[-] self@awful.systems 57 points 1 month ago

fucking wild you busted out a Dollar Tree word like abstruse but came here to brag about how you didn’t read the article because you couldn’t understand its extremely simply worded headline

[-] self@awful.systems 79 points 2 months ago* (last edited 2 months ago)

this is a gentle reminder to posters in this thread that the fediverse in general is nowhere near secure from an opsec perspective; don’t post anything that compromises yourself or us.

with that said, happy December 4th to those who celebrate. post commemorative cocktail recipes here.

e: remember, they call it the fediverse cause it’s full of feds

17
submitted 3 months ago by self@awful.systems to c/morewrite@awful.systems

we have a WriteFreely instance now! I wrote up a guide to why it exists, why it's so fucking janky, and what we can do to fix it.

1

this is somewhat of a bigger update, and it's the product of a few things that have been in progress for a while:

email

email should be working again as of a couple months ago. good news: our old provider was, ahem, mildly inflating our usage to get us off their free plan, so this part of our infrastructure is going to cost a lot less than anticipated.

backups

we now have a restic-based system for distributed backups, thanks to a solid recommendation from @froztbyte@awful.systems. this will make us a lot more resilient to the possibility of having our host evaporate out from under us, and make other disaster scenarios much less lethal.

writefreely

I used some of the spare capacity on our staging instance to spin up a new WriteFreely instance where we can post long-form articles and other stuff that's more suitable for a blog. post your gibberish at gibberish.awful.systems! contact me if you'd like an invite link; WriteFreely instances are particularly vulnerable to being turned into platforms for spam and nothing else, so we're keeping this small-scale for instance regulars for now.

alongside all the ordinary WriteFreely stuff (partial federation, a ton of jank), our instance has a special feature: if you have an account, you can make a PR on this repository and once it's merged, gibberish will automatically pull its frontend files from that repo and redeploy WriteFreely. currently this is only for the frontend, but there's a lot you can do with that -- check out the templates, pages, less, and static directories on the repo to see what gets pulled. check it out if you see some jank you want to fix! (also it's the only way to get WriteFreely to host images as part of a post, no I'm not kidding)

what's next?

next up, I plan to turn off Hetzner's backups for awful.systems and use that budget to expand the node's storage by 100GB, which should increase the monthly bill by around 2.50 euros. I want to go this route to expand our instance's storage instead of using an object store like S3 or B2 because using block storage makes us more resilient to Hetzner or Backblaze evaporating or ending our service, and because it's relatively easy to undo this decision if it proves not to scale, but very hard to go from using object storage back to generic block storage.

after that, it'll be about time to carefully upgrade to the current version of Lemmy, and to get our fork (Philthy) in a better state for contributions.

as always, see our infrastructure deployment flake for more documentation and details on how all of the above works.

40
submitted 3 months ago by self@awful.systems to c/techtakes@awful.systems

this post has been making the rounds on Mastodon, for good reason. it’s nominally a post about the governance and community around C++, but (without spoiling too much) it’s written as a journey packed with cathartic sneers at a number of topics and people we’ve covered here before. as a quick preview, tell me this isn’t relatable:

This is not a feel good post, and to even call it a rant would be dismissive of the absolute unending fury I am currently living through as 8+ years of absolute fucking horseshit in the C++ space comes to fruition, and if I don’t write this all as one entire post, I’m going to physically fucking explode.

fucking masterful

an important moderator note for anyone who comes here looking to tone police in the spirit of the Tech Industry Blog Social Compact: lol

[-] self@awful.systems 43 points 4 months ago

really stretching the meaning of the word release past breaking if it’s only going to be available to companies friendly with OpenAI

Orion has been teased by an OpenAI executive as potentially up to 100 times more powerful than GPT-4; it’s separate from the o1 reasoning model OpenAI released in September. The company’s goal is to combine its LLMs over time to create an even more capable model that could eventually be called artificial general intelligence, or AGI.

so I’m calling it now, this absolute horseshit’s only purpose is desperate critihype. as with previous rounds of this exact same thing, it’ll only exist to give AI influencers a way to feel superior in conversation and grift more research funds. oh of course Strawberry fucks up that prompt but look, my advance access to Orion does so well I’m sure you’ll agree with me it’s AGI! no you can’t prompt it yourself or know how many times I ran the prompt why would I let you do that

