Mastodon Feed: Posts

Mastodon Feed

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

@NfNitLoop yep. it turns out some fundamental norms break when the thing generating the code is not the thing that experiences the consequences of it breaking. we have all these programming paradigms and yet none of them anticipates "what if pure nonsense is in the arena"

Mastodon Feed

Boosted by NfNitLoop ("Cody Casterline 🏳️‍🌈"):
MaryAnnJohanson@mas.to ("MaryAnn Johanson (she/her)") wrote:

If you've ever said "Fuck ICE" on social media, congrats: You're on a government watchlist.

Fuck ICE. And fuck Trump and the fascist oligarchs he rode in on, too. #FuckICE #FuckTrump

https://www.rawstory.com/social-media-subpoena/

Mastodon Feed

Boosted by NfNitLoop ("Cody Casterline 🏳️‍🌈"):
dbattistella@mstdn.ca ("DB 🌱💦") wrote:

The man who wrote "Don't be evil" said he chose it specifically so it would be hard to remove. Paul Buchheit, the engineer who later built Gmail, suggested the phrase at a Google corporate values meeting on July 19, 2001.

Then in early 2018, internal documents leaked showing that Google had signed a Pentagon contract to build AI to analyze drone footage. By April, over 3,000 Google employees had signed a letter to Sundar Pichai demanding the contract be cancelled.

The letter specifically cited "Don't be evil" as the standard the company was failing to meet. Dozens of engineers resigned in protest.

Sometime between late April and early May, the slogan disappeared from the code of conduct's preface.

#Google #AI #Tech

[![Screenshot of a Wikipedia page: Don't be evil "Don't be evil" was Google's former motto, and a phrase used in Google's corporate code of conduct.[1[213 In 2018, Google removed the original motto from the preface of its code of conduct, [°]](https://files.mastodon.social/cache/media_attachments/files/116/664/159/577/507/603/original/938a10ed12de32a0.png)]7

Mastodon Feed

Boosted by NfNitLoop ("Cody Casterline 🏳️‍🌈"):
Natasha_Jay@tech.lgbt ("Natasha :mastodon:🇪🇺") wrote:

A Brand Manager getting ready for Pride Month 🌈

by Smooth Dunk

8 panel comic Panel 1: Yellow guy is working on a laptop in the background. In the foreground is a Brand Manager. Yellow Guy: “Sir, shall I put up our rainbow profile picture?”  Panel 2: Brand Manger snaps his head round and hisses “Cool your jets, Eugene. I’m checking the Profitometer”  Panel 3: Gonna do my best to explain this. We see The Profitometer. A huge machine that whirs and ticks. A digital display at the top of the machine reads “Supporting Pride equals”Beneath this display is a huge gauge with “Max Loss” on the left and “Max profit” on the right. There is a needle that is pointing exactly half way between the two extremes.  Panel 4: Close up on Brand Manager’s face. His teeth are gritted. Sweat trickles down his brow.  Panel 5: Close up on gauge. The needle remains exactly central  Panel 6: Close up on Brand Manager’s eyes. They are concentrating intensely. Also, more sweat.  Panel 7: Close up on the gauge. The needle clicks ONE DEGREE towards profit  Panel 8: Brand Manager screams like a football coach “GO GAY EUGENE”

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

RE: https://neuromatch.social/@jonny/116667022863059853

"The tests are theater".

Goodbye TDD, hello TDT - Theater Driven Testing.

I have seen this professionally. Hundreds of lines that test that the data that was passed into mocks comes out of the mocks. Tests that test nothing important. Or don't test what they claim.

But if you say "Hey, these tests look pretty useless, let's delete?", someone will chime in with "Tests are good! We want code coverage!"

Rather than argue, it's easier to just delete the tests when some refactor breaks them. 😮‍💨

Mastodon Feed

Boosted by NfNitLoop ("Cody Casterline 🏳️‍🌈"):
baldur@toot.cafe ("Baldur Bjarnason") wrote:

> Without significant changes to how these systems are developed and deployed, they will continue to undermine international human rights laws and standards. Whether proprietary or open source, these systems require urgent attention to establish clear frameworks for accountability and human rights protection before their impacts become more deeply entrenched in society.

https://www.amnesty.org/en/documents/pol40/0996/2026/en/

This is a pretty good report overall. Hard to disagree with any of the conclusions.

Mastodon Feed

Boosted by glyph ("Glyph"):
iris_meredith ("Iris") wrote:

RE: https://mastodon.sprawl.club/@ludicity/116672511881939477

If you possibly can help this person, please do so: I can second that they are very cool, and unlike me, they're geographically situated in a sensible time zone that'll let them collaborate effectively and have US working rights.

Mastodon Feed

Boosted by glyph ("Glyph"):
iris_meredith ("Iris") wrote:

LinkedIn is so strange. You see people who will be reviled for committing some of the worst crimes against humanity ever in half a century posting, and you're expected to treat it as "thought leadership"

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

@dysfun Oh noes, my million-monkeys-typing-into-a-statistical-model machine banged on some buttons that I didn't want it to. I'd better blame the passing butterfly that diverted their attention, instead of the boss who hired said monkeys.

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

@dysfun A friend ran into this on a random web search today.

At first we all had a good chuckle about it. "Please paste this RCE into your shell to prove you are a stupid human."

But then someone suggested that maybe LLMs were the target.

Screenshot: Complete these verification steps: 1. Press Command + Space to open Spotlight Search. 2. Type "Terminal" and press  Return. 3. In opened Terminal, paste the command ( Command + V ) and press  Return. You will observe and agree: | am not a robot - reCAPTCHA Verification ID: 915930" Press the verify button to proceed.

