ago
Pivotal Tracker ; Users loved it, it had an excellent model for tracking work and limiting work in progress on software projects. There is no real good alternative and the usual suspects for tracking project work are horrible in comparison.
60.
by kesor 1y ago
Google Wave ; It had a bunch of agents participating in editing the text together with you, making spelling fixes, finding additional information to enrich your content, and so much more.

More ›