Mastodon Feed: Posts

Boosted by adam@social.lol ("Adam"):
neatnik@social.lol ("Neatnik") wrote:

url.town has surpassed 800 entries! If you haven’t yet visited the omg.lol community’s little home-grown web directory, stop by and poke around! It’s a delight.

https://url.town

dysfun@treehouse.systems ("gaytabase") wrote:

sudo

Boosted by zkat@toot.cat ("Katerina Marchán"):
drmambobob@ecoevo.social ("Raptor's Nest (He/Him)") wrote:

There’s a massive misunderstanding about how Japanese eat noodles. We don’t slurp it up, we inhale it. There’s a huge difference. The former makes a wet slurpy sound, but the latter is a sharp airy sound.
#japan #ramen #food

dysfun@treehouse.systems ("gaytabase") wrote:

warning: viewer indiscretion advised

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
adactio ("Jeremy Keith") wrote:

Journal: Mistrust

How Apple’s penchant for breaking the web has given me more empathy towards developers who are suspicious of the web platform.

🔗https://adactio.com/journal/22507

dysfun@treehouse.systems ("gaytabase") wrote:

light pen (finger)

Boosted by zkat@toot.cat ("Katerina Marchán"):
RnDanger@infosec.exchange ("2xfo") wrote:

Stolen to add alt text

Boosted by zkat@toot.cat ("Katerina Marchán"):
mttaggart@infosec.exchange ("Taggart") wrote:

The Linux Foundation getting in bed with Coinbase to develop a web payments standard.

No thanks very much

https://www.linuxfoundation.org/press/linux-foundation-is-launching-the-x402-foundation-and-welcoming-the-contribution-of-the-x402-protocol

Boosted by zkat@toot.cat ("Katerina Marchán"):
expensive@shitpost.trade ("Expensive Shitpost LLC") wrote:

this account is quantity over quality. we throw shitposts at the wall till they stick.

Boosted by zkat@toot.cat ("Katerina Marchán"):
andnull@social.nouveau.community ("... and and and and ...") wrote:

I really like how watching the uxn channel on the concatenative discord has really left a permanent mark in my brain. Me and the rest of the mods basically let it operate as an automous zone in the discord. I even had it muted for the most part. Unbenounced to everyone outside that channel however, an AI booster and vibecoder had joined.

In a matter of a month or so, he killed the whole channel. Strangled it with his refusal to think. Every word someone said was feed straight into an LLM. Then, thrown back at chat cause the LLM could predict the correct code, and he could understand anything.

Eventually everyone grew so tired they began leaving. One fucking person leeched off a whole community. Never. Again. Ever inch you cede to LLM contributions in your community is another step towards its implosion. There is a reason so much of the art community has vehemently rejected whole sale.

But programmers need their new toys. And programmers demand every space follows industry trends and recommendation. And programmers cannot handle the idea of someone saying "no, we don't do that here". Cause nobody gets to say No in the software world. Things are always forced to evolve, consequences be damned.

I made the mistake once. I will not be making it twice. Do not help build the new car depedent infrastructure or everything you care about will be paved over.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
mhoye@cosocial.ca wrote:

An argument I've been stewing on for a while is that the place stochastic language modelling will be most effective as a dowsing rod for _classes_ of both errors and overlooked needs and growth-paths. I think that @cwebber is partly correct but that attacker/defender is a reactive framing; not wrong, but: narrow.

We have a mirror here that can show us ourselves, reflected in patterns. What would we want to see if we could see anything? How would we change?

https://social.coop/@cwebber/116335331872866600

dysfun@treehouse.systems ("gaytabase") wrote:

5 days of not smoking. starting to breathe a little easier, which is to say breathing at all up to a point.

very mixed feelings tbh, but inclined to consider this is a permanent change now.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
mnl@hachyderm.io ("mnl mnl mnl mnl mnl") wrote:

@aredridel one thing that grounds me is that this is mostly an online keyboard-driven debate. Normal people in the real world a) don't give a shit where the software comes from b) are delighted when you build them a little tool that solves their problem c) feel empowered when they realize they can build bigger things.

And now that software copyright is dead I legit dgaf anymore. I'm not going to launder opensource software for the sake of pretending I made something, but I'm sure as hell going to launder anything proprietary/company backed I run across.

legit where is the "hack the planet" spirit? opensource has been coopted by bigtech for decades now, every opensource project landing page looks like a fucking startup, but the stuff llm-assisted devs do is _weird af_ and quirky and creative. And usually deeply human, mirroring the creators personality, which is so ironic.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
mnl@hachyderm.io ("mnl mnl mnl mnl mnl") wrote:

@aredridel as if copyright as a hack to open more sourcecode through the GPL has worked in the last 20 years. All I've seen in practice is that there companies will rather build shitty in-house replacements of GPL'd software than open any of their own source.