That timing lines up with a cryptic post on X by OpenAI Altman, in which he said he was “excited for the winter constellations to rise soon.” If you ask ChatGPT o1-preview what Altman’s post is hiding, it will tell you that he’s hinting at the word Orion, which is the winter constellation that’s most visible in the night sky from November to February (but it also hallucinates that you can rearrange the letters to spell “ORION”).

there’s something incredibly embarrassing about the fact that Sammy announced the name like a lazy ARG based on a GPT response, which GPT proceeded to absolutely fuck up when asked about. a lot like Strawberry really — there’s so much Binance energy in naming the new version of your product after the stupid shit the last version fucked up, especially if the new version doesn’t fix the problem

[-] self@awful.systems 59 points 4 months ago

I love Blade Runner, but I don’t know if we want that future. I believe we want that duster he’s wearing, but not the, uh, not the bleak apocalypse.

there’s nothing more painful than when capitalists think they understand cyberpunk

58
submitted 5 months ago by self@awful.systems to c/techtakes@awful.systems

this article is about how and why four of the world’s largest corporations are intentionally centralizing the internet and selling us horseshit. it’s a fun and depressing read about crypto, the metaverse, AI, and the pattern of behavior that led to all of those being pushed in spite of their utter worthlessness. here’s some pull quotes:

Web 3.0 probably won’t involve the blockchain or NFTs in any meaningful way. We all may or may not one day join the metaverse and wear clunky goggles on our faces for the rest of our lives. And it feels increasingly unlikely that our graphic designers, artists, and illustrators will suddenly change their job titles to "prompt artist” anytime soon.

I can’t stress this point enough. The reason why GAMM and all its little digirati minions on social media are pushing things like crypto, then the blockchain, and now virtual reality and artificial intelligence is because those technologies require a metric fuckton of computing power to operate. That fact may be devastating for the earth, indeed it is for our mental health, but it’s wonderful news for the four storefronts selling all the juice.

The presumptive beneficiaries of this new land of milk and honey are so drunk with speculative power that they'll promise us anything to win our hearts and minds. That anything includes magical virtual reality universes and robots with human-like intelligence. It's the same faux-passionate anything that proclaimed crypto as the savior of the marginalized. The utter bullshit anything that would have us believe that the meek shall inherit the earth, and the powerful won't do anything to stop it.

1

we’ve exceeded the usage tier for our email sending API today (and they kindly didn’t email me to tell me that was the case until we were 300% over), so email notifications might be a bit spotty/non-working for a little bit. I’m working on figuring out what we should migrate to — I’m leaning towards AWS SES as by far the cheapest option, though I’m no Amazon fan and I’m open to other options as long as they’ve got an option to send with SMTP

[-] self@awful.systems 81 points 6 months ago

Copilot then listed a string of crimes Bernklau had supposedly committed — saying that he was an abusive undertaker exploiting widows, a child abuser, an escaped criminal mental patient. [SWR, in German]

These were stories Bernklau had written about. Copilot produced text as if he was the subject. Then Copilot returned Bernklau’s phone number and address!

and there’s fucking nothing in place to prevent this utterly obvious failure case, other than if you complain Microsoft will just lazily regex for your name in the result and refuse to return anything if it appears

[-] self@awful.systems 52 points 6 months ago

But don’t worry! Google’s AI summaries will soon have ads!

dear fuck, pasting ads onto the part of Google search that’s already known to be unreliable and annoying at best seems like a terrible idea. for a laugh, let’s see if there’s any justification for this awful shit in the linked citation

Ads have always been an important part of consumers’ information journeys.

oh these people are on the expensive drugs huh

69
submitted 6 months ago by self@awful.systems to c/techtakes@awful.systems

after the predictable failure of the Rabbit R1, it feels like we’ve heard relatively nothing about the Humane AI Pin, which released first but was rapidly overshadowed by the R1’s shittiness. as it turns out, the reason why we haven’t heard much about the Humane AI pin is because it’s fucked:

Between May and August, more AI Pins were returned than purchased, according to internal sales data obtained by The Verge. By June, only around 8,000 units hadn’t been returned, a source with direct knowledge of sales and return data told me. As of today, the number of units still in customer hands had fallen closer to 7,000, a source with direct knowledge said.

it’s fucked in ways you might not have seen coming, but Humane should have:

Once a Humane Pin is returned, the company has no way to refurbish it, sources with knowledge of the return process confirmed. The Pin becomes e-waste, and Humane doesn’t have the opportunity to reclaim the revenue by selling it again. The core issue is that there is a T-Mobile limitation that makes it impossible (for now) for Humane to reassign a Pin to a new user once it’s been assigned to someone.