Mastodon Feed

Gargron ("Eugen Rochko") wrote:

I've reuploaded with fixed audio and updated the link in the post.

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

So I'll stick to Rust for now.

But it was fun "thinking in Zig" for a while. (Check out the Zigglings tutorial for a nice interactive tutorial!) And I wish the Zig team luck. I look forward to seeing where it goes!

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

However, after using #Rust, Zig's lack of memory safety features is a deal breaker for me. I was reading some Zig docs (which now I can't find) that recommended always documenting the owner when passing parameters by pointer, so you know who should free some memory.

Yes! That's what Rust's ownership is for! It forces me to document that explicitly in every case, in a way that the compiler can check for me, instead of me having to load and keep all of that information in my brain.

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

Given so many shared principles, I spent some time giving Zig another look this weekend.

It's definitely improved quite a lot since the last time I tried it out. It's of course a giant improvement over C. The way it handles generics via complie-time code generation is pretty cool. Though it feels *very* weird to see types saved to `const` variables. 😆

Mastodon Feed

Gargron ("Eugen Rochko") wrote:

If you learned something you didn't know from the video, please comment!

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

* A big drawback of LLMs is that they're centrally controlled by a few companies.
* Software made by corporations has a hostile relationship with users.
* Using nondeterministic LLM tools is worse than using deterministic tools that you can trust. ("I still have to review the code.")

Andrew also has some great tips for dealing with burnout. 😄

Mastodon Feed

isagalaev ("Ivan Sagalaev :flag_wbw:") wrote:

@NfNitLoop also was mostly nodding in agreement through the whole thing. Like he pointed out this technology would be interesting if it wasn't trapped by four big companies.

Mastodon Feed

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

If you can get past the clickbait headline and editing, this is a good interview with the creator of #Zig, Andrew Kelley.

A lot of what he said in that interview aligns with my own thoughts:
* AI contributions are low effort and low-value, if not *negative* value.
* A big part (Andrew says "the main point") of code review is mentorship.
* "Vibecoding blogs are a snoozefest"
* (re: LLMs working "surprisngly well") "I don't want to be *surprised* by the absence of a bug."

https://www.youtube.com/watch?v=iqddnwKF8HQ

Mastodon Feed

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
mothcompute@merping.synth.download ("security expert") wrote:

new spectre mitigation. this branch predictor always tells the truth and this branch predictor always lies

Mastodon Feed

Gargron ("Eugen Rochko") wrote:

After doing a couple of private workshops on how to use #Mastodon, I've decided I could just record a series of tutorials and have everyone benefit from it. Mastodon has a lot of features, and sometimes even long time users aren't aware of all of them! The first episode is about making posts on Mastodon--going over every aspect.

https://www.youtube.com/watch?v=zCQsRNt9pqg

Mastodon Feed

Boosted by cstanhope@social.coop ("The Luddites were right"):
brohrer@recsys.social ("Brandon Rohrer") wrote:

Spirit fuel for me are the toots of you who are building your own compiler, your own operating system, your own shell, your own frontend framework, your own language. Just for the pleasure and the education and the hell of it. Your journey is my jam.

Mastodon Feed

Boosted by cstanhope@social.coop ("The Luddites were right"):
troublewithwords@wandering.shop ("Steven Hoefer") wrote:

Curb Cut Effect would be a pretty good band name.

Mastodon Feed

fribbledom ("muesli") wrote:

GitHub Copilot got 25x more expensive over night.

I just felt a great disturbance in the force, as if millions of low-effort PRs suddenly cried out and were silenced.

Maintainers worldwide breathe a sigh of relief. Businesses are about to wake up to some very interesting invoices.

Mastodon Feed

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

#monsterdon I miss the days when you could get broken media working by futzing with it instead of "welp, SD card is dead"

Mastodon Feed

Boosted by kornel ("Kornel"):
Kensan wrote:

RE: https://mstdn.ca/@teledyn/116652708401285794

Mastodon Feed

fromjason ("fromjason.xyz ❤️ 💻 ✍️ 🥐 🇵🇷") wrote:

I made a LinkedIn-specific video and I feel dirty lol. I'm taking a swing in the dark at making content for other freelancers. Idk how I feel about it. I dont want to spend a lot of time on LinkedIn

Mastodon Feed

jwz wrote:

"Blink twice if Zuckerberg is an asshole"

Facebook whistleblower Sarah Wynn-Williams was forced to sit in silence on stage at an event at Hay festival, after lawyers advised her not to speak because of ongoing legal action brought by Meta....
https://jwz.org/b/yk8F

Screenshot

Mastodon Feed

Boosted by kornel ("Kornel"):
fribbledom ("muesli") wrote:

After enough years in software, you learn that LGTM doesn't actually stand for "Looks Good To Me".

Only the pros know its real meaning:

"Looked. Got tired. Merged."

Mastodon Feed

jsonstein@masto.deoan.org ("Jeff Sonstein") wrote:

‘Great replacement politicians promote intent-based definitions that characterize a policy as “discriminatory” only when it has explicitly racial language expressing the intention to discriminate. Intent-based definitions allow great replacement theorists to imagine that the policies and programs devoid of racial language but leading to racial inequity are “not racist,” while policies and programs created to reduce racial inequity are “racist” because they explicitly use racial language.’

from “Chain of Ideas” by Prof. Kindi

Mastodon Feed

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

When he's underwater, does he get wet? Or does the water get him instead?

#monsterdon #ultraman