Show more

Are you a FLOSS maintainer? What does the current state of LLMs do to you, personally? Does it help you? Does it give your sleepless nights? How do you feel about companies using your code to train their models?

I'm looking for some (quotable) feelings for a slide for my hackumenta talk (CCC) - Agent leck Aier (Agent lick b**ls), the influence of LLMs on our society.

Edit: And tell me if its OK to attach your name and a link to your comment :) - default: anonymous quote

Pls boost or tag a maintainer (who is OK with being tagged).

#php #rust #zig #floss #foss #OpenSource #FreeSoftware

People talking about "the worst security hole of the year" are being optimistic tbh

Delivery robots scan faces and collect customer biometric data, lawsuit alleges - mashable.com/tech/delivery-rob if they don't, they soon will #privacy #surveillance

Judge rejects Trump’s wild claim EPA can dismiss data center pollution lawsuits - arstechnica.com/tech-policy/20 " Trump’s EPA can’t unilaterally dismiss SpaceXAI pollution lawsuit, court rules. "

✨Your Daily Reminder✨: We don't have to build technology to surveil and kill. We can build technology that helps us fight for a better future.

@bagder I have very mixed feelings about this. Yes, AI does find real bugs (turns out a lot of bugs and exploits fit common patterns, hooray humans). But I also think taking a dump of an AI conversation, or even a “summary” of the problem, and dropping it on the doorstep of a maintainer is…VERY RUDE.

It should be socially expected that a filer of a bug continue to write a succinct summary of the issue.

I tend to reject reports like this, not because the error is not real, but because increasing automation means that the bug WILL SOON BE FOUND AGAIN BY SOMEONE WITH BETTER MANNERS who will take the time to understand the problem and write a succinct paragraph to describe it. This latter filer is MY FRIEND, and someone I want to build a relationship with and reward by crediting with finding the bug.

If maintainers don’t reject slop-dumps (sorry, LLM-dumps), they will be DDoS’d by thousands of auto-filers. As AI bug reports get better, it must get easier to reject those who don’t put any effort into filing.

So today some of the worst people on earth using the most environmentally destructive tools on earth brought down a huge part of the internet by exploiting a potential problem found many years ago in PHP which was previously not practical for anyone to exploit.

It only became "practical" to exploit this problem because "AI" came along to automate the boring and repetitive work required to do so. Absurd amounts of computing power were applied specifically to break into other peoples' websites.

Now the same people who are behind this are trying to pretend that they're heroes for finding an exploit.

"AI uncovers critical vulnerabilities", their headline reads. But the vulnerability was only made to be critical by the very same people as hide behind that headline.

Many thousands of websites are offline, many people are losing money, many thousands of people are working hard to try to patch or work around the problem, many small guys are finding it almost impossible to patch, and we're all supposed to be thankful to OpenAI and others involved ? Nah. Fuck off.

Everything about "AI" involves a lot of externalized costs, but I've still yet to see even a single positive thing that has been achieved by "AI". That has definitely not changed today.
#php2shell #php #fuckai #openai

There is a dog in the park. I am jealous, for the dog knows nothing of php

@drahardja I am an extremely stupid man, certainly less calibre than the IA could hire, and I fixed this shit without asking *any* LLMs. Not even my local ones. Come on guys.

Don’t let the door hit your ass on the way out, Zuck

VOTE YES ON PROP 40

VOTE NO ON PROP 41 and 42 (poison pills on 40).

businessinsider.com/meta-mark-

“Mark Zuckerberg considered leaving California over the billionaire tax”

#threateningMeWithAGoodTime #FuckZuck #California #YesOnProp40 #NoOnProp41 #NoOnProp42

eMudhra's press release tries to insinuate that AI is somehow partly to blame for their CA fuckup:

« Last weekend, an independent researcher found that the weak keys could be reverse-engineered possibly with AI-assisted techniques. »

In fact, the researcher just ran all the CA certs through an existing key linter: github.com/google/paranoid_cry (not actually expecting to find anything!)

emudhra.com/en/press-release/e

"The video shows a plainclothes agent wearing a red flannel shirt pursuing the man with a handgun, followed by two masked agents in tactical gear. I’m a U.S. citizen the man tells the agents as they surround him. When asked why he was running, he responds, “Because I don’t know who you guys are.” sfgate.com/bayarea/article/ice

Print this on an index card, stick it in your pocket, and take it to the polls with you when you vote.

18 US Code §592: Troops at Polls, prohibited by law.

[You can copy the text in the Alt-txt box directly.]

I am reminded again how in 1935, the US Congress passed a tax law that created a new top tax bracket of nearly 80%.

That tax bracket contained exactly one person, John D. Rockefeller, and was designed to stop him from causing so much trouble.

In case anyone needs the precedent.

@ketan Morally this is somewhere in the ballpark of the car company promising to set "nature lovers free" when they buy the big-ass SUV and have a blast running it roughshod all over the woodland and prairie. At some point the driver stops, gets out and pets a small plant, then gets back in the SUV and floors it. This was from ~20ya.

Show more
Librem Social

Librem Social is an opt-in public network. Messages are shared under Creative Commons BY-SA 4.0 license terms. Policy.

Stay safe. Please abide by our code of conduct.

(Source code)

image/svg+xml Librem Chat image/svg+xml