Mastodon Feed: Posts

Boosted by glyph ("Glyph"):
gsuberland@chaos.social ("Graham Sutherland / Polynomial") wrote:

and I want to be really really clear here: good software is not a label. you do not label a thing as good software. no committee or organisation decides that a thing is good software or not. break out of that mindset.

good software is the process and philosophy of making software that is good and does good.

Boosted by glyph ("Glyph"):
gsuberland@chaos.social ("Graham Sutherland / Polynomial") wrote:

and if your immediate thought is "we can solve this with licenses" then you are looking at it the wrong way.

licenses are the language of state power, which you as an individual almost certainly cannot wield. you cannot rely on their ambient deterrence as a social norm because your opponents are exactly the sorts of people who either don't give a shit about social norms or actively get off on violating them.

good software isn't about enforcement. it's about the process of doing good for people

Boosted by glyph ("Glyph"):
gsuberland@chaos.social ("Graham Sutherland / Polynomial") wrote:

we need to start a new movement called good software. software that is good, and does good, for the public good.

the freedom part is the wrong focus. it's a misaligned value judgement. you can write free software that bombs children. you can have a free software project maintained by a fascist or a rapist. software that bombs children, by the very definition, can never be good. projects where fascists and rapists are allowed into positions of power are, by the very definition, not good.

Boosted by glyph ("Glyph"):
malwaretech@infosec.exchange ("Marcus Hutchins :verified:") wrote:

Machine Speed is a lie: stop trying to fight AI with AI

Marketing departments are screaming that attackers are moving at "machine speed" and only AI-powered defenses can stop them. But cyberattacks have always been automated, and proactive defense beats reactive AI.

https://malwaretech.com/2026/09/machine-speed-is-a-lie-stop-trying-to-fight-ai-with-ai.html

jonny@neuromatch.social ("jonny (nonvenomous)") wrote:

RE: https://dair-community.social/@timnitGebru/117209767659620534

the thing the AI vendors actually sell is Brute Force as a Service

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
timnitGebru@dair-community.social ("Timnit Gebru (she/her).") wrote:

I'm posting here what I had written to a news outlet in April during the whole Mythos hype 🙄 and people keep on being like BUT why are you not making CLEAR arguments against Bla BLa Bla? Cause I'm learning that its a trap to make me waste my time debunking claims while they get trillions to implement their dystopian future and we're stuck talking about whatever they're doing rather than implementing our visions. 🧵

Boosted by kornel ("Kornel"):
WeirdWriter@caneandable.social ("Robert Kingett") wrote:

RE: https://mas.to/@carnage4life/117206712672621535

Genius!

fribbledom ("muesli") wrote:

1Password sponsoring Omarchy is a match made in heaven!

A password manager for idiots, sponsoring a Linux distro for noobs.

EmilyEnough@hachyderm.io ("Emily 🏳️‍🌈🏳️‍⚧️") wrote:

I am here for the incredibly niche STL memes.

jwz wrote:

I pay my parking tickets like any other Corporate Person: a million dollars at a time.

Waymo has been slapped with more than 8,300 parking tickets since January 2025, with fines of nearly $1 million for various violations.

Waymo spokesperson Chris Bonelli told Gazetteer in an email that the company pays parking tickets "like any other driver".
https://jwz.org/b/yk%5Fe

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
emma@orbital.horse ("ƒ: 🪿️→ ℝ") wrote:

Big 💪️ to Wikimedia Workers! Congrats on your union! https://www.wired.com/story/wikipedia-workers-vote-on-whether-to-unionize/

jscalzi@threads.net ("John Scalzi") wrote:

"'AI' queries per almond" is a weird fucking measurement

ireneista@irenes.space ("Irenes (many)") wrote:

the traps of individual words don't have this problem, because they only have to work with specific words that we already know how they're implemented

neatnik@social.lol ("Neatnik :prami:") wrote:

omg.lol Office Hours for Asia and Australia, and the USA starting in just about an hour! https://home.omg.lol/office-hours

ireneista@irenes.space ("Irenes (many)") wrote:

that's a little annoying because we want to put a prefix in front of the word body, to run before it, but we want to do that after the whole body is in place. that's fine though, our compilation semantics explicitly allow code to be slid around after it's generated; that's how our high-level control flow works.

ireneista@irenes.space ("Irenes (many)") wrote:

because at the start of a word definition, we don't yet know whether it's going to be a Forth word or an assembly word. it's only when we get to the end and see whether it's ; or ;asm that we know that.

ireneista@irenes.space ("Irenes (many)") wrote:

okay, yeah, if we really want to trace every call, this is going to require putting it in hex-semicolon-alternate not hex-colon-alternate (the definition of a Forth word begins with : and ends with ;, and the hex transform provides its own "alternate" implementations of these and other word-defining words)

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
ial_2@sciences.re ("Yaël Moussouni") wrote:

The #IgNobel Prize of #Technology is awarded to a team from China, Canada, USA, Kuwait nad Egypt, for pioneering the use of mosquito proboscises as 3D printing nozzles, a process known as "necroprinting".

