Jarkko Sakkinen
Posts
4432Following
316Followers
471OpenPGP: 3AB05486C7752FE1
Jarkko Sakkinen
repeated
repeated
Fedi.Tips
FediTips@social.growyourown.services
You can use Mastodon on many retro computer systems:
Apple II - https://www.colino.net/wordpress/en/mastodon-for-apple-ii
Apple Macintosh (pre-OS X) - https://github.com/smallsco/macstodon
Commodore 64 - https://github.com/Havoc6502/MOStodon
Commodore Amiga - https://github.com/BlitterStudio/amidon
MS-DOS - https://github.com/SuperIlu/DOStodon
MVS - https://github.com/mainframed/BREXXTODON
Palm OS - https://github.com/knickish/heffalump
Windows 95 - https://github.com/meyskens/mastodon-for-workgroups
5
5
1
Jarkko Sakkinen
jarkko
Firefox has been using Rust since 2017 and about fifth of its code that was previously in C/C++ has been written in Rust (see Oxidation sub-project).
I'd expect slower cycle for Linux than Firefox. I think this would be likely truth for operating system with decades of legacy (e.g. BSD derivatives).
This makes me think that making any conclusions after two years is simply not a pragmatic mindset.
#rust #linux #kernel
I'd expect slower cycle for Linux than Firefox. I think this would be likely truth for operating system with decades of legacy (e.g. BSD derivatives).
This makes me think that making any conclusions after two years is simply not a pragmatic mindset.
#rust #linux #kernel
2
6
9
Jarkko Sakkinen
jarkko
First thought I had this morning was that I started wondering if you could make "ChatGPT - The Board Game".
Each player would have 4x4 grid that would be educated during game, taking inspiration from mechanical matchbox tic-tac-toe player from 60s. And decision making would based on the conclusions of the AI and player.
Simplifying machine learning feedback ideas could foster some cool ideas for board games ;-)
Each player would have 4x4 grid that would be educated during game, taking inspiration from mechanical matchbox tic-tac-toe player from 60s. And decision making would based on the conclusions of the AI and player.
Simplifying machine learning feedback ideas could foster some cool ideas for board games ;-)
1
0
1
Jarkko Sakkinen
jarkko
Edited 10 months ago
Fresh ideas for remote attestation (vs Intel and AMD CA):
https://pom.ata.network/
#confidential #computing #machinehood
https://pom.ata.network/
#confidential #computing #machinehood
0
0
0
Jarkko Sakkinen
repeated
repeated
Jarkko Sakkinen
jarkko
Edited 10 months ago
I have good news!
The best available documentation on Rust-Linux still exists (when compared to anything else available, including Linux own documentation):
https://www.nccgroup.com/us/research-blog/rustproofing-linux-part-14-leaking-addresses/
I emailed to NCC Group and they've had just changed the URL. It is four part series, and the first part contains the URLs for rest of this series.
Documentation can be written from either "how"-angle or "why"-angle. For me the "how"-angle is irrelevant because I always read the source code from end to end :-) So I don't get whole a lot from existing kernel documentation.
#rust #linux #kernel
The best available documentation on Rust-Linux still exists (when compared to anything else available, including Linux own documentation):
https://www.nccgroup.com/us/research-blog/rustproofing-linux-part-14-leaking-addresses/
I emailed to NCC Group and they've had just changed the URL. It is four part series, and the first part contains the URLs for rest of this series.
Documentation can be written from either "how"-angle or "why"-angle. For me the "how"-angle is irrelevant because I always read the source code from end to end :-) So I don't get whole a lot from existing kernel documentation.
#rust #linux #kernel
2
3
8
Jarkko Sakkinen
jarkko
Edited 10 months ago
Why NPC's suck in games if the AI is great?
When Fallout 4 came out I stopped playing story oriented games because I started to feel like that I'm in a universe consisting of FSM's, which based on quiz trigger animation sequences ;-) No joke. Have not played story oriented games after 2015.
I guess you could improve a lot of taking some ideas of caching, tessellation algorithms and occlusion culling and apply them to AI. Like make "Quake solution" for AI. The best version would be used for the nearest entities, as it doesn't matter if the dudes far away are drunk/high ;-)
When Fallout 4 came out I stopped playing story oriented games because I started to feel like that I'm in a universe consisting of FSM's, which based on quiz trigger animation sequences ;-) No joke. Have not played story oriented games after 2015.
I guess you could improve a lot of taking some ideas of caching, tessellation algorithms and occlusion culling and apply them to AI. Like make "Quake solution" for AI. The best version would be used for the nearest entities, as it doesn't matter if the dudes far away are drunk/high ;-)
0
0
1
Jarkko Sakkinen
jarkko
Everybody knows Phil Collins, i.e. the guy made gated reverb popular :-)
https://www.youtube.com/watch?v=m2siy25EWto
https://www.youtube.com/watch?v=m2siy25EWto
0
0
0
Jarkko Sakkinen
jarkko
I'm working as a contractor through my own company so I guess I put that to my social media profiles instead of who I have a contract with :-) I pay my own salary for myself starting from October.