90
submitted 6 months ago by self@awful.systems to c/techtakes@awful.systems
37
submitted 6 months ago by self@awful.systems to c/techtakes@awful.systems

as I was reading through this one, the quotes I wanted to pull kept growing in size until it was just the whole article, so fuck it, this one’s pretty damning

here’s a thin sample of what you can expect, but it gets much worse from here:

Internal conversations at Nvidia viewed by 404 Media show when employees working on the project raised questions about potential legal issues surrounding the use of datasets compiled by academics for research purposes and YouTube videos, managers told them they had clearance to use that content from the highest levels of the company.

A former Nvidia employee, whom 404 Media granted anonymity to speak about internal Nvidia processes, said that employees were asked to scrape videos from Netflix, YouTube, and other sources to train an AI model for Nvidia’s Omniverse 3D world generator, self-driving car systems, and “digital human” products. The project, internally named Cosmos (but different from the company’s existing Cosmos deep learning product), has not yet been released to the public.

46
submitted 7 months ago by self@awful.systems to c/techtakes@awful.systems

so Andreessen Horowitz posted another manifesto just over a week ago and it’s the most banal fash shit you can imagine:

Regulatory agencies have been green lit to use brute force investigations, prosecutions, intimidation, and threats to hobble new industries, such as Blockchain.

Regulatory agencies are being green lit in real time to do the same to Artificial Intelligence.

does this shit ever get deeper than Regulation Bad? fuck no it doesn’t. is this Horowitz’s attempt to capitalize on the Supreme Court’s judiciary coup? you fucking bet.

here’s some more banal shit:

We find there are three kinds of politicians:

Those who support Little Tech. We support them.

Those who oppose Little Tech. We oppose them.

Those who are somewhere in the middle – they want to be supportive, but they have concerns. We work with them in good faith.

I find there are three kinds of politicians:

  • those who want hamburger. I give them hamburger.
  • those who abstain from hamburger. I do not give them hamburger.
  • those who have questions about hamburger. I refer them to the shift supervisor in good faith.
[-] self@awful.systems 43 points 7 months ago

In April 2014, Gerard created a RationalWiki article about Effective Altruism, framing the subculture as “well-off libertarians congratulating each other on what wonderful human beings they are for working rapacious [s---]weasel jobs but choosing their charities well, but never in any way questioning the system that the problems are in the context of,” “a mechanism to push the libertarian idea that charity is superior to government action or funding,” and people who “will frequently be seen excusing their choice to work completely [f---]ing evil jobs because they're so charitable.”

it's fucking amazing how accurate this is, and almost a decade before SBF started explaining himself and never stopped

[-] self@awful.systems 60 points 7 months ago

At the same time, most participants felt the LLMs did not succeed as a creativity support tool, by producing bland and biased comedy tropes, akin to ``cruise ship comedy material from the 1950s, but a bit less racist''.

holy shit that’s a direct quote from the paper

[-] self@awful.systems 116 points 7 months ago* (last edited 7 months ago)

there’s this type of reply guy on fedi lately who does the “well actually querying LLMs only happens in bursts and training is much more efficient than you’d think and nvidia says their gpus are energy-efficient” thing whenever the topic comes up

and meanwhile a bunch of major companies have violated their climate pledges and say it’s due to AI, they’re planning power plants specifically for data centers expanded for the push into AI, and large GPUs are notoriously the part of a computer that consumes the most power and emits a ton of heat (which notoriously has to be cooled in a way that wastes and pollutes a fuckton of clean water)

but the companies don’t publish smoking gun energy usage statistics on LLMs and generative AI specifically so who can say

2
submitted 9 months ago by self@awful.systems to c/freeasm@awful.systems

it can’t be overstated how important the Nix evaluator is to the Nix ecosystem; it implements the Nix language and package manager, maintains the store, has a hand in the low-level workings of every Nix tool, and is the focus of the push by Eelco and friends to commercialize Nix and keep it appealing to military-industrial interests.

all of the above is why I joined the Aux CLI SIG, which focuses on maintaining a fork of the Nix evaluator for the Aux ecosystem. but just now I saw the announcement for Lix, a Nix evaluator fork that focuses on modernizing the codebase (including gradually replacing C++ with Rust), maintaining correctness (something the upstream evaluator has been notoriously struggling with lately), and doing right by its community. I found myself nodding along to their description of the project and feeling something I haven’t felt since I read the open letter — I’m finally feeling excited for the future of the technology behind Nix.

I have no idea if Lix will become Aux’s chosen evaluator fork, though the Aux CLI SIG can help determine that collectively (and I’ll have many more details on Aux in a post later tonight). here’s what’s truly exciting though: by following Lix’s install steps and pulling auxpkgs-unstable, we can have a package ecosystem and NixOS fork that’s completely independent of the Nix community, and we can have it right now. I’m so excited by that news that I’m going to spin up a host just to give Lix+auxpkgs a try later tonight.

here’s the Aux thread about Lix; so far, there’s a lot of high-level support and excitement for using it as Aux’s evaluator.

2
submitted 10 months ago by self@awful.systems to c/freeasm@awful.systems

this thread fucking sucks for me to have to post, but the linked open letter is an important read. none of the systemic issues pertaining to marginalized folks and commercial/military-industrial interests in the Nix community I’ve previously written about on TechTakes have been solved; in fact, they’ve gotten worse to the point where the Nix community moderation team is essentially in the process of quitting. that’s the beginning to an awful end for a project I like a whole lot.

even if you don’t give a fuck about Nix, the open letter is an important read because the toxicity, conflicts of interest, and underhanded tactics detailed in it are incredibly common in the open source space. this letter could have been written about a multitude of infamously toxic open source projects; Nix is lucky that it has marginalized folks involved who care about the direction of the project and want to make things better, but those people are actively leaving, after being burnt out by the toxic people and structures entrenched in Nix’s community. that’s a fucking tragedy.

50
submitted 10 months ago by self@awful.systems to c/techtakes@awful.systems

who could have seen this coming, other than everyone who told the homebrew tree inverter guy this was a bad idea they absolutely shouldn’t do

[-] self@awful.systems 52 points 10 months ago

"We pay our AI artist 15,000 USD per month for exactly 10 hours of work," reads an X post from the official Champions TCG account. "Why? In that time, he still makes HUNDREDS of AMAZING bits of artwork—ASTRONOMICALLY FASTER than ANY team of traditional artists.

[…]

The anonymous artist "has 15 years of digital art experience" and doesn't use social media, he says.

"For us to get this with a team of traditional artists it would cost us a lot more money, and time," said Malec. "The guy's a pro and he charges what he's worth. We are well connected in the space and no one comes close to the quality he delivers."

I’m convinced this is money laundering, and pcgamer is negligent for not putting “NFT card game” anywhere in the title for this one so people have sufficient warning of that before they read half the article

"His art is 100% AI generated, yet it has no extra fingers, no generic designs, no mistakes... It has consistent evolutions, skins, alt art styles—literally no one is on his level. We don't care how he makes it, we only care that the end user enjoys our game."

extreme disagree on these not being generic — they look like shit, to be honest, and most of them seem to be poorly framed on the card at the very least. according to the article, a bunch of them also have extra fingers and mistakes, so there’s that too.

Instead, the company challenged artists to complete a series of "art tests" in 48 hours, claiming that anyone who can match the quality of its AI prompt writer will be considered for a job as their assistant.

aaaand there it is. do unpaid labor for us to prove your worth in this ridiculously unfair competition, and if you’re lucky you’ll get the chance to make less than minimum wage photoshopping the extra fingers out of thousands of shitty AI generated card images

[-] self@awful.systems 80 points 10 months ago

the secret sauce is always hiding labor exploitation behind a thick layer of bad ideas

[-] self@awful.systems 63 points 11 months ago

for anyone who wants to increase Amazon’s GPT bill by generating dildo limericks, it looks like this is only enabled for Amazon’s app, not their website

view more: next ›

self

joined 2 years ago
MODERATOR OF