I am absolutely flabbergasted, as someone who wants everybody to get value out of their computational devices, to get them to do what they want when they want on their own terms, and seeing how LLMs facilitate exactly the kind of "a single html file with local storage application that i share with my friends" malleable software so often dreamed of, to then be "but i have copyright on this pedestrian piece of software that presumably is 'remixed' by LLMs so no." and then attacking everybody studying and using the tools as "hubris driven tech bros wanting to keep their toys" when no, I want everybody to play with my toys.

dysfun@treehouse.systems ("gaytabase") wrote:

hi i'm nilegreen and this week i'll be synthesising computrons out of some dangerous chemicals and last week's jam.

dysfun@treehouse.systems ("gaytabase") wrote:

ludicrous as in wouldn't really be acceptable to most people rather than really good, obviously.

dysfun@treehouse.systems ("gaytabase") wrote:

c'mon, you knew i would have some absolutely ludicrous desktop setup, obviously.

dysfun@treehouse.systems ("gaytabase") wrote:

finally got around to putting a fan on my cpu cooler. it makes quite a difference....

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
mnl@hachyderm.io ("mnl mnl mnl mnl mnl") wrote:

@aredridel definitely something driving me up the walls. The hype-AI and the anti-AI debate is... accepting the same framework that AI tools are labor-replacing thinking-replacing magical widgets. Either irredeemably bad, or unconditionally amazing.

It's however much easier to have a nuanced discussion with the hype-AI crowd, because by virtue of them using the tool are very aware of the current shortcomings of the technology and of the spectrum of skill using them.

aredridel@kolektiva.social ("Mx. Aria Stewart") wrote:

RE: https://mastodon.social/@danluu/116333185599071224

This. This is how it is.

Startups can ship slop faster than ever _and that is the correct decision for them for the most part_. If we structure business so that they are existentially constrained like they are, this is the correct answer for how they will behave.

And if we want real engineering, the tool isn't the problem, it's the incentives.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
danluu ("Dan Luu") wrote:

@regehr @ryan This matches what I hear from friends in industry.

I hear a lot of stories of projects where someone outsourced thinking to an LLM and spent days or weeks debugging something and not fixing it when applying a bit of human judgment would solve the problem. A friend of mine makes good money as a consultant who cleans up people's LLM-assisted messes.

On the flip side, the growth rate of startups is faster than ever because you can ship at previously impossible velocities.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
regehr ("John Regehr") wrote:

My own experience is that it’s not at all easy to get reasonable code out of even the latest LLMs. Makefiles and that sort of thing— easy peasy. Medium or larger code you’d actually maintain? Very tricky

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
regehr ("John Regehr") wrote:

I’m not sure how to feel about this but I suspect it’s good news for CS types, it means all the managers and other random vibe coders aren’t really competition

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
regehr ("John Regehr") wrote:

this spring I've been teaching undergrads to use LLM agents. my rationale for doing this was that it would give me a chance to covertly teach lots of real software engineering, which is what I've done.

meanwhile, I've been watching the students closely to try to figure out whether a coding agent is a leveling factor (reducing differences in effectiveness between different students) or an anti-leveling factor (amplifying differences). at this point I'm 99% sure it's the second thing.

aredridel@kolektiva.social ("Mx. Aria Stewart") wrote:

God I hate ending up defending these tools because some of the takes are _so damn bad_. Like you could not do better as a psyop to advertise these tools and cement the ideological foundations.

aredridel@kolektiva.social ("Mx. Aria Stewart") wrote:

I knew it was this way but it's really hitting me today how much the Open Source movement and copyright maximalism supplanted the idea of free software, and again how much the Free Software movement turned from a close ideological cousin of the remix and open culture movement into a culture of legalism. At the same time, copyright law itself has been extended to be near-immortal copyrights rather than brief monopolies to spur creation by enabling profit from creating works.

pzmyers@freethought.online ("pzmyers 🕷") wrote:

Trying to bomb a nation of over 90 million people into the stone age isn't going to happen, and ought not to be the purpose of a war. It's also not going to salvage his status here at home.

https://freethoughtblogs.com/pharyngula/2026/04/02/yay-the-war-is-over/

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
pluralistic@mamot.fr ("Cory Doctorow") wrote:

Youtube's notice-and-staydown system is Content ID, an incredibly baroque system that allows copyright holders (and people pretending to be copyright holders) to "claim" video and sound files, and block others from posting them. No one - not even the world's leading copyright experts - can figure out how to use this system to uphold copyright:

https://pluralistic.net/2024/06/27/nuke-first/#ask-questions-never

20/

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
aral@mastodon.ar.al ("Aral Balkan") wrote:

So Anthropic employees are using Claude Code to contribute AI-generated code to open source repositories and hiding the fact using their own internal “undercover mode”.

Totally trustworthy people.

(Any open source project that at the very least requires disclosure of AI-authored contributions should immediately ban Anthropic employees on principle.)

#AI #Anthropic #ClaudeCode #subterfuge

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
GossiTheDog@cyberplace.social ("Kevin Beaumont") wrote:

Probably going to get a viral blog out of this experience, I'm trying to report a 4tb exposed cloud bucket to a company using their responsible disclosure programme... but they replaced the people with a GenAI ticket system that refuses to discuss the case as it thinks exploring open buckets is unethical and against its rules.