Right now I'm learning basics of #Inkscape because I don't have money to pay for a logo :-) It won't be great but it will be my own.
I'm using #IBM Plex font in this. I got the understanding that it is open source but have to check if the licensing terms allow to use it in a company logo.
Right now I'm learning basics of #Inkscape because I don't have money to pay for a logo :-) It won't be great but it will be my own.
I'm using #IBM Plex font in this. I got the understanding that it is open source but have to check if the licensing terms allow to use it in a company logo.
0
0
4
Jarkko Sakkinen
jarkko
Where the love for programming language, tool or technology begins, engineering ends ;-)
Engineering is art an of identifying things that suck, and finding ways to make them suck less.
#engineering #programming
Engineering is art an of identifying things that suck, and finding ways to make them suck less.
#engineering #programming
1
2
3
Jarkko Sakkinen
jarkko
Edited 10 months ago
I wish XCOM 3 will arrive some day. Every time I start playing Phoenix Point I get far but at some point I have too many stations and ships, get frustrated and give up.
Something in-between Phoenix Point and XCOM 2 in terms strategy/arcade factor would be nice.
BTW, speaking of games, why do NPC's suck so much if AI is so great? I cannot play story oriented games at all these days, not even the most visually pleasing like Cyberpunk, because despite all the nice graphics, I'm surrounded by FSMs. They just trigger stupid scripts causing pre-sequenced behavior. Why this is not fixed now that AI takes over everything? :-)
Something in-between Phoenix Point and XCOM 2 in terms strategy/arcade factor would be nice.
BTW, speaking of games, why do NPC's suck so much if AI is so great? I cannot play story oriented games at all these days, not even the most visually pleasing like Cyberpunk, because despite all the nice graphics, I'm surrounded by FSMs. They just trigger stupid scripts causing pre-sequenced behavior. Why this is not fixed now that AI takes over everything? :-)
0
0
0
Jarkko Sakkinen
jarkko
Edited 10 months ago
Two years of being in the mainline is a short durationin kernel. Some feature patch sets take longer. For most part I'm wondering why people go crazy of not much happening on Rust side.
More patch sets like this are best possible effort for Rust enablement (and BTW, it is in C ;-) ):
https://lore.kernel.org/lkml/20240815173903.4172139-21-samitolvanen@google.com/
Once GCC-Rust catches up I'd expect more to happen as that will unlock for instance embedded toolchains and give at least theoretical chance to have something with Rust in a defconfig.
I'd contribute and/or wait another 3-5 years before making any fast conclusions. Linux is no worse in this than any other operating system, which decades of legacy to worry about. There is no unfortunately fast path here anywhere.
More patch sets like this are best possible effort for Rust enablement (and BTW, it is in C ;-) ):
https://lore.kernel.org/lkml/20240815173903.4172139-21-samitolvanen@google.com/
Once GCC-Rust catches up I'd expect more to happen as that will unlock for instance embedded toolchains and give at least theoretical chance to have something with Rust in a defconfig.
I'd contribute and/or wait another 3-5 years before making any fast conclusions. Linux is no worse in this than any other operating system, which decades of legacy to worry about. There is no unfortunately fast path here anywhere.
2
1
2
Jarkko Sakkinen
jarkko
Edited 10 months ago
In addition to the NCC Groups awesome article series RTIC demonstrates how a low-level architecture can be documented without requiring enormous stretch in understanding all the nitty gritty details of Rust:
https://rtic.rs/2/book/en/
RTIC is a hard real-time OS framework. The documentation has exactly the domain information that helps to match the core concepts to the Rust artifacts and quickly also grasp those.
#rust #linux #kernel #rtic
https://rtic.rs/2/book/en/
RTIC is a hard real-time OS framework. The documentation has exactly the domain information that helps to match the core concepts to the Rust artifacts and quickly also grasp those.
#rust #linux #kernel #rtic
0
1
2
Jarkko Sakkinen
repeated
repeated
daniel:// stenberg://
bagder@mastodon.social
Edited 10 months ago
I (accidentally) removed two symbols from a curl header that have been deprecated and without functionality for over 17 years.
Result: of course there is still code out there using those that now reports build failures...
Keeping absolute backwards compatibility is real work.
2
1
1
Jarkko Sakkinen
jarkko
Edited 10 months ago
At the gates of hell - The Lake of Hell national park
0
1
3
Jarkko Sakkinen
repeated
repeated
allison
aparrish@friend.campi got so angry after reading this paper on LLMs and African American English that i literally had to stand up and go walk around the block to cool off https://www.nature.com/articles/s41586-024-07856-5 it's a very compelling paper, with a super clever methodology, and (i'm paraphrasing/extrapolating) shows that "alignment" strategies like RLHF only work to ensure that it never seems like a white person is saying something overtly racist, rather than addressing the actual prejudice baked into the model
1
5
1