The #Olfaction (!) Prize is awarded to a team from Poland, Germany and Sweden, for collecting evidence that babies smell wonderful to their parents; but teenagers do not.

Boosted by cstanhope@social.coop ("Your weary 'net denizen"):
poleguy ("poleguy looking for lost tools") wrote:

How can this article only have the word "environment" in the context of "environment variables."

No Elephant in this room.

https://teachmeidea.com/ethical-considerations-using-ai-tools-in-coding/

Who's behind this top hit? I haven't looked, but I bet it's a shill with big money paying for search placement.

#AI #ethics

Dear #lazymastodon who has links/resources on reading about ethical concerns of AI use?

Boosted by firepoet@tech.lgbt ("Stephen 🌈 (he/him)"):
YukiDeer@furry.engineer ("YukiDeer :verifiedtrans: 🇫🇷 🇺🇦🍉") wrote:

And, I am looking at the generous community of Fedi yet again with a request to help me find a job.

I am an IT Support for L2/L3 for SaaS and Infrastructure with 2 years of experience. Work primarily with Linux/Windows and some MacOS devices (though I can ultimately learn it all, I am a fast learner)

Worked for both small startups and enterprises (for example Cameco, BNP Paribas and so on)

My stack: Windows Server 2016-2022, Azure AD, MS Office 365, Azure,
Microsoft Intune, Linux (Debian, RHEL, CentOS), Mac OS X, PowerShell, Bash, Python, REST
API, MongoDB, MySQL, Atlassian Confluence, VMware vSphere, Docker, AnyDesk, Jira, ITIL,
ServiceNow, SalesForce, Apache, SCMS, DNS, DHCP, TCP/UDP routing, LAN/WLAN, HTML

I keep my CSAT at around 85-90%

I speak 3 languages in professional fluency: English, French and Ukrainian

My last experience is in 2024, since I have moved. Couldn't work for a while. Open for all offers, ideally in France and EU.

For more info: YukiDeer.01 (Signal)

#fedihire #lgbt

fribbledom ("muesli") wrote:

"Word wrapping is easy. Just count characters and split on spaces, right?"

Sure.

Until you account for Unicode grapheme clusters, terminal cell width, combining characters, ZWJ, emoji, CJK, tabs, ANSI sequences, OSC 8 hyperlinks, indentation, hard/soft breaks, unbreakable words, hyphenation, and trailing whitespace.

And then a scrollbar enters the chat.

Easy.

ireneista@irenes.space ("Irenes (many)") wrote:

wait, do we need to use our compiled-trap infrastructure for this.... we're also the interpreter (the hex transform, like all our transforms, takes control by running instead of the usual Forth interpret word), we may be able to observe it more directly... hmm

glyph ("Glyph") wrote:

okay uh… wait. what does this function _do_

>>> socket.getfqdn('google.com.')
'pnsfoa-af-in-x0e.1e100.net'

ireneista@irenes.space ("Irenes (many)") wrote:

it's hard to trap exit because it's written in machine code, and our internal ABI isn't set up to understand the idea of calling from machine code into Forth code

we did already solve that problem with a different word, though, so we can just imitate what we did there

ireneista@irenes.space ("Irenes (many)") wrote:

either way, we're going to have to solve the problem of trapping exit, the word which "returns" from words that are implemented in Forth (as opposed to machine-code words)

"returns" is in quotes because of the usual thing about the Forth threaded-execution model. words don't so much return as they jump directly into whatever their caller wants to do next.

ireneista@irenes.space ("Irenes (many)") wrote:

because we already have a way to trap every single word, observing calls is easy, and we could notionally have the hex transform keep its own dynamic state about that

... but it might wind up being easier to use the stack-walking tooling we already have, which is used to pretty-print stack traces on segfaults

Boosted by glyph ("Glyph"):
gabe@mendeddrum.org ("Gabe") wrote:

@glyph That's amazing! I've got the same IP address on my luggage!

Boosted by jwz:
mattblaze@federate.social ("Matt Blaze") wrote:

So there are basically two ways for this to go, and I honestly have no idea which one it will be:

(sane option): Everyone recognizes that this can't happen, and the EO has no effect until the courts have ruled on the merits (after November)

(chaos option): USPS refuses to accept/deliver blank ballots in the coming weeks, despite having no mechanism for states to comply, and we end up in completely uncharted territory.

@heidilifeldman

ireneista@irenes.space ("Irenes (many)") wrote:

... we want to do the most conceptually difficult remaining piece, which is the logic to generate comments for properties that are only practical to determine by observing the whole call stack at compile-time

ireneista@irenes.space ("Irenes (many)") wrote:

okay. we dusted off the Evocation code we were working on a couple days ago and finished out that task. the hex transform now outputs appropriate comments for string literals and alignment padding.

we could make it do integer literals as well with the same methodology, but those are a little hard to disambiguate from ... everything else ... so before we